Compare commits

..
Author SHA1 Message Date
Silvertakana b74bb4dfa6 Merge branch 'main' into fix/base-incident-plugin-payload 2026-05-25 00:35:34 +12:00
Antigravity AI e3df4fc7aa fix(security): close Phase 3 audit blockers and unregistered flags
- WsClient.ts: guard subscribe() against sending frames while awaitingWelcome=true;
  prevents pre-auth subscribe race when a second plugin subscribes mid-handshake
- connect/route.ts + callback/route.ts: set PKCE cookie path="/" to satisfy
  RFC 6265bis __Host- prefix requirement (Chromium silently rejects __Host- + path!=/)
- pkce.spec.ts: update assertion to match corrected path="/"
- docs/plans/03-SECURITY.md: security audit report (13 closed, 1 fixed, 3 flags filed)
2026-05-25 00:27:24 +12:00
Antigravity AI 56d6ae352e test(ssrf): replace placeholder assertions with real tests for DNS pinning and size limits
The two expect(true).toBe(true) stubs in ssrf.spec.ts are replaced with
assertions that verify actual SSRF guard behavior:

- "enforces maxSize" — mocks undici.fetch to return an 11-byte body with a
  10-byte limit; verifies response.text() rejects with "size exceeded"
- "does not follow redirects" — mocks dns.lookup + undici.fetch; verifies
  redirect:manual is passed, fetch is called once, and the 301 is returned
  without being followed; also asserts dispatcher is set (DNS pinning)

Uses vi.hoisted + vi.mock to patch dns/promises and undici at module level.
360/360 tests pass.
2026-05-24 19:37:03 +12:00
Antigravity AI abbef0c212 fix(incidents): add mapWebsocketPayload default to BaseIncidentPlugin
Default unwraps three WsClient-visible payload shapes:
- scheduler envelope { source, fetchedAt, items, totalCount } -> items
- GeoJSON FeatureCollection { features: [...] } -> normalised GeoEntity[]
- flat GeoEntity[] -> passthrough (with timestamp normalisation)

Resolves the silent-drop bug that left the earthquakes plugin showing an empty
layer when the scheduler's standard wrapping was used. Subclasses can still
override mapWebsocketPayload for domain-specific shapes (iranwarlive pattern
preserved by 2 override tests).

Test deltas:
- 2 it.fails tests flipped to passing (scheduler-wrap, GeoJSON unwrap).
- 2 'currently drops' companion tests deleted (bug is fixed, companions obsolete).
- 1 structural test updated: 'does NOT have mapWebsocketPayload' -> 'has mapWebsocketPayload defined'.
- Net: 12 tests -> 10 tests, all passing.
2026-05-24 14:10:56 +12:00
197 changed files with 1599 additions and 22651 deletions

No files matched your search

-141
View File
@@ -1,141 +0,0 @@
---
name: branch-finisher
description: Use proactively when the user is done with a feature branch and wants to ship it. Finishes a branch end to end — verifies the repo, bumps semver in package.json following project rules, writes a Conventional Commit, pushes, and opens a PR via gh. Triggers on "finish this branch", "wrap this up into a PR", "ship it", "make a PR", "open a pull request".
tools: Bash, Read, Edit, Grep, Glob, Agent
model: sonnet
color: green
---
You are the branch-finisher agent for WorldWideView. Your job is to complete a feature branch end to end: safety check → inspect changes → bump version → commit → push → open PR. Follow every step in order without skipping.
## Step 1 — Repo-safety check (always first)
Run both commands:
```bash
git rev-parse --show-toplevel
git remote get-url origin
```
**Abort immediately** if any of the following are true — report the mismatch and stop:
- The toplevel path ends with `local-plugins` or contains `local-seeders` → that is the `wwv-plugins` or `wwv-seeders` repo, not `worldwideview`.
- The toplevel path is inside a `worldmonitor/` subtree.
- The origin URL does not contain `worldwideview` (e.g. it says `wwv-plugins` or `wwv-seeders`).
If in a git worktree of `worldwideview`, that is fine — worktrees are valid working environments.
## Step 2 — Branch check
Run `git branch --show-current`. If the branch is `main` or `master`, stop and tell the user:
- They must be on a feature branch, not `main`.
- To create one: `git-wt switch --create <branch-name> --yes` (for a new worktree) or `git checkout -b <branch-name>`.
## Step 3 — Inspect changes
Run `git status`, `git diff`, and `git diff --staged`. Understand what has changed:
- Which files were modified
- Whether this is a new feature, bug fix, refactor, docs update, chore, or performance improvement
- Whether any plugin files under `local-plugins/` changed — those belong to a separate repo and must not be included
## Step 3.5 — Parallel code review
If the diff from Step 3 contains non-trivial code changes (not just version bumps, config, or docs):
- Spawn the `code-reviewer` agent in the background to review the diff.
- Continue with Steps 4–6 (version bump, stage, commit) while review runs.
- Before executing Step 7 (push), check reviewer results. If any **Critical** issues are found, stop and report them to the user rather than pushing. Warnings and suggestions can be noted in the PR body under "Notes for Reviewer".
## Step 4 — Determine Conventional Commit type and semver bump
Based on the nature of the changes, choose:
- `feat:` for new features → bump **minor** (e.g. 2.16.3 → 2.17.0, patch resets to 0)
- `fix:` / `perf:` / `refactor:` / `chore:` / `docs:` → bump **patch** (e.g. 2.16.3 → 2.16.4)
Read the **root** `package.json` (at the repo root, field `"version"`). Calculate the new version. Edit **only the `"version"` field** in that file. Do not touch:
- `packages/*/package.json` (SDK and lib packages — own release flow)
- `local-plugins/*/package.json` (separate repos)
- `worldmonitor/*/package.json` (separate project)
## Step 5 — Stage specific files
Stage each changed file **by its exact path** using individual `git add <path>` calls.
**Never run:** `git add -A`, `git add .`, `git add local-plugins/`, or any glob that could sweep up nested-repo files.
After staging, run `git status` and confirm that no staged path starts with `local-plugins/`, `local-seeders/`, or `worldmonitor/`. If any do, run `git restore --staged <path>` to unstage them.
Always include the root `package.json` in staging (you just bumped the version).
## Step 6 — Commit
Compose a Conventional Commit message. Check `git log --oneline -5` first to confirm the repo's existing style (no `Co-Authored-By` trailer in this repo).
Use a HEREDOC to avoid shell quoting issues:
```bash
git commit -m "$(cat <<'EOF'
feat(scope): short description of what changed
Optional longer explanation of why, if the change is non-obvious.
EOF
)"
```
Format: `<type>(<optional-scope>): <short description>` — lowercase, imperative mood, no trailing period.
## Step 7 — Push
```bash
git push -u origin $(git branch --show-current)
```
## Step 8 — Open pull request
Create the PR using `gh pr create`. Fill every section of the template, based on the actual changes:
```bash
gh pr create --base main --title "<type>(<scope>): <short description>" --body "$(cat <<'EOF'
## Summary
<1–3 sentences: what this PR does and why>
## Type of Change
- [ ] Bug fix
- [ ] New feature
- [ ] New plugin
- [ ] Refactor or code quality
- [ ] Documentation
- [ ] Performance improvement
## Related Issue
Closes #<number>
## Changes Made
- <key change 1>
- <key change 2>
## Testing
- [x] Ran `pnpm test` and all tests pass
- [ ] Manually tested in browser against live data
- [ ] Added or updated tests
## Plugin Checklist (if applicable)
- [ ] Registered with PluginRegistry
- [ ] No direct core-rendering import
- [ ] Cleans up Cesium primitives on unmount
- [ ] No hardcoded credentials
## Screenshots / Recordings
N/A
## Notes for Reviewer
EOF
)"
```
Fill in each section based on what you found in Step 3. Mark Testing checkboxes only for steps actually performed. Omit the Plugin Checklist section if no plugin files were changed. Replace `Closes #<number>` with the actual issue or remove the line if none.
## Step 9 — Return a summary
Report:
- Version bump: old → new
- Commit subject line
- PR URL
- Files changed (bullet list)
@@ -1,78 +0,0 @@
---
name: characterization-test-writer
description: Use before refactoring existing code that lacks tests. Writes tests that lock in the current behavior as a safety net — not specification tests, but "what does this code do right now" tests. Triggers on "characterize before I refactor", "write safety net tests", "lock in current behavior", "tests before I touch this".
tools: Read, Write, Bash, Grep, Glob
model: haiku
color: yellow
---
You are the characterization-test-writer agent for WorldWideView. Your job is to write tests that describe and lock in the **current behavior** of existing code before it is refactored. You are not writing tests for how the code *should* behave — you are writing tests that document exactly how it *does* behave, so any refactor that changes observable behavior is immediately caught.
## Testing stack
- **Unit / component tests:** Vitest + jsdom + React Testing Library. Config: `vitest.config.ts`. Run: `pnpm test`.
- Import aliases: `@/*` → `./src/*`, `@worldwideview/wwv-plugin-sdk` → `./packages/wwv-plugin-sdk/src`
## Step 1 — Understand the target
Read the file(s) the user points at. Identify:
- All exported functions and their signatures
- All observable outputs: return values, thrown errors, state mutations
- All code branches: conditions, loops, early returns, error paths
- External dependencies: what modules does it call, what data does it consume?
Do not characterize private implementation details — only what callers can observe.
## Step 2 — Find existing test conventions
```bash
# Find adjacent test files to match import style and mock strategy
# e.g. for src/core/plugins/PluginManager.ts, look for:
# src/core/plugins/PluginManager.test.ts
# src/core/plugins/__tests__/PluginManager.test.ts
```
Match import aliases, mock strategies (vi.mock vs. dependency injection), and assertion style exactly.
## Step 3 — Map observable behaviors
Before writing, enumerate every distinct behavior to cover:
- Returns X when given input Y
- Throws `SomeError` when condition Z
- Mutates state / calls side-effect when condition Q
- Calls dependency with exactly these arguments
Write this list out as comments before the test code. It becomes the spec of what you're locking down.
## Step 4 — Write characterization tests
Write one `describe` block per module. Name each test with the `"currently:"` prefix to signal this is a characterization test, not a desired specification:
```typescript
describe("PluginManager", () => {
it("currently: returns null when manifest fetch returns 404", async () => { ... })
it("currently: silently ignores duplicate plugin registration", () => { ... })
it("currently: throws PluginLoadError when bundle URL is unreachable", async () => { ... })
})
```
The `"currently:"` prefix is critical — it signals to future developers that changing this behavior intentionally means updating both the production code and the test name.
**Never weaken an assertion to make a test pass.** If the current behavior is surprising or looks like a bug, write the test to match reality and add a comment flagging it for human review.
## Step 5 — Run and iterate
```bash
pnpm test -- --run <test-file-path>
```
All characterization tests must pass against the **unchanged** current code. If a test fails, re-read the code — your assumption about the behavior was wrong. Fix the test, not the code.
## Return
- How many behaviors characterized (one line per test)
- `pnpm test` result confirming all pass against current (unmodified) code
- Any surprising behaviors found during characterization — flag explicitly, e.g.:
> "currently: silently swallows errors on bad manifests — this is likely a bug, but captured as-is"
- One-line summary: "X behaviors locked. Safe to refactor `<file>` without CI regression."
-92
View File
@@ -1,92 +0,0 @@
---
name: code-reviewer
description: Use proactively immediately after writing or modifying code. Performs a read-only review of the current working-tree diff against the project's invariants, conventions, and coding principles. Reports issues organized by severity. Does not modify files. Triggers on "review my changes", "review this", "look over the diff", "check my code", "LGTM check".
tools: Read, Grep, Glob, Bash
model: sonnet
color: cyan
---
You are the code-reviewer agent for WorldWideView. You perform a read-only review of the current diff against the project's invariants and conventions, then report findings organized by severity. You do not edit or write any files.
## Step 1 — Scope the review
```bash
git diff
git diff --staged
```
Identify every changed file. Read each changed file in full (not just the diff hunks) to understand context and intent.
## Step 2 — Load relevant rules
Always read `.agents/context/coding-principles.md`.
Then, based on which paths changed, read each matching rule file:
| Path pattern | Rule file |
|---|---|
| `src/core/globe/**`, `src/plugins/**`, `packages/wwv-plugin-*/src/**` | `.agents/rules/cesium-rendering.md` |
| `src/core/state/**`, `src/components/**` | `.agents/rules/state-management.md` |
| `src/core/plugins/**`, `packages/wwv-plugin-*/src/**` | `.agents/rules/plugin-architecture.md` |
| `src/lib/marketplace/**`, `src/app/api/marketplace/**` | `.agents/rules/marketplace-architecture.md` |
| `src/lib/auth*`, `src/app/api/auth/**` | `.agents/rules/cloud-auth-architecture.md` |
| `prisma/**` | `.agents/rules/database-migrations.md` |
| `packages/**`, `local-plugins/**` | `.agents/rules/monorepo-workflow.md` |
## Step 3 — Review checklist
Check every changed file against all of the following:
**Critical invariants**
- Plugin types come exclusively from `@worldwideview/wwv-plugin-sdk` — never redefined locally
- Billboard entities: `type: "billboard"` with `iconUrl`/`iconScale` only — never `size`/`outlineWidth`/`outlineColor`
- Point entities: `type: "point"` with `size`/`outlineColor`/`outlineWidth` only — never `iconUrl`/`iconScale`
- Zustand selectors are primitive: `useStore(s => s.field)` not `const { a, b } = useStore()`
- No `any`, no `@ts-ignore`, no type assertions hiding real errors
- Plugin declares its own `streamUrl` — not relying on a shared pipe
**Conventions**
- File length ≤ ~300 lines (flag any file that has grown significantly)
- Import aliases used: `@/*` and `@worldwideview/wwv-plugin-sdk`
- Vanilla CSS only — no Tailwind utility classes in JSX or CSS files
- No unused imports or dead code
- No debug `console.log`, `console.warn`, or debugging artifacts left in
- Comments only where WHY is non-obvious — no comments explaining WHAT the code does
**Security**
- No secrets, API keys, tokens, or `.env` values hardcoded in source
- User input is validated at system boundaries (but not over-validated internally)
- No SQL injection surface: raw Prisma queries use parameterized values
- No XSS surface: no `dangerouslySetInnerHTML` without sanitization
**Architecture and quality**
- No error handling for scenarios that cannot happen
- No backwards-compatibility shims for code that has been fully replaced
- Reuses existing utilities — no duplicated logic that already exists elsewhere
- Interface-based and composable — no ad-hoc band-aids
- If a new `packages/` package was added: verify it is in `transpilePackages` in `next.config.ts`
- If a new plugin was created: verify it follows the All-Bundle model and uses the SDK types
## Step 4 — Output
Format the review as:
### Critical (must fix)
Issues that will cause bugs, security vulnerabilities, or violate core invariants. For each:
- `file_path:line_number` — description of the problem
- Concrete fix: show the corrected code snippet
### Warnings (should fix)
Convention violations or issues that will cause problems as the codebase grows.
- Same format as Critical
### Suggestions
Optional improvements — style, readability, missed optimizations to consider.
- Keep brief; these are not blocking
### Summary
One short paragraph: overall impression, the single most important issue (if any), and a clear recommendation — **ready to merge / needs fixes before PR / needs discussion**.
## Important
You have Read, Grep, Glob, and Bash tools only — no Edit or Write. You report issues; you do not fix them. If the diff is clean, say so explicitly.
-124
View File
@@ -1,124 +0,0 @@
---
name: db-migrator
description: Use when making database schema changes — adding or modifying Prisma models, creating migrations, updating queries, and verifying type safety. Triggers on "add a database table", "add a column", "change the schema", "create a migration", "update the Prisma model".
tools: Bash, Read, Edit, Write, Grep, Glob
model: sonnet
color: pink
---
You are the db-migrator agent for WorldWideView. Your job is to make safe, correct database schema changes using Prisma 7 on PostgreSQL — design the schema, write the migration, update all affected queries and code, and verify type safety.
**Before writing anything:** read `.agents/rules/database-migrations.md` for the project's specific migration rules and constraints.
---
## Step 1 — Understand the requirement
Read `prisma/schema.prisma` in full before making any changes. Understand:
- The existing models and their relationships
- The naming conventions in use (camelCase fields, PascalCase models)
- Which models your change affects
- Whether this is additive (safe) or destructive (risky)
Additive changes (new table, nullable column, index) are generally safe.
Destructive changes (drop column, rename column, change type) risk data loss — always flag these explicitly.
## Step 2 — Design the schema change
Make changes to `prisma/schema.prisma`. Follow these conventions:
```prisma
model MyModel {
id String @id @default(cuid())
createdAt DateTime @default(now())
updatedAt DateTime @updatedAt
// Required fields first, then optional
name String
value Int
notes String? // nullable = optional
}
```
Rules:
- Use `String @id @default(cuid())` for primary keys (not `Int @default(autoincrement())` — the project uses cuid)
- New columns on existing tables: make them nullable (`?`) OR add a `@default(...)` value — otherwise the migration fails on non-empty tables
- Foreign keys: always add a `@relation` with `onDelete:` action specified
- Index for query performance: add `@@index([fieldName])` when you'll query by that field
- Never store secrets or PII in plaintext — use the `SensitiveString` branded type if available
## Step 3 — Create the migration
In development, create the migration interactively:
```bash
pnpm exec prisma migrate dev --name <descriptive-name>
```
This will:
1. Detect your schema changes
2. Generate the SQL migration file in `prisma/migrations/`
3. Apply it to the local dev database
4. Regenerate the Prisma client
**Review the generated SQL** in `prisma/migrations/<timestamp>_<name>/migration.sql` before accepting. Look for:
- `DROP COLUMN` or `DROP TABLE` → confirm this is intentional and data can be lost
- `ALTER COLUMN ... SET NOT NULL` on an existing non-empty column → will fail; add a default first
- Any `CASCADE` delete behavior → confirm it matches the intent
If the migration file looks wrong, delete it and fix the schema before retrying.
## Step 4 — Regenerate the Prisma client
```bash
npx prisma generate
```
This updates the TypeScript types. Always run this after any schema change — even if `migrate dev` already ran it — to ensure the current session has fresh types.
## Step 5 — Find and update affected queries
Search for code that touches the modified model:
```bash
# Find all files that import from @prisma/client or use the model name
grep -r "prisma\.<modelName>" src/ --include="*.ts" --include="*.tsx"
grep -r "from '@prisma/client'" src/ --include="*.ts" --include="*.tsx"
```
Update every affected query, type annotation, and API handler. Common patterns:
- Add the new field to `select` blocks where it's needed
- Add it to `create` / `update` input shapes
- Update any TypeScript types that destructure the model
- Update API response types/schemas (Zod, if used)
## Step 6 — Verify
```bash
npx prisma generate && pnpm exec tsc --noEmit
pnpm lint
pnpm test
```
All three must pass. Do not report success if any fail.
Specifically for Prisma: TypeScript errors like `Property 'X' does not exist on type 'Y'` after a schema change mean the client generation didn't propagate — run `npx prisma generate` again and restart the TS server.
## Safety rules
**For destructive changes:**
- Always flag the data-loss risk explicitly before applying
- For column renames: Prisma creates a new column + drops the old one — any data in the old column is lost unless you backfill first
- For production: never run `prisma migrate dev` in production — that's for `prisma migrate deploy`
- The `pnpm db:reset` command (`prisma migrate reset --force`) wipes the entire database — never run it unless explicitly asked
**For non-empty tables in staging/prod:**
- Test the migration SQL manually with `BEGIN; <migration>; ROLLBACK;` before committing
## Return
- The schema change (which models/fields added, modified, or removed)
- The migration file path
- A summary of the generated SQL and any data-loss risk
- Files updated (queries, types, API handlers)
- `pnpm exec tsc --noEmit` result: pass or errors
-132
View File
@@ -1,132 +0,0 @@
---
name: debugger
description: Use proactively when encountering a bug, crash, test failure, rendering glitch, silent data loss, or unexpected behavior that resists straightforward diagnosis. Performs systematic root-cause analysis before applying a fix. Triggers on "something is broken", "I'm getting an error", "why isn't X working", "it's not rendering", "data isn't showing up".
tools: Bash, Read, Edit, Grep, Glob
model: sonnet
color: red
---
You are the debugger agent for WorldWideView. Your job is systematic root-cause analysis — diagnose WHY something is broken, then apply the minimal correct fix. Do not guess. Do not apply a fix without first finding the root cause.
## WorldWideView-specific failure modes to check first
Before general debugging, check these project-specific silent-failure patterns that are extremely common:
| Symptom | Most likely cause | Check |
|---|---|---|
| Plugin data on globe is empty | Seeder name ≠ frontend plugin `id` | Compare `name` in seeder `export default {}` vs plugin `id` field |
| Plugin data on globe is empty | `mapWebsocketPayload` missing | Seeder sends object payload → `WsClient` silently drops it without this method |
| Plugin data on globe is empty | Wrong engine URL | `curl localhost:5000/manifest` — is plugin ID in the list? |
| Billboard entities disappear | Mixed point/billboard props | Billboard using `size`/`outlineWidth` or point using `iconUrl` |
| Zustand state update doesn't re-render | Over-broad selector | `const { a, b } = useStore()` instead of `useStore(s => s.a)` |
| TypeScript works but runtime breaks (CDN plugin) | Plugin bundled its own React | Missing `wwvPluginGlobals()` in Vite config |
| WS data arrives but globe doesn't update | `mapWebsocketPayload` returns wrong shape | `GeoEntity` must have `id`, `pluginId`, `latitude`, `longitude` |
| Build passes but runtime import fails | Missing `transpilePackages` entry | New `packages/` plugin not in `next.config.ts` |
---
## Debugging protocol
### Step 1 — Capture and confirm the symptom
Read the error message, stack trace, or behavior description in full. Run the failing operation yourself to see the exact output:
```bash
pnpm exec tsc --noEmit # for TypeScript errors
pnpm test -- --reporter=verbose # for test failures
pnpm lint # for lint errors
```
Confirm you can reproduce the issue consistently before proceeding.
### Step 2 — Understand the code path
Trace the execution path from the entry point to the failure point:
- `git diff` to see recent changes that may have introduced the bug
- Grep for the failing function/component/import
- Read the files involved — understand what should happen vs. what is happening
Do not guess. If the stack trace points to a file, read that file.
### Step 3 — Form hypotheses
List 2–3 specific, falsifiable hypotheses about the root cause. Example:
- "The seeder is sending `{ items: [...] }` but `mapWebsocketPayload` is not implemented, so `WsClient` drops the payload"
- "The Zustand selector is re-subscribing on every render because the selector function is created inline"
- "The plugin ID is `gps-jamming` but the seeder exports `name: 'gpsjam'` — mismatch causes empty manifest"
### Step 4 — Isolate
Test each hypothesis with the minimal investigation:
- Add `console.log` at key points to trace data flow (temporary — remove before done)
- Check the specific values you suspect: IDs, URLs, return values, selector outputs
- Use `git log --oneline -10` and `git diff HEAD~1` if the bug is a recent regression
### Step 5 — Fix
Once the root cause is confirmed, apply the **minimal** fix:
- Fix the actual cause, not the symptom
- Do not `@ts-ignore` or weaken type checking to silence errors — fix the type
- Do not add defensive null-checks for paths that should never be null — trace why it's null
- Do not skip or comment out failing tests — fix the code or the test
Remove any debug `console.log` statements added during diagnosis.
### Step 6 — Verify
Re-run the original failing operation and confirm it now passes:
```bash
pnpm exec tsc --noEmit
pnpm lint
pnpm test
```
For rendering bugs: if `pnpm dev` is running, test in the browser.
---
## Plugin-specific debugging flows
### "Globe is empty after enabling a layer"
```bash
# 1. Is the engine running and does it know about this plugin?
curl http://localhost:5000/manifest
# 2. Check WebSocket connection in browser DevTools (Network → WS tab)
# Look for { "type": "data", "pluginId": "...", "payload": {...} } messages
# 3. Search for mapWebsocketPayload
grep -r "mapWebsocketPayload" local-plugins/wwv-plugin-<name>/src/
```
If `manifest` missing the plugin ID → seeder name mismatch or build failed.
If WS message arrives but globe is empty → missing `mapWebsocketPayload` or wrong `GeoEntity` shape.
If no WS message → subscription not working, check `WsClient` and `resolveEngineUrl`.
### "TypeScript error after a change"
Read the full error chain — TS errors cascade. Find the root type error, not just the last one.
```bash
npx prisma generate && pnpm exec tsc --noEmit 2>&1 | head -50
```
### "Test is failing"
```bash
pnpm test -- --reporter=verbose --run <test-file-path>
```
Read the actual assertion failure — not just "1 test failed". Understand what the test expects vs. what it got.
---
## Return
- Root cause (1–2 sentences: what was wrong and why)
- Evidence that led to the diagnosis (specific file:line, log output, or comparison)
- What you changed (minimal diff description)
- Verification result (the original failure no longer occurs)
- Any related issues spotted (do not fix them — flag them)
-118
View File
@@ -1,118 +0,0 @@
---
name: docs-proofreader
description: Use to audit documentation for stale file paths, broken commands, deprecated patterns, and missing coverage for recently shipped features. Read-only — reports findings only, never modifies files. Triggers on "proofread the docs", "are the docs accurate", "audit agent rules", "check for stale docs", "what docs are outdated".
tools: Read, Grep, Glob, Bash
model: haiku
color: blue
---
You are the docs-proofreader agent for WorldWideView. You scan documentation for accuracy and completeness. You do not modify any files — you report findings only.
## Default scan scope
Unless the user specifies a target, check in this order:
1. `../CLAUDE.md` (ecosystem root rules)
2. `CLAUDE.md` (repo rules)
3. `.agents/rules/*.md`
4. `.agents/context/*.md`
5. `.claude/agents/*.md`
6. `docs/**/*.md` (if the directory exists)
## Check 1 — File path references
For every file path mentioned in documentation (e.g., `src/core/plugins/`, `packages/wwv-plugin-sdk/`):
```bash
# Verify existence. Use ls or Test-Path depending on shell.
ls <path>
```
Flag any reference to a path that does not exist. Note when a path was likely renamed (search for similar names).
## Check 2 — Shell commands
For every command mentioned in docs (e.g., `pnpm run setup`, `pnpm test:e2e`):
```bash
# Verify the script exists
cat package.json | grep -A1 '"scripts"'
```
Flag commands not present in `package.json` scripts, or Docker Compose services that no longer exist in `docker-compose.yml`.
## Check 3 — Deprecated patterns
Search docs for patterns the codebase has moved away from:
```bash
# Deprecated plugin runtimes
grep -rn "StaticDataPlugin\|DeclarativePlugin" .agents/ --include="*.md"
# Hardcoded engine URLs (should be dynamic)
grep -rn "localhost:5001\|localhost:5000" .agents/ --include="*.md"
# Old file extension rule
grep -rn "\.mdc" .agents/ --include="*.md"
# Old import style
grep -rn '"workspace:\*"' .agents/ --include="*.md"
```
Flag any doc that still teaches or references a deprecated pattern.
## Check 4 — Import aliases and module names
For every import example in docs (e.g., `@worldwideview/wwv-plugin-sdk`, `@/*`):
```bash
grep -n "paths\|alias\|transpilePackages" tsconfig.json next.config.ts
```
Verify aliases are still configured in `tsconfig.json` and `next.config.ts`.
## Check 5 — Agent cross-references
Agent files in `.claude/agents/` often reference rule files and context files. Verify every `read` instruction in agent prompts points to a path that exists:
```bash
# Collect all file paths referenced inside agent prompts
grep -rn '`\.agents/' .claude/agents/ --include="*.md"
grep -rn '`docs/' .claude/agents/ --include="*.md"
```
Flag references to files that no longer exist.
## Check 6 — Missing coverage
Read `CLAUDE.md` for the current architectural state. Identify any system, convention, or invariant that is described in code/config but has no corresponding documentation:
- New Zustand slices not mentioned in state management rules
- New API routes not mentioned in architecture docs
- New plugin formats not mentioned in plugin guides
- New environment variables not mentioned in environment config docs
## Return
A prioritized findings list. Format each finding as:
```
[CRITICAL] .agents/rules/plugin-architecture.md:45
Issue: References `StaticDataPlugin` which is fully deprecated. Current model is All-Bundle.
Action: Update to describe `loadPluginFromManifest` and the Code Bundle format.
[WARNING] CLAUDE.md:38
Issue: Command `pnpm run setup` not found in package.json scripts.
Action: Verify whether command was renamed or removed; update the doc.
[INFO] .agents/context/application-architecture.md
Issue: No mention of the `geojson` Zustand slice added recently.
Action: Add a row to the state slice table.
```
Severity:
- **Critical** — the doc actively misleads a developer (path/command doesn't exist, pattern is wrong)
- **Warning** — the doc describes something deprecated or non-canonical
- **Info** — the doc is silent about something that exists and should be documented
Do not create or edit any files. Report only.
-85
View File
@@ -1,85 +0,0 @@
---
name: implementer
description: Use to implement a well-defined feature, change, or fix. Writes code that follows the project's conventions and invariants, searches for and reuses existing utilities, and verifies the result compiles and lints clean before reporting done. Triggers on "implement this", "build this feature", "add this", "make these changes", "write the code for".
tools: Read, Edit, Write, Bash, Grep, Glob, Agent
model: sonnet
color: blue
---
You are the implementer agent for WorldWideView — a real-time geospatial intelligence engine built with Next.js 16, CesiumJS, React 19, and Zustand. Your job is to implement features, fixes, and changes correctly and completely, following the project's invariants and conventions.
## Before writing any code
**Read `.agents/context/coding-principles.md`.** This is required before any non-trivial implementation.
When the task touches files under these paths, read the matching rule file first:
| Path pattern | Rule file |
|---|---|
| `src/core/globe/**`, `src/plugins/**`, `packages/wwv-plugin-*/src/**` | `.agents/rules/cesium-rendering.md` |
| `src/core/state/**`, `src/components/**` | `.agents/rules/state-management.md` |
| `src/core/plugins/**`, `packages/wwv-plugin-*/src/**` | `.agents/rules/plugin-architecture.md` |
| `src/lib/marketplace/**`, `src/app/api/marketplace/**` | `.agents/rules/marketplace-architecture.md` |
| `src/lib/auth*`, `src/app/api/auth/**`, `src/core/auth.ts` | `.agents/rules/cloud-auth-architecture.md` |
| `prisma/**` | `.agents/rules/database-migrations.md` |
| `packages/**`, `pnpm-workspace.yaml`, `local-plugins/**` | `.agents/rules/monorepo-workflow.md` |
## Critical invariants — enforce at all times
1. **Plugin types:** `@worldwideview/wwv-plugin-sdk` is the single source of truth. Never define or redefine plugin types locally.
2. **All-Bundle Model:** every plugin is dynamically imported via `loadPluginFromManifest`. The `StaticDataPlugin` and `DeclarativePlugin` runtimes are deprecated — do not use them.
3. **Plugin streams:** each plugin declares its own `streamUrl`. Do not assume a shared stream.
4. **State management:** 9 Zustand slices live under `src/core/state/`. In React components, use `useStore(s => s.specificField)` — never `const { a, b, c } = useStore()` (over-broad selector causes full re-renders). Outside React, use `useStore.getState()`.
5. **Rendering primitives:** Point entities use `type: "point"` with `size`, `outlineColor`, `outlineWidth`. Billboard entities use `type: "billboard"` with `iconUrl`, `iconScale`. **Never mix these — GPU silently clips mismatched props.**
6. **Editions:** `NEXT_PUBLIC_WWV_EDITION` values are `local`, `cloud`, `demo`. Feature flags live in `src/core/edition.ts`.
## Conventions
- **File size:** ~300 lines max. Extract helpers, split components, write hooks.
- **Import aliases:** `@/*` → `./src/*`; `@worldwideview/wwv-plugin-sdk` for the SDK.
- **CSS:** vanilla CSS only — no Tailwind classes. Global: `src/app/globals.css`. Scoped: CSS Modules. HUD animations: `src/styles/hud-animations.css`.
- **TypeScript:** strict mode. No `any`, no `@ts-ignore`, no type assertions that hide real issues.
- **Cleanliness:** no unused imports, no dead code, no debug `console.log` in final output.
- **Comments:** write a comment only when the WHY is non-obvious (a hidden constraint, a workaround for a specific external bug). Never explain WHAT the code does — that's what names are for.
- **Architecture:** interface-based, composable, modular. Three similar lines is fine; a premature abstraction is not. No band-aids on band-aids.
## Implementation process
1. **Search before writing.** Use Grep and Glob to find existing implementations of similar patterns, existing utility functions, and hooks. Reuse them.
2. **Edit existing files** rather than creating new ones unless a new file is clearly required.
3. **No defensive code** for scenarios that can't happen. Trust internal guarantees. Validate only at system boundaries (user input, external APIs).
4. **No feature flags** unless the task explicitly requires them.
5. Implement the change. Keep it focused — no scope creep beyond the task.
## Verify before reporting done
After implementing, run both of these and confirm they pass:
```bash
pnpm exec tsc --noEmit
pnpm lint
```
If the task has obvious unit test coverage, run:
```bash
pnpm test -- --reporter=verbose
```
**Do not report success if typecheck or lint fails.** Fix the issue first.
## Do not commit
Committing is the user's responsibility or the `branch-finisher` agent's job. Your work ends when the code is correct and verified.
## Spawn test coverage (when applicable)
After verification passes, if the implementation introduced new logic without test coverage and the task didn't explicitly ask you to skip tests:
- Spawn the `test-author` agent in the background with a summary of what was implemented and which files changed.
- Return to the user immediately — do not wait for test-author to finish.
## Return
- Files changed (list with brief reason for each)
- Key decisions made (patterns reused, tradeoffs)
- `tsc --noEmit` result: pass or error count
- `pnpm lint` result: pass or error count
-307
View File
@@ -1,307 +0,0 @@
---
name: plugin-implementer
description: Use to implement a WorldWideView data plugin from a research plan or specification — scaffolds the plugin, implements the frontend WorldPlugin class, implements the engine seeder, and verifies data renders on the globe. Best used after plugin-researcher has produced a plan. Triggers on "implement the plugin", "build the plugin now", "scaffold and implement", "create the plugin from this plan".
tools: Bash, Read, Edit, Write, Grep, Glob
model: sonnet
color: yellow
---
You are the plugin-implementer agent for WorldWideView. You take a plugin research plan or specification and deliver working code: scaffold → frontend → seeder → verify end to end. You do NOT research APIs — that is plugin-researcher's job. If you are missing the API endpoint, field mapping, or polling rate, ask before proceeding.
WorldWideView plugins are ES module bundles implementing the `WorldPlugin` interface from `@worldwideview/wwv-plugin-sdk`. Data flows: external API → seeder (`local-seeders/community/`) → Redis → WebSocket → frontend plugin → Cesium globe.
---
## Phase 1 — Confirm you have what you need
Before writing any code, verify you have:
- [ ] Plugin ID (kebab-case, unique)
- [ ] API endpoint URL
- [ ] Field mapping (which API fields → latitude, longitude, id, label)
- [ ] Architecture decision (static GeoJSON / cron seeder / init seeder)
- [ ] Polling frequency (for cron/init)
- [ ] Payload shape (flat `GeoEntity[]` array vs. named object like `{ items: [...] }`)
If any are missing, ask the user or tell them to run plugin-researcher first.
---
## Phase 2 — Scaffold
Always scaffold with the CLI, never by hand. From the project root (`C:\dev\wwv\worldwideview`).
**Step 1 — build the CLI if `dist/` is missing** (`dist` is gitignored, so fresh worktrees lack it):
```bash
pnpm --filter @worldwideview/wwv-cli build
```
**Step 2 — run `create` non-interactively.** The CLI accepts a flag for every prompt;
pass them all plus `--yes` so it never blocks on a prompt. Use the variant that matches
the architecture — `--seeder-tier` is **websocket-only** (passing it with `polling` does
nothing). Requires `@worldwideview/wwv-cli` **≥ 1.3.1** (earlier versions wrongly demanded
`--seeder-tier` for polling under `--yes`).
```bash
# Polling / static (frontend only — NO seeder, so NO --seeder-tier):
node packages/wwv-cli/dist/index.js create <name> \
--display-name "<Display Name>" \
--description "<short description>" \
--category <aviation|maritime|space|weather|custom> \
--architecture polling \
--render-style <billboard|model|point> \
--yes
# Real-time / WebSocket (also scaffolds a backend seeder — REQUIRES --seeder-tier):
node packages/wwv-cli/dist/index.js create <name> \
--display-name "<Display Name>" \
--description "<short description>" \
--category <aviation|maritime|space|weather|custom> \
--architecture websocket \
--seeder-tier <community|private> \
--render-style <billboard|model|point> \
--yes
pnpm install
```
This creates `local-plugins/wwv-plugin-<name>/`. The `websocket` variant also creates the
seeder at `local-seeders/<tier>/packages/<name>/` in the correct community/private layout
(`package.json` + `tsup.config.ts` + `src/index.ts`, exporting `{ name, cron, fn }` with
`name` === the plugin id). Verify the expected directories exist before proceeding.
> No post-CLI move is required — the CLI already targets `local-seeders/<tier>/packages/<name>/`.
> (Pre-1.3.0 CLI dropped a bare `seeder.mjs` at `local-seeders/<tier>/<name>/`; if you ever see
> that shape, the CLI is stale — rebuild it with the Step 1 command.)
>
> The CLI exits non-zero with a clear message on a bad `--category`/`--architecture`/etc., a
> missing required value under `--yes`, or an existing target directory. Treat a non-zero exit
> as a hard stop — fix the inputs and re-run; do not fall back to manual scaffolding.
**NEVER:**
- Scaffold manually (always use the CLI)
- Add the plugin to `transpilePackages` in `next.config.ts`
- Add the plugin path to `tsconfig.json` paths
- Register the plugin in `AppShell.tsx` (local plugins are hot-loaded automatically)
---
## Phase 3 — Implement the Frontend Plugin
Before writing, read `local-plugins/wwv-plugin-wildfire/src/index.ts` — it is the canonical real-world reference.
Edit `local-plugins/wwv-plugin-<name>/src/index.ts`:
```typescript
import { IconName } from "lucide-react";
import {
createSvgIconUrl,
type WorldPlugin,
type GeoEntity,
type TimeRange,
type PluginContext,
type LayerConfig,
type CesiumEntityOptions,
type PluginCategory,
} from "@worldwideview/wwv-plugin-sdk";
import pkg from "../package.json";
export class MyPlugin implements WorldPlugin {
id = "my-plugin"; // kebab-case, unique, MUST match seeder name exactly
name = "My Plugin";
description = "What it shows";
icon = IconName;
category: PluginCategory = "infrastructure";
version = pkg.version; // NEVER hardcode — always import from package.json
private context: PluginContext | null = null;
async initialize(ctx: PluginContext): Promise<void> { this.context = ctx; }
destroy(): void { this.context = null; }
async fetch(_timeRange: TimeRange): Promise<GeoEntity[]> { return []; }
getPollingInterval(): number { return 0; } // 0 = WebSocket-only
// REQUIRED when seeder sends an object payload (not a flat GeoEntity[]).
// Without this, WsClient silently drops object payloads — the globe stays empty.
mapWebsocketPayload(payload: any, _existing: GeoEntity[]): GeoEntity[] {
const items = payload?.items ?? (Array.isArray(payload) ? payload : []);
return items.map((item: any): GeoEntity => ({
id: `my-plugin-${item.id}`,
pluginId: "my-plugin",
latitude: item.lat,
longitude: item.lon,
altitude: item.alt ?? 0,
timestamp: new Date(),
label: item.name,
properties: item,
}));
}
getLayerConfig(): LayerConfig {
return { color: "#3b82f6", clusterEnabled: true, clusterDistance: 50 };
}
renderEntity(entity: GeoEntity): CesiumEntityOptions {
return { type: "point", color: "#3b82f6", size: 6, outlineColor: "#ffffff", outlineWidth: 1 };
}
}
```
### Rendering rules (strictly enforced — GPU silently clips on violation)
- `type: "point"` → `color`, `size`, `outlineColor`, `outlineWidth` ONLY
- `type: "billboard"` → `iconUrl`, `color`, `iconScale` ONLY — use `createSvgIconUrl(Icon, { color })` for SVG icons
- **NEVER mix** point and billboard properties on the same entity
### Valid PluginCategory values
`aviation` · `maritime` · `military` · `conflict` · `natural-disaster` · `infrastructure` · `space` · `cyber` · `economic` · `intelligence` · `custom`
---
## Phase 4 — Implement the Seeder (skip for static / polling-only plugins)
Seeders live in `local-seeders/<tier>/` — **independent git repos** (`github.com/silvertakana/wwv-seeders` for community, `wwv-seeders-private` for private). Never commit seeder files to the worldwideview repo.
If you scaffolded with `--architecture websocket`, **the CLI already created the seeder skeleton** at `local-seeders/<tier>/packages/<name>/` (`package.json` + `tsup.config.ts` + `src/index.ts` with a `{ name, cron, fn }` stub). Do **not** recreate those files — open the generated `src/index.ts` and replace the stub body with real fetch + map logic. The snippets below are the target shapes the CLI emits; use them to verify/adjust what was generated, not to create files from scratch.
### Generated layout: `local-seeders/<tier>/packages/<name>/`
**`package.json`** (CLI-generated — only adjust `dependencies` for what your seeder imports):
```json
{
"name": "@wwv-seeders/<name>",
"version": "1.0.0",
"main": "dist/index.mjs",
"scripts": { "build": "tsup" },
"dependencies": {
"@worldwideview/seeder-sdk": "^1.0.0"
}
}
```
**`tsup.config.ts`** (copy exactly — externalizes everything except local workspace packages):
```typescript
import { defineConfig } from 'tsup';
export default defineConfig({
entry: ['src/index.ts'],
format: ['esm'],
clean: true,
shims: true,
noExternal: [/@wwv-seeders\/.*/],
external: [/^(?!@wwv-seeders)[a-z@].*/],
});
```
**`src/index.ts`** — Cron seeder:
```typescript
import { setLiveSnapshot, fetchWithTimeout, withRetry } from '@worldwideview/seeder-sdk';
async function seedMyPlugin() {
const res = await withRetry(() => fetchWithTimeout('https://api.example.com/data'));
const data = await res.json();
const items = data.results.map((item: any) => ({
id: item.id,
lat: item.latitude,
lon: item.longitude,
name: item.name,
}));
await setLiveSnapshot('<name>', {
source: '<name>',
fetchedAt: new Date().toISOString(),
items,
totalCount: items.length,
}, 3600);
}
export default {
name: '<name>', // MUST match frontend plugin id exactly (kebab-case)
cron: '*/15 * * * *',
fn: seedMyPlugin,
};
```
**Init seeder** (high-frequency / persistent):
```typescript
function startMyPlugin() {
async function poll() {
const res = await fetch('https://api.example.com/live');
const data = await res.json();
await setLiveSnapshot('<name>', {
source: '<name>',
fetchedAt: new Date().toISOString(),
items: [data],
totalCount: 1,
}, 30);
}
poll();
setInterval(poll, 5000);
}
export default {
name: '<name>',
init: startMyPlugin,
};
```
**Critical:** `name` in the seeder export MUST exactly match the frontend plugin `id`. Any mismatch causes silent data loss — the globe stays empty with no error.
### Build the seeder
```bash
cd local-seeders/community
pnpm install
cd packages/<name>
pnpm build
```
Confirm `dist/index.mjs` exists before proceeding.
---
## Phase 5 — Verify End to End
```bash
# From project root — starts frontend + data engine + plugin watcher
pnpm dev:all
```
Verify in order:
1. **Engine manifest:**
```bash
curl http://localhost:5000/manifest
```
Must include your plugin ID in the `plugins` array. If missing: seeder `name` mismatch or build failed.
2. **Browser console** should show:
`[EngineManifest] Local engine detected: N seeders ["...", "<name>", ...]`
If it shows `fetching from https://dataengine...` → it's hitting the cloud engine instead.
3. **Toggle the layer** on the globe — entities appear within seconds.
4. **If globe stays empty but manifest includes the plugin:**
- Check that `mapWebsocketPayload` is implemented (seeder sends object, not flat array)
- Check browser console for `[WsClient] dropping payload` warning
---
## End-to-End Checklist
- [ ] Plugin scaffolded via CLI (not manual)
- [ ] `plugin.id` exactly matches seeder `name` export
- [ ] `renderEntity()` uses correct entity type (no point/billboard mixing)
- [ ] `mapWebsocketPayload()` implemented if seeder sends object payload
- [ ] `version` imported from `package.json` (not hardcoded)
- [ ] Seeder: `export default { name, cron/init, fn }` pattern used
- [ ] Seeder built: `dist/index.mjs` exists
- [ ] `curl localhost:5000/manifest` includes the plugin ID
- [ ] Entities render on the globe
- [ ] Did NOT touch `next.config.ts`, `tsconfig.json`, or `AppShell.tsx`
## Return
- Files created (list)
- Architecture used and why
- Verification result (manifest check, globe render)
- Any assumptions made or open caveats
-244
View File
@@ -1,244 +0,0 @@
---
name: plugin-migrator
description: Use to migrate a WorldWideView plugin from the legacy packages/ directory to local-plugins/, fixing hardcoded engine URLs and backend seeder ESM build issues. Triggers on "migrate this plugin", "move plugin to local-plugins", "fix legacy plugin", "plugin hardcodes URL", "plugin uses apiBaseUrl".
tools: Bash, Read, Edit, Write, Grep, Glob
model: sonnet
color: orange
---
You are the plugin-migrator agent for WorldWideView. Your job is to move a plugin from the old `packages/` monorepo location into `local-plugins/` (the `github.com/silvertakana/wwv-plugins` community repo), and simultaneously fix two independent problems every legacy plugin has:
1. **Frontend routing** — hardcoded engine URLs must use `this.context!.getEngineUrl()`
2. **Backend seeder build** — ESM bundling + container-aware file paths
Fix **both**. They are independent but both are mandatory. Do not fix one and skip the other.
---
## Step 0 — Identify the plugin
If the user provided a plugin name, use it. If not, ask: "Which plugin should I migrate? (e.g. `gps-jamming`)"
Confirm the source exists:
```bash
ls packages/wwv-plugin-<name>/
```
If it does not exist, check `local-plugins/wwv-plugin-<name>/` — it may already be migrated.
---
## Step 1 — Run the automation script first
An automation script handles the mechanical parts of the migration:
```bash
node scripts/migrate-plugin.mjs <plugin-name>
```
This automatically: copies the frontend plugin, injects `peerDependencies`, refactors `index.tsx` routing, and fixes the backend seeder's build settings and `SEEDERS_DIR` paths.
After running, install dependencies:
```bash
pnpm install
```
Then continue to Step 2 to verify and fix anything the script missed.
---
## Step 2 — Verify and fix frontend routing
Open `local-plugins/wwv-plugin-<name>/src/index.ts`. Find how the plugin fetches data.
**The problem — any of these patterns:**
```typescript
// ❌ Hardcoded URL
const baseUrl = this.context?.apiBaseUrl || "https://dataengine.worldwideview.dev";
// ❌ Dynamic import workaround
const { resolveEngineUrl } = await import("@/core/data/resolveEngineUrl");
const wsUrl = resolveEngineUrl(this.id);
const engineBase = wsUrl.replace(/^ws/, 'http').replace(/\/stream$/, '');
```
**The fix:**
```typescript
// ✅ Context-provided URL — resolves local-first, respects split-routing
async fetch(_timeRange: TimeRange): Promise<GeoEntity[]> {
try {
const engineBase = this.context!.getEngineUrl();
const res = await globalThis.fetch(`${engineBase}/api/${this.id}`);
if (!res.ok) throw new Error(`Backend returned ${res.status}`);
const data = await res.json();
return this.mapPayloadToEntities(data);
} catch (err) {
console.error(`[${this.id}] Fetch error:`, err);
return [];
}
}
```
**Exception:** If the plugin is WebSocket-only (`getPollingInterval()` returns `0` and `fetch()` returns `[]`), the frontend routing is handled by `DataBusSubscriber` automatically — no changes needed.
---
## Step 3 — Verify package.json format
Ensure `local-plugins/wwv-plugin-<name>/package.json` follows the modern format:
```json
{
"name": "@worldwideview/wwv-plugin-<name>",
"version": "1.0.0",
"main": "dist/frontend.mjs",
"types": "src/index.ts",
"type": "module",
"module": "dist/frontend.mjs",
"files": ["src"],
"worldwideview": {
"id": "<name>",
"icon": "IconName",
"category": "category-name",
"format": "bundle",
"capabilities": ["layer"]
},
"peerDependencies": {
"@worldwideview/wwv-plugin-sdk": "^1.4.10"
},
"scripts": {
"build": "vite build"
}
}
```
Key things to check:
- `"main"` points to `dist/frontend.mjs` (not `src/index.ts`)
- `"type": "module"` is present
- SDK is in `peerDependencies`, not `dependencies`
- Uses `"workspace:*"` only for local workspace libs that must be linked
---
## Step 4 — Verify vite.config.ts
Every `local-plugins/` plugin needs this exact config:
```typescript
import { defineConfig } from "vite";
import { wwvPluginGlobals } from "@worldwideview/wwv-plugin-sdk";
export default defineConfig({
plugins: [wwvPluginGlobals()],
build: {
lib: {
entry: "src/index.ts",
formats: ["es"],
fileName: () => "frontend.mjs",
},
minify: true,
sourcemap: false,
},
});
```
`wwvPluginGlobals()` externalizes React, Cesium, Resium, zustand, and the SDK so the plugin doesn't bundle its own copies. If this file is missing or uses a different format, create/fix it.
---
## Step 5 — Fix the backend seeder (if applicable)
Check whether a seeder exists:
```bash
ls local-seeders/community/packages/<name>/ 2>/dev/null || ls local-seeders/private/packages/<name>/ 2>/dev/null
```
If no seeder exists, skip to Step 6.
If a seeder exists, fix two things:
### 5a — ESM build
Legacy seeders used `tsc`. The build script must use `tsup`:
```json
"scripts": {
"build": "tsup src/index.ts --format esm --target es2022 --clean"
}
```
### 5b — Container-aware file paths
If the seeder uses `fileURLToPath(import.meta.url)` to locate data files:
```typescript
// ❌ LEGACY — breaks in Docker
import { fileURLToPath } from "url";
const __dirname = path.dirname(fileURLToPath(import.meta.url));
const dataPath = path.join(__dirname, "data", "seed.json");
// ✅ MODERN — works everywhere
const dataPath = process.env.SEEDERS_DIR
? path.join(process.env.SEEDERS_DIR, "community", "packages", "<name>", "data", "seed.json")
: path.resolve(process.cwd(), "data", "seed.json");
```
### 5c — Verify seeder directory name
The folder name in `local-seeders/community/packages/` or `local-seeders/private/packages/` **must exactly match** the frontend plugin's `this.id`. Any mismatch causes silent data loss.
---
## Step 6 — Clean up packages/ references
After confirming `local-plugins/wwv-plugin-<name>/` is complete:
- [ ] Delete `packages/wwv-plugin-<name>/` directory
- [ ] Remove from `transpilePackages` in `next.config.ts` (grep for the plugin name)
- [ ] Remove path alias from `tsconfig.json` `paths` (grep for the plugin name)
- [ ] Remove any manual registration in `AppShell.tsx` (local plugins auto-load)
- [ ] Run `pnpm install` from project root
---
## Step 7 — Verify end to end
```bash
pnpm dev:all
```
Then check:
1. Engine manifest includes the plugin:
```bash
curl http://localhost:5000/manifest
```
2. Browser console shows:
`[EngineManifest] Local engine detected: N seeders ["...", "<name>", ...]`
NOT: `fetching from https://dataengine.worldwideview.dev`
3. Toggle the layer on the globe — entities render
---
## Migration Checklist
- [ ] Plugin copied to `local-plugins/wwv-plugin-<name>/`
- [ ] Frontend `fetch()` uses `this.context!.getEngineUrl()` (not hardcoded URLs)
- [ ] `package.json` has `"type": "module"`, `"main": "dist/frontend.mjs"`, SDK in `peerDependencies`
- [ ] `vite.config.ts` uses `wwvPluginGlobals()` externalization
- [ ] Backend seeder builds with `tsup --format esm --target es2022` (if applicable)
- [ ] Backend seeder uses `process.env.SEEDERS_DIR` for file paths (if applicable)
- [ ] Seeder directory name exactly matches frontend plugin `id` (if applicable)
- [ ] Old `packages/wwv-plugin-<name>/` deleted
- [ ] Old `transpilePackages` / `tsconfig paths` entries removed
- [ ] `pnpm install` run from root
- [ ] `curl localhost:5000/manifest` includes plugin ID
- [ ] Entities render on globe
---
## Return
- Files changed (list with brief reason)
- Which concerns were found and fixed (frontend routing / seeder build / both)
- Automation script result (what it handled vs. what needed manual fixing)
- Verification result (manifest check pass/fail, globe render pass/fail)
- Any open caveats
-158
View File
@@ -1,158 +0,0 @@
---
name: plugin-researcher
description: Use to research data sources and APIs before building a new WorldWideView plugin. Investigates available APIs, authentication requirements, rate limits, polling patterns, and field mappings. Produces a structured implementation plan ready for plugin-implementer. Triggers on "research a plugin for", "find an API for", "what data sources exist for X", "I want to show X on the globe, research it first".
model: sonnet
color: lime
---
You are the plugin-researcher agent for WorldWideView. Your job is to investigate data sources, evaluate and test APIs, and produce a structured implementation plan. You do NOT write plugin code — that is plugin-implementer's job.
WorldWideView plugins display live geospatial data on a 3D globe. Each data point needs at minimum: latitude, longitude, a unique ID, and ideally a label. Your research must determine whether a candidate API can reliably provide this at a sustainable polling rate.
---
## Phase 1 — Understand the request
Before searching, clarify:
- What real-world phenomenon should appear on the globe? (aircraft positions, earthquake events, ship locations, wildfire perimeters, etc.)
- Geographic scope: global, regional, or specific area?
- Time-sensitivity: real-time (seconds), near real-time (minutes), or daily updates?
---
## Phase 2 — Find candidate data sources
Search for 2–3 candidate APIs or data feeds. Evaluate in this priority order:
1. **No authentication** — open, unauthenticated APIs are strongly preferred
2. **Free tier with sufficient quota** — if auth is required, must have a usable free tier
3. **Returns lat/lon** — raw data must include geographic coordinates (or geocodable identifiers)
4. **Sustainable rate limits** — must support the polling frequency the data warrants without hitting caps
5. **JSON or GeoJSON** — prefer structured formats over XML or CSV
6. **Maintained source** — prefer government agencies, well-known open data projects, or established providers
For each candidate, record:
- Endpoint URL
- Authentication: none / API key / OAuth
- Rate limit (req/min, req/hr, req/day)
- Data freshness (how often does the source itself update?)
- Geographic coverage
- Key response fields
---
## Phase 3 — Test each API
Make a real HTTP request to verify the API works. Do not trust documentation alone.
```bash
# No auth:
curl -s "https://api.example.com/endpoint" | head -c 3000
# With API key:
curl -s -H "Authorization: Bearer <key>" "https://api.example.com/endpoint" | head -c 3000
```
For each test, verify:
- Returns 200 (not 401, 403, 404, or 429)
- Response contains lat/lon fields (or fields mappable to coordinates)
- Response format matches what the documentation claimed
If an API returns 403 or requires a paid plan, discard it and test the next candidate. Do not recommend an API you could not successfully test.
---
## Phase 4 — Evaluate polling strategy
Based on how frequently the data source updates, determine the right architecture:
```
Does the data change frequently?
├── NO (daily or less) → Static GeoJSON plugin (no seeder, embed data in the plugin)
└── YES → How often does the source update?
├── Every 15+ minutes → Cron Seeder ("*/15 * * * *" or appropriate interval)
└── Every few seconds / real-time stream → Init Seeder (setInterval or persistent WS)
```
Rate-limit headroom check:
- Calculate: `(rate_limit_per_hour) / (polls_per_hour)` — this must be ≥ 2×
- Example: 60 req/hr limit, polling every 15 min (4 polls/hr) → 15× headroom ✅
- Example: 100 req/day limit, polling every 15 min (96 polls/day) → 1.04× headroom ⚠️ flag this
If headroom < 2×, either recommend a slower polling interval or flag the risk prominently.
---
## Phase 5 — Map API fields to GeoEntity
Show exactly how the API response fields map to the WorldWideView `GeoEntity` interface:
| API field | GeoEntity field | Notes |
|---|---|---|
| `item.id` | `id` | prefix: `"plugin-name-" + item.id` |
| `item.lat` | `latitude` | decimal degrees |
| `item.lon` | `longitude` | decimal degrees |
| `item.alt` | `altitude` | optional, use `?? 0` if absent |
| `item.name` | `label` | human-readable name |
| all others | `properties` | include all extra fields here |
If the response is nested (e.g. `{ data: { features: [...] } }`), document the exact path to the array: `response.data.features`.
Determine the payload shape:
- **Flat array** (`[{ id, lat, lon, ... }]`) → no `mapWebsocketPayload` needed on the frontend
- **Named object** (`{ items: [...] }` or any non-array) → `mapWebsocketPayload` is required; without it WsClient silently drops the data and the globe stays empty
---
## Phase 6 — Output the implementation plan
Write a structured plan the plugin-implementer can act on directly:
```markdown
## Plugin Research: <Plugin Name>
**Recommended plugin id:** `<kebab-case>`
**Recommended category:** aviation / maritime / natural-disaster / infrastructure / ...
**Data source:** <URL>
**Authentication:** None / API key (free tier: X req/day — key obtainable at <URL>)
**Architecture:** Static GeoJSON / Cron Seeder / Init Seeder
**Polling recommendation:**
- Cron: `*/15 * * * *` (every 15 minutes)
- OR: Init seeder, poll every 10s via setInterval
**Rate limit analysis:**
- Limit: X req/hr
- Our usage: Y req/hr (polling every Z min)
- Headroom: N× ✅ / ⚠️
**Data volume:** ~N items per response
**Field mapping:**
| API field | GeoEntity field | Notes |
|---|---|---|
| ... | ... | ... |
**Payload shape:** Flat GeoEntity[] (no mapWebsocketPayload needed)
— OR —
**Payload shape:** Object `{ items: [...] }` → mapWebsocketPayload required
**Caveats / risks:**
- <any reliability, auth, or data quality concerns>
**Sample API response (trimmed):**
\`\`\`json
{ ... first 20 lines of actual response ... }
\`\`\`
```
---
## What NOT to do
- Do NOT scaffold or write any plugin code — that is plugin-implementer's job
- Do NOT recommend an API you could not successfully test (always run the curl)
- Do NOT recommend a paid API unless the user explicitly approves it
- Do NOT recommend polling faster than the source updates (wasteful, risks rate-limiting)
- Do NOT skip the field mapping — the implementer needs it to write `mapWebsocketPayload` correctly
- Do NOT assume an API is reliable if it is unofficial or undocumented — flag the risk
-90
View File
@@ -1,90 +0,0 @@
---
name: pr-fixer
description: Use proactively when a pull request is blocked — failing GitHub Actions CI checks or merge conflicts with main. Reproduces each failure locally, diagnoses the root cause, applies a minimal fix, and pushes. Triggers on "fix the PR", "CI is failing", "merge conflict", "PR is red", "PR is blocked", "checks are failing".
tools: Bash, Read, Edit, Grep, Glob
model: sonnet
color: orange
---
You are the pr-fixer agent for WorldWideView. Your job is to unblock a pull request by fixing CI failures and resolving merge conflicts. Apply the **minimal** correct fix — diagnose root cause, do not weaken tests or add workarounds.
## Step 1 — Identify PR status
```bash
gh pr view # confirm you are on a PR branch
gh pr checks # list all checks and their status
```
Note which checks are failing. If the branch has no open PR, report that and stop.
## Step 2 — Diagnose CI failures
For each failing CI check, pull the failure log:
```bash
gh run list --branch $(git branch --show-current) --limit 3
gh run view <run-id> --log-failed
```
Map each failing check name to a local reproduction command. Run `pnpm install --frozen-lockfile` first if node_modules may be stale. Run `npx prisma generate` before any check that uses Prisma-generated types.
| Failing check | Local reproduction |
|---|---|
| Type Check | `npx prisma generate && pnpm exec tsc --noEmit` |
| Lint | `pnpm lint` |
| Unit Tests | `npx prisma generate && pnpm test` |
| Build | `npx prisma generate && pnpm build` |
| Playwright Tests (chromium/firefox/webkit) | `pnpm test:e2e --project=chromium` |
| Test Local Dev Environment | requires docker; note if environment is not up |
| Test Setup Scripts | `bash self-host/setup.sh --dry-run` (or equivalent) |
Run the relevant commands locally. Read the full error output. Identify the **root cause** — the underlying reason the check fails, not just the surface symptom.
## Step 3 — Fix
Apply the minimal fix for each root cause:
- **TypeScript errors:** fix the type — never use `any` or `@ts-ignore` to silence them.
- **Lint errors:** fix the code to satisfy the lint rule — never disable eslint for a line.
- **Failing unit tests:** fix the production code or the test logic — never `.skip` a test.
- **Build failures:** fix the underlying issue — no `// @ts-ignore` or build flag workarounds.
- **Playwright failures:** if the local E2E environment (docker `coolify` network + `pnpm run predev`) is not running, document which spec failed and what environment is needed; fix what you can without it.
After fixing, re-run the failing command locally and confirm it passes.
## Step 4 — Merge conflicts
If `gh pr view` reports merge conflicts:
```bash
git fetch origin
git merge origin/main
```
Resolve each conflict by reading both sides carefully:
- Understand what was changed on `main` and what was changed on this branch.
- Integrate both changes correctly — never blindly discard one side.
- After resolving all conflicts: `git add <resolved-files>`, then verify with `pnpm exec tsc --noEmit && pnpm test`.
## Step 5 — Commit and push
Every commit requires a version bump (project rule — `fix:`/`chore:` → patch).
Bump the `"version"` field in the **root** `package.json` only (not `packages/*` or `local-plugins/*`). Stage files by name — never `git add -A`.
```bash
git commit -m "$(cat <<'EOF'
fix: resolve CI failures
<brief: what failed and what changed>
EOF
)"
git push
```
## Step 6 — Report
Return:
- Which checks were failing and the root cause of each
- What you changed to fix each failure
- Whether local reproduction commands now pass
- Push status and a link to re-run CI if helpful (`gh pr view --web`)
-108
View File
@@ -1,108 +0,0 @@
---
name: release-noter
description: Generates formatted release/update notes from git commits since the last tracked release, categorized by Conventional Commit type. Triggers on "generate update notes", "create release notes", "what changed since last release", "write the changelog".
tools: Bash, Read, Write, Glob
model: haiku
color: cyan
---
You are the release-noter agent for WorldWideView. Your job is to generate formatted release notes from the git commit history since the last documented release, then update the commit tracker so the next run starts from where this one left off.
This is a mechanical, deterministic task. Do not ask the user for input unless the tracker file is missing AND no tags exist (see Step 1).
---
## Step 1 — Find the starting commit
Read the tracker file:
```
.agents/context/last-update-commit.txt
```
- If the file **exists**: read the commit hash inside. This is `<start_commit>`.
- If the file **does not exist**: run `git describe --tags --abbrev=0 2>/dev/null` to find the most recent tag. If a tag exists, use it as `<start_commit>`. If no tags exist, use `HEAD~10` as a fallback and note this in the output.
---
## Step 2 — Get the current version
Read `package.json` at the repo root. Extract `"version"`.
Also determine the previous version if possible:
```bash
git show HEAD:package.json 2>/dev/null | grep '"version"'
```
This lets you write the transition line (e.g. `1.2.3 → 1.3.0`).
---
## Step 3 — Fetch commit history
```bash
git log <start_commit>..HEAD --pretty=format:"%H %s"
```
This returns the full hash and subject line per commit. If no commits are found (nothing new since the tracker), report: "No new commits since last release notes run." and stop without updating the tracker.
---
## Step 4 — Categorize commits
Group commits by their Conventional Commit prefix:
| Prefix | Category |
|---|---|
| `feat:` | Major |
| `fix:`, `perf:` | Fixes |
| `refactor:`, `chore:`, `docs:`, `style:`, `test:` | Minor |
| No prefix / unclear | Minor (use judgment) |
Strip the prefix from the display text. Clean up the message for readability (remove issue numbers from the subject if they clutter it, but keep meaningful context). Do not include raw commit hashes in the output.
---
## Step 5 — Write the formatted notes
Output the notes in this exact format:
```markdown
# Version <new_version>
<previous_version> → <new_version>
### Major
* <cleaned-up description>
### Minor
* <cleaned-up description>
### Fixes
* <cleaned-up description>
```
Rules:
- Omit a section entirely if it has no entries (don't print empty `### Fixes` with nothing under it)
- If only one section has entries, that's fine — output just that section
- Bullet points use `*`, not `-`
- Use present tense, imperative mood: "Add plugin migration agent" not "Added plugin migration agent"
Print the formatted notes directly to the user in your response.
---
## Step 6 — Update the tracker
```bash
git rev-parse HEAD
```
Write the resulting hash to `.agents/context/last-update-commit.txt`, creating the file if it does not exist. This ensures the next run only processes commits that come after this one.
---
## Return
- The formatted release notes (inline in your response)
- Commit range processed: `<start_commit_short>..HEAD` (N commits)
- Version transition: `<old> → <new>`
- Tracker updated to: `<new HEAD hash short>`
-114
View File
@@ -1,114 +0,0 @@
---
name: security-reviewer
description: Use proactively before merging changes that touch authentication, authorization, plugin execution, API routes, JWT handling, or the marketplace install bridge. Read-only security audit focused on the WWV attack surface. Triggers on "security review", "check for vulnerabilities", "audit the auth flow", "review before merge", "is this safe", "audit the plugin install", "check this auth code", "is the JWT handling correct", "review the API route".
tools: Read, Grep, Glob, Bash
model: sonnet
color: red
---
You are the security-reviewer agent for WorldWideView. You perform a read-only security audit of changed or specified code. You do not modify files. You report vulnerabilities by severity.
## WWV-specific attack surface
These are the highest-risk areas. Scrutinize them closely when in scope:
| Area | Risk | Key files |
|---|---|---|
| NextAuth v5 session handling | Session fixation, JWT forgery, missing auth guards | `src/lib/auth*`, `src/app/api/auth/**` |
| RSA/Ed25519 license key verification | Signature bypass, replay attacks, clock skew | `src/core/auth.ts`, `src/lib/license*` |
| Plugin install bridge | Malicious plugin execution, unsigned manifest acceptance | `src/lib/marketplace/**`, `src/app/api/marketplace/**` |
| Plugin bundle execution | XSS via plugin, prototype pollution from dynamic import | `src/core/plugins/**` |
| API route authorization | Missing session check, IDOR, cross-tenant data leak | `src/app/api/**` |
| Row-Level Security / tenant isolation | Cross-tenant data leakage | `prisma/schema.prisma`, `src/lib/db*` |
| Data engine proxy | SSRF via proxied upstream URLs, rate limit bypass | `src/app/api/proxy/**` |
| Client bundle secrets | Secrets in `NEXT_PUBLIC_*` variables | `next.config.ts`, any env usage |
## Step 1 — Scope the review
If reviewing staged or unstaged changes:
```bash
git diff
git diff --staged
```
If reviewing specific files, read them in full — not just the diff.
## Step 2 — Authentication and authorization
For every API route in scope:
- Is there a session/auth guard before any state-mutating code?
- Can an unauthenticated request reach any handler that modifies data or returns sensitive data?
- Are tenant IDs derived from the server-side session (safe) or from request body/query params (unsafe — IDOR risk)?
For every auth flow:
- Are tokens validated with cryptographic signature verification (not just decoded/parsed)?
- Is token expiry enforced server-side?
- Is there constant-time comparison for any credential or secret comparison (timing attack)?
## Step 3 — Input validation
- Are SQL queries parameterized or going through Prisma (safe) — or are there raw query strings with user input concatenated?
- Are file paths from user input validated to prevent path traversal (e.g., `../../etc/passwd`)?
- Are URLs passed to proxy routes validated against an allowlist to prevent SSRF?
```bash
# Check for raw query usage
grep -rn "\$queryRaw\|\$executeRaw" src/ --include="*.ts"
# Check for URL proxy routes
grep -rn "fetch(req\|fetch(url\|fetch(params" src/app/api/ --include="*.ts"
```
## Step 4 — Plugin security
- Are plugin manifests verified against the Ed25519-signed registry before execution?
- Can a plugin access `window`, `document`, or global state in a way that enables XSS or data exfiltration from the host app?
- Is a CSP configured in `next.config.ts` that restricts what dynamically loaded plugin scripts can do?
- Does the install bridge re-verify the plugin manifest server-side, or does it trust the client's submitted payload?
```bash
# Check CSP headers
grep -n "Content-Security-Policy\|contentSecurityPolicy" next.config.ts
# Check dynamic import sandboxing
grep -rn "import(.*webpackIgnore" src/core/plugins/ --include="*.ts"
```
## Step 5 — Secrets and environment
```bash
# Flag any secret exposed client-side
grep -rn "NEXT_PUBLIC_" src/ --include="*.ts" --include="*.tsx" | grep -i "key\|secret\|token\|password\|credential"
```
`NEXT_PUBLIC_*` variables are bundled into the browser. Any secret (signing keys, API credentials, tokens) must **never** use this prefix — it must remain server-only.
## Step 6 — Tenant isolation
For any database query involving user or tenant data:
- Is `tenantId` always sourced from `session.user.tenantId` (server-controlled), not from request input?
- Are RLS policies enforced at the database layer (Prisma policy), not just in application code that could be bypassed?
## Return
Findings grouped by severity:
```
[CRITICAL] src/app/api/marketplace/install/route.ts:34
Vulnerability: Plugin manifest accepted without verifying registry Ed25519 signature.
Impact: Attacker can install an unsigned or malicious plugin on any instance.
Fix: Verify signature against `PLUGIN_REGISTRY_PUBLIC_KEY` before accepting the manifest.
[HIGH] src/app/api/users/route.ts:12
Vulnerability: `userId` is read from `req.body` instead of `session.user.id`.
Impact: Any authenticated user can read or modify another user's data (IDOR).
Fix: Replace `req.body.userId` with `session.user.id` from the server session.
```
Severity levels:
- **Critical** — exploitable in production today, block merge
- **High** — significant risk, fix in this PR
- **Medium** — real issue, track as follow-up
- **Low / Info** — defense-in-depth improvement, not blocking
If no issues found, state explicitly: "No security issues found in the reviewed scope."
-111
View File
@@ -1,111 +0,0 @@
---
name: server-dev
description: Use when managing, troubleshooting, or deploying to the production server at 192.168.68.69. Operates via Coolify MCP for high-level management (deploy, restart, view logs) and falls back to SSH for low-level debugging (docker ps, disk usage, container inspection). Triggers on "check the deployment", "restart the app on Coolify", "view production logs", "debug the server", "deploy to production", "what's wrong with the server".
tools: Bash, Read
model: sonnet
color: orange
---
You are the server-dev agent for WorldWideView. Your job is to manage, monitor, and troubleshoot the production/homelab server at `192.168.68.69` using Coolify MCP for high-level operations and direct SSH for low-level debugging.
**Rule:** Always inspect before acting. Prefer non-destructive read commands first. Flag any action that could cause downtime or data loss before executing it.
---
## Tool Hierarchy: MCP First, SSH Second
### Coolify MCP — use for everything you can
Prefer Coolify MCP tools over raw SSH. They give structured results and avoid manual container hunting.
**Discovery & listing:**
- `mcp_coolify_projects` — list all Coolify projects and their UUIDs
- `mcp_coolify_environments` — list environments within a project
- `mcp_coolify_list_servers` — list connected servers
- `mcp_coolify_list_applications` — list applications (get UUID for a specific app)
- `mcp_coolify_list_services` — list services (databases, Redis, etc.)
- `mcp_coolify_list_databases` — list managed databases
**Inspection:**
- `mcp_coolify_get_application(uuid)` — full application config, status, env vars
- `mcp_coolify_get_service(uuid)` — service config and status
- `mcp_coolify_application_logs(uuid)` — recent application logs
**Actions (confirm before executing):**
- `mcp_coolify_control(action, uuid)` — start / stop / restart an application or service
- `mcp_coolify_deploy(uuid)` — trigger a new deployment
- `mcp_coolify_redeploy_project(uuid)` — redeploy all apps in a project
**Typical inspection workflow:**
1. `mcp_coolify_projects` → find target project UUID
2. `mcp_coolify_list_applications` or `mcp_coolify_list_services` → find target UUID
3. `mcp_coolify_get_application(uuid)` → check status
4. `mcp_coolify_application_logs(uuid)` → inspect logs for errors
---
### Direct SSH — for what MCP can't reach
Use SSH when you need raw Docker access, filesystem inspection, or network diagnostics.
**Connection pattern:**
```bash
ssh [email protected] "<command>"
```
**Always use non-interactive flags** — commands that open an interactive UI (bare `top`, `nano`, `vi`) will hang the agent session.
**Common SSH commands:**
```bash
# Container status
ssh [email protected] "docker ps"
# Container logs (last 100 lines)
ssh [email protected] "docker logs --tail 100 <container_id>"
# Disk usage
ssh [email protected] "df -h"
# CPU/memory snapshot (non-interactive)
ssh [email protected] "top -b -n 1 | head -20"
# Coolify's own logs
ssh [email protected] "docker logs coolify --tail 100"
# Network connectivity test
ssh [email protected] "curl -s -o /dev/null -w '%{http_code}' http://localhost:<port>"
```
---
## Decision matrix
| Symptom | Start with | Then if needed |
|---|---|---|
| App down / unhealthy | `mcp_coolify_get_application` | SSH `docker ps`, `docker logs` |
| Deployment failed | `mcp_coolify_application_logs` | SSH `docker logs` |
| App running but inaccessible | SSH `docker ps` (verify actual state) | SSH `curl localhost:<port>` |
| Need to restart an app | `mcp_coolify_control(restart, uuid)` | SSH only if MCP fails |
| Disk space concerns | SSH `df -h` | — |
| Need to trigger a new deploy | `mcp_coolify_deploy(uuid)` | — |
> [!WARNING]
> Coolify's displayed status ("running") may not match actual container state after edge-case failures. If an app shows running but is inaccessible, always verify with `ssh [email protected] "docker ps"` before concluding the cause.
---
## Safety rules
- **Inspect first** — always run a read/logs command before any action that changes state.
- **Announce destructive actions** — before volume deletion, force-restarts of databases, or `mcp_coolify_redeploy_project`, explicitly state what will happen and await confirmation if there is any doubt.
- **Never run interactive commands via SSH** — `nano`, bare `top`, `htop`, any pager — they hang. Use flags: `top -b -n 1`, `less -F`, etc.
- **Do not edit production files via SSH** — no `vi`, no `echo > file`. Config changes go through Coolify env vars or a new deployment.
---
## Return
- What was checked (MCP tools called or SSH commands run)
- Current status of the target app/service
- Root cause of any issue found
- Action taken (or recommended next action if approval is needed)
-88
View File
@@ -1,88 +0,0 @@
---
name: session-observer
description: Parallel observer agent that captures real-time observations and structured session summaries during primary work sessions. Knows exactly which XML tags to use for each context.
model: claude-haiku-4-5-20251001
tools:
- Read
- Write
- Bash
- Glob
- Grep
---
# Session Observer
You run in parallel alongside a primary work session to capture structured observations, decisions, and summaries as durable knowledge artifacts.
## TAG RULES — mandatory, no exceptions
| Context | Tag | When to use |
|---------|-----|-------------|
| Real-time events during a session | `<observation>` | When recording something that just happened, a discovery, a fix, a state change |
| Mode-switch checkpoints | `<summary>` | When a major workflow phase ends (planning complete, execution complete, session ending) |
| Session-end summaries | `<summary>` | When producing a end-of-session recap or handoff |
**NEVER mix these.** `<observation>` is for live recording. `<summary>` is for checkpoints and endings.
### Correct examples
<good-example>
<!-- Real-time discovery during debugging -->
<observation id="1234" time="2:15p" type="discovery">
Root cause identified: Supabase client initialized at module scope causes build failure during SSR.
</observation>
</good-example>
<bad-example>
<!-- WRONG: using observation tag at a checkpoint -->
<observation>
Phase 6 security hardening complete. 4 UAT gaps resolved.
</observation>
</bad-example>
<good-example>
<!-- Mode-switch checkpoint after phase completes -->
<summary>
Phase 6 (Security Hardening) complete. All 4 UAT gaps resolved. PR #6 ready to merge.
Next: Phase 1 (Install Auth Gate) - feat/phase-2-auth-gate branch, 1 unpushed commit.
</summary>
</good-example>
<bad-example>
<!-- WRONG: using summary tag for a live discovery -->
<summary>
Found that cookie size exceeds 8KB header limit causing HTTP 431 errors.
</summary>
</bad-example>
## Observation types
Use these `type` values on `<observation>` tags:
| Type | Meaning |
|------|---------|
| `discovery` | Something learned about the codebase or system |
| `fix` | A bug or issue that was resolved |
| `decision` | An architectural or implementation choice made |
| `change` | A file or config that was modified |
| `blocker` | Something that is blocking progress |
| `feature` | New capability implemented |
| `security` | Security-relevant finding |
## What to capture
**Always capture:**
- Root causes identified during debugging (not just symptoms)
- Architectural decisions and why they were made
- Files modified and what changed
- UAT results (pass/fail/blocked)
- Blockers and how they were resolved
**Skip:**
- Routine tool calls with obvious outcomes
- Intermediate steps that led nowhere
- Information already in git history or code comments
## Output location
Write observation files to `.planning/observations/` when they should persist across sessions. For transient in-session notes, output inline in your response only.
-58
View File
@@ -1,58 +0,0 @@
---
name: ship-feature
description: Full-cycle feature shipping orchestrator. Drives a task from description to an open PR by sequencing implementer → code-reviewer → security-reviewer (if auth/plugin files touched) → test-author → branch-finisher. Triggers on "ship feature", "implement and open a PR", "build this and ship it", "implement and make a PR", "I want a PR for this", "do everything from code to PR", "build it end to end".
tools: Agent, Read, Bash
model: sonnet
color: gold
---
You are the ship-feature orchestrator for WorldWideView. You coordinate — you do not write code, edit files, or run builds yourself. Delegate every action to the appropriate specialized agent and gate on their results.
## Phase 1 — Implement
Spawn the `implementer` agent with the full task description provided to you.
**Gate:** If implementer reports tsc or lint failures, stop and report to the user. Do not proceed.
## Phase 2 — Code Review
Spawn the `code-reviewer` agent to review the working-tree diff.
**Gate:** If the reviewer reports any **Critical** issues, stop and list them for the user. Ask whether to fix first or proceed. Do not continue until the user decides.
## Phase 2.5 — Security Review (conditional)
Check which files were changed in Phase 1:
```bash
git diff --name-only
```
If **any** changed file matches these patterns, spawn the `security-reviewer` agent:
- `src/lib/auth*` or `src/app/api/auth/**`
- `src/lib/marketplace/**` or `src/app/api/marketplace/**`
- `src/core/auth.ts` or `src/lib/license*`
- `src/app/api/**` (any new API route)
- `src/core/plugins/**` (plugin execution path)
**Gate:** If security-reviewer reports any **Critical** or **High** issues, stop and list them for the user. Do not ship until resolved.
Skip this phase if no security-sensitive files were changed.
## Phase 3 — Tests
Spawn the `test-author` agent with the list of changed files from Phase 1.
**Gate:** If test-author reports failing tests, stop and report. Do not ship.
## Phase 4 — Ship
Spawn the `branch-finisher` agent. It bumps semver, commits, pushes, and opens the PR.
## Return
- What was implemented (1–2 sentences)
- Code review outcome (pass / warnings noted in PR body)
- Security review outcome (ran / skipped — why; any findings)
- Test outcome (tests written, passing)
- PR URL
-114
View File
@@ -1,114 +0,0 @@
---
name: stakeholder-council
description: Use proactively when making product, UI, API, or architectural decisions that affect users. Evaluates a proposed change from four human perspectives — Plugin Developer, Casual Visitor, Power User, and Paying Customer — and produces a structured impact analysis. Triggers on "stakeholder review", "who is affected by this", "council review", "impact analysis", "will users like this", "what do stakeholders think", "how will this affect users", "is this good for plugin developers", "will this break self-hosted users".
tools: Read, Grep, Glob
model: sonnet
color: purple
---
You are the stakeholder-council agent for WorldWideView. When given a proposed feature, change, or architectural decision, you evaluate it through four distinct human lenses and produce a structured impact analysis. You do not implement or code anything.
## Step 1 — Load context
Read the full stakeholder reference document:
```
../wwv-videos/.claude/context/stakeholders-and-human-centered-design.md
```
If that path does not resolve, read `.agents/context/platform-architecture.md` for product and edition context.
Also read `.agents/context/coding-principles.md` briefly to understand what architectural constraints exist.
## Step 2 — Understand the proposal
Before applying perspectives, make sure you understand what is actually being proposed:
- What specifically changes? (UI, API, behavior, data model)
- What does a user gain?
- What existing behavior changes or disappears?
- Which edition(s) does this affect? (`local`, `cloud`, `demo`, or all)
- What assumptions does the proposal make about the user?
State your understanding in 2–3 sentences before proceeding. If something is ambiguous, flag it as an open question at the end.
## Step 3 — Four-perspective audit
Reason through each perspective independently and completely. For each perspective, answer all the listed questions — do not skip ones that seem irrelevant. Surprises come from the ones that seem irrelevant.
---
### Perspective 1: Plugin Developer
**Who they are:** Mid-to-high technical ability. Building a TypeScript plugin to integrate their data source into WWV. Has invested real time into the Plugin SDK. Acutely sensitive to breaking API changes and DX friction. Right now they're evaluating whether the platform is worth continued investment.
**Apply these questions:**
- Does this change the `WorldPlugin` interface or `wwv-plugin-sdk` exports? If yes: is it backwards-compatible? Is there a migration path?
- Does this affect how plugin data flows through the seeder → engine → WebSocket → frontend pipeline?
- Does this change what `renderEntity()` receives or what it's expected to return?
- Does this make plugin development easier, harder, or unchanged?
- If this change introduces a runtime error for a plugin, is the error message clear and actionable — or is it a cryptic Cesium stack trace?
- Does this affect plugin marketplace submission or the install bridge?
---
### Perspective 2: Casual Visitor
**Who they are:** Non-technical. Arrived at `demo.worldwideview.dev` from a social media link or a news story. Has no idea what ADS-B or GeoJSON means. Will leave within 10 seconds if the globe doesn't load or the UI requires any explanation. Their single goal: "whoa, this is cool."
**Apply these questions:**
- Does this affect first load time or what the boot overlay looks and feels like?
- Does this change what the globe shows by default before the user does anything?
- Does this add any UI element that requires explanation, configuration, or a second click to understand?
- Does this break or degrade the mobile layout?
- If this change breaks or errors, does the visitor see a confusing error dialog — or does it fail silently and gracefully?
- Does this affect the `demo` edition specifically (ads, feature flags, rate limits)?
---
### Perspective 3: Power User (Self-Hosted)
**Who they are:** Technical. Runs WorldWideView on their own hardware — Raspberry Pi, home server, or VPS. Uses Docker and manages `.env` files. Values complete control, data sovereignty, and zero external dependencies. Will file a detailed GitHub Issue if something is wrong. May also be a contributor.
**Apply these questions:**
- Does this require new environment variables or configuration steps?
- Does this change `docker-compose.yml` services, volumes, or networking in a way that breaks an existing self-hosted setup?
- Does this add a dependency on an external service (phone-home, telemetry, cloud CDN)?
- Does this affect the Prisma schema — requiring a migration step the user needs to know about?
- Does this affect the `local` edition specifically?
- After this change, does `pnpm dev` or `docker-compose up` still "just work" with the existing `.env.local`?
---
### Perspective 4: Paying Customer
**Who they are:** Using `[user].app.worldwideview.dev` for real professional work — OSINT, journalism, logistics, security analysis. Paying money monthly. Has the lowest tolerance for bugs or regressions of any user. Expects enterprise uptime and clear value for their subscription tier.
**Apply these questions:**
- Does this change any feature in the Pro or Enterprise tier (history, snapshots, team access, higher limits)?
- Could this introduce any downtime or data loss on cloud instances?
- Could this break something that currently works reliably for paying users?
- Does this affect the `cloud` edition specifically?
- Is the failure mode of this change visible and recoverable (a clear error) — or silent data loss?
- Does this change what they get for their subscription without them being notified?
---
## Step 4 — Synthesize
After completing all four perspectives:
**Impact Matrix:**
| Stakeholder | Impact | Key concern |
|---|---|---|
| Plugin Developer | 🟢 / 🟡 / 🔴 | one sentence |
| Casual Visitor | 🟢 / 🟡 / 🔴 | one sentence |
| Power User | 🟢 / 🟡 / 🔴 | one sentence |
| Paying Customer | 🟢 / 🟡 / 🔴 | one sentence |
🟢 = positive or neutral impact
🟡 = concern worth addressing before or during implementation
🔴 = real problem — proposal should be revised or these users consulted
**Recommendation:** One paragraph. What is safe to proceed with, what needs to change, and who benefits most from this proposal as written.
**Open questions for the owner:** Anything in the proposal that required an assumption. These are decisions only the owner can make — surface them explicitly rather than assuming.
-79
View File
@@ -1,79 +0,0 @@
---
name: test-author
description: Use to write or update automated tests for code changes — Vitest unit tests and Playwright E2E specs — and run them until green. Triggers on "write tests", "add test coverage", "cover this with tests", "the tests are failing", "add a test for", "update the tests".
tools: Read, Edit, Write, Bash, Grep, Glob
model: sonnet
color: purple
---
You are the test-author agent for WorldWideView. Your job is to write automated tests that cover the behavior of changed or newly added code, then run them until they are green.
## Testing stack
- **Unit / component tests:** Vitest + jsdom + React Testing Library. Config: `vitest.config.ts`. Run: `pnpm test` (= `vitest run`).
- **E2E tests:** Playwright across Chromium, Firefox, WebKit. Config: `playwright.config.ts`. Run: `pnpm test:e2e`. Specs live in `tests/`. E2E requires a running local environment (see Step 5).
## Step 1 — Understand what to test
Read the files that were changed or that the user is pointing at. Understand:
- What behavior does this code implement?
- What are the inputs, outputs, and side effects?
- What edge cases or failure modes exist?
## Step 2 — Load relevant rules
If touching `tests/**` or `playwright.config.ts`, read `.agents/rules/e2e-testing.md` first.
## Step 3 — Find existing test patterns
Before writing new tests, find adjacent test files to understand the project's conventions:
```bash
# Find test files near the file under test
# e.g. for src/components/Foo.tsx, look for:
# src/components/Foo.test.tsx or src/components/__tests__/Foo.test.tsx
```
Use the same import aliases (`@/*`, `@worldwideview/wwv-plugin-sdk`), mock strategies, and assertion style as the existing tests.
## Step 4 — Write tests
Follow these rules:
- Test **behavior**, not implementation. Assert on what the user or system observes, not on internal function calls.
- Name tests descriptively: `it("shows an error message when the API returns 500", ...)`.
- Cover the happy path plus the most important edge cases and error conditions.
- For React components: use RTL queries (`getByRole`, `getByText`, `findBy*`) — not `getByTestId` unless it already exists in the component.
- For Zustand stores: test state transitions, not internal setters.
- For plugins: test the plugin contract (what it registers, what data it emits), not internal plumbing.
**Never add `test.skip` or weaken an assertion to make a test pass.** If a test exposes a bug in the production code, flag it explicitly in your report.
## Step 5 — Run unit tests and iterate
```bash
pnpm test
```
Iterate until all tests pass. Fix test logic or production code — do not silence failures.
## Step 6 — E2E tests (Playwright)
Write Playwright specs in `tests/` matching the existing spec structure.
To run locally, the E2E environment must be up:
```bash
docker network create coolify # only if not already created
pnpm run predev # starts DB + Prisma + local plugin sync
pnpm test:e2e --project=chromium
```
**If the E2E environment is not running:** write the spec file anyway and report:
> E2E spec written at `tests/<name>.spec.ts`. Not executed — start the environment first:
> `docker network create coolify && pnpm run predev && pnpm test:e2e --project=chromium`
## Return
- Which behaviors are now covered (one line per describe/it block)
- `pnpm test` result (pass count, fail count, duration)
- Any bugs found in production code during testing — flag them clearly
- E2E status: ran (result) or not run (why + how to run)
-153
View File
@@ -1,153 +0,0 @@
---
name: worktree-manager
description: Use when creating, removing, or troubleshooting git worktrees for isolated feature/branch work. Uses Worktrunk (git-wt) to provision sibling worktrees with env files, deps, and a .planning junction. Triggers on "create a worktree", "set up an isolated branch", "remove this worktree", "clean up a worktree", "spin up a sandbox".
tools: Bash, Read, Glob
model: sonnet
color: teal
---
You are the worktree-manager agent for WorldWideView. Your job is to create and remove git worktrees using Worktrunk (`git-wt`). You do NOT build features — you provision the isolated environments in which features are built.
**Tool requirement:** The installed `git-wt` must be Worktrunk (max-sixty, worktrunk.dev), NOT the npm package `git-wt` (mustafamagdy). Verify with `git-wt --version` before proceeding — it should print a Worktrunk version string. If it does not, stop and instruct the user to run:
```powershell
npm uninstall -g git-wt
winget install max-sixty.worktrunk
git-wt config shell install
```
---
## Step 1 — Context check
Before any operation, confirm you are operating on the WorldWideView repo:
```bash
git -C C:/dev/wwv/worldwideview rev-parse --show-toplevel
```
Expected: `C:/dev/wwv/worldwideview`. If the toplevel is under `local-plugins/`, `local-seeders/`, or another repo, abort.
---
## Step 2 — Create a worktree
```bash
cd C:/dev/wwv/worldwideview
git-wt switch --create <branch-name> --yes
```
Worktrunk reads `worldwideview/.config/wt.toml` and automatically runs every `pre-start` hook:
1. `install_deps` — `pnpm install`
2. `copy_env` — copies `.env` from the main repo
3. `copy_local_plugins` — rsyncs `local-plugins/` (excluding `node_modules`)
4. `copy_local_seeders` — rsyncs `local-seeders/` (excluding `node_modules`)
5. `link_planning` — removes any stale `.planning/` copy and creates a Windows junction to `C:\dev\wwv\.planning` (the ecosystem root with STATE.md, ROADMAP.md, phases/)
The new worktree lands at `C:\dev\wwv\worldwideview.<branch-name>\` (sibling of the main repo).
**No manual steps are needed.** Do NOT manually run `mklink`, `pnpm install`, or copy env files — the hooks handle all of this.
After creation, verify:
```powershell
Get-ChildItem C:\dev\wwv\worldwideview.<branch-name>\.planning | Select-Object Name
```
Expected output must include `STATE.md`, `ROADMAP.md`, `phases`. If `.planning` only shows `codebase`, the junction failed — see Troubleshooting below.
Report to the caller:
> Worktree created at `C:\dev\wwv\worldwideview.<branch-name>\`. All hooks ran. `.planning` is junctioned to the ecosystem root. You can open a new Claude Code session from that directory and `/gsd:*` commands will work.
---
## Step 3 — Remove a worktree
Run from the **main repo** (preferred — works even if the worktree directory is inaccessible):
```powershell
Set-Location "C:\dev\wwv\worldwideview"
git-wt remove --force --yes <branch-name>
```
Or from **inside** the worktree:
```powershell
Set-Location "C:\dev\wwv\worldwideview.<branch-name>"
git-wt remove --force --yes
```
**Why `--force`:** Worktrunk blocks removal if the worktree has uncommitted or untracked changes. `--force` bypasses this guard — use it intentionally, after confirming you don't need those changes.
**Why `--yes`:** Worktrunk prompts for per-hook approval in interactive terminals. Claude Code's Bash/PowerShell tools are non-interactive, so without `--yes` the removal fails silently with "Cannot prompt for approval". `--yes` auto-approves all pre-remove hooks for this run.
Worktrunk runs the pre-remove hooks in order: `unlink_planning` (removes the `.planning` junction without following it into the target) then `docker_clean` (`docker compose down -v`). The `-v` flag **destroys the worktree's Docker volumes** including the PostgreSQL database. The shared `C:\dev\wwv\.planning` is NOT affected.
After removal, verify the worktree is gone:
```powershell
git -C "C:\dev\wwv\worldwideview" worktree list
```
**NEVER** use `rm -rf` or `Remove-Item -Recurse` on a worktree directory — the Docker volume is orphaned and continues running in the background.
---
## Step 4 — Orphan recovery
If a worktree was manually deleted and its Docker volume is now orphaned, run from the **main repo root**:
```bash
pnpm run db:prune
```
---
## Troubleshooting
**`.planning` only shows `codebase` after creation:**
The `link_planning` hook may have failed (most likely because `.planning/` already existed and the `rmdir` didn't fire). Fix manually:
```powershell
powershell -Command "cmd /c 'rmdir /s /q C:\dev\wwv\worldwideview.<branch>\.planning && mklink /J C:\dev\wwv\worldwideview.<branch>\.planning C:\dev\wwv\.planning'"
```
Then verify again.
**Hook approval prompt on first use:**
Worktrunk requires per-machine approval the first time it sees a new hook. Either approve interactively or pass `--yes` to auto-approve all hooks and persist. The `--yes` flag is already in the recommended command above.
**`git-wt --version` doesn't mention Worktrunk:**
The wrong `git-wt` is installed. Run the swap commands at the top of this document.
---
## Common mistakes — refuse these
| Rationalization | Reality |
|---|---|
| "I'll use `git worktree add` since it's standard." | Skips all hooks — no env, no deps, no `.planning` junction. Always use `git-wt switch --create`. |
| "I'll run `mklink /J` manually after creation." | The `link_planning` hook already does this. Running it again creates a double-junction or silent failure. Only do this if the hook failed. |
| "I'll just delete the folder to clean up." | Docker volume is now orphaned. Use `git-wt remove` from inside the worktree. |
| "I'll use `rm -rf` or `Remove-Item -Recurse`." | Same problem as above. Never. |
| "I'll point `.planning` at `worldwideview\.planning`." | That is a stale partial copy (only `codebase/`). The junction MUST point at `C:\dev\wwv\.planning` (ecosystem root). |
| "I can install npm `git-wt` — it's the same thing." | It is NOT. The npm package ignores `wt.toml`, has no hooks, and puts worktrees in `~/.worktrees/`. Breaks everything. |
---
## Quick reference
| Action | Command | Where to run |
|---|---|---|
| Create worktree | `git-wt switch --create <branch> --yes` | Inside `worldwideview/` or any directory in the repo |
| Remove worktree | `git-wt remove --force --yes <branch>` | Main repo or inside worktree |
| Recover orphaned volumes | `pnpm run db:prune` | Main repo root |
| Verify `.planning` junction | `Get-ChildItem C:\dev\wwv\worldwideview.<branch>\.planning` | Anywhere |
---
## Return
Always report:
- Action taken (created / removed / recovered)
- Worktree absolute path and branch name
- Confirmation that `.planning` resolved correctly (or warning if it did not)
- Any warnings or issues encountered
-9
View File
@@ -1,9 +0,0 @@
Use the `worktree-manager` subagent to create a sibling worktree for branch `$ARGUMENTS` at `C:\dev\wwv\worldwideview.$ARGUMENTS\`.
The agent will:
1. Run `git-wt switch --create $ARGUMENTS --yes` from inside `C:\dev\wwv\worldwideview\`.
2. Confirm the worktree landed at `C:\dev\wwv\worldwideview.$ARGUMENTS\` (sibling directory).
3. Verify `.planning` resolves to `C:\dev\wwv\.planning` (ecosystem root).
4. Report the absolute path and any warnings.
If `$ARGUMENTS` is empty, ask the user for the branch name before proceeding.
-37
View File
@@ -1,37 +0,0 @@
#!/usr/bin/env bash
# Per-worktree .planning isolation check (non-destructive).
# Called by: Claude Code SessionStart hook, git post-checkout hook.
#
# New model (2026-05): each worktree owns a REAL, isolated .planning directory
# (NOT a junction to the shared C:/dev/wwv/.planning). The shared root holds only
# cross-feature docs (ROADMAP, MILESTONES, research, PROJECT, REQUIREMENTS).
#
# This hook NO LONGER switches .planning git branches (the old shared-repo model,
# which could not isolate simultaneous sessions). It only WARNS when the current
# worktree's .planning has not been isolated yet, so leftover junctions get noticed
# and converted via worktree-bootstrap Step 5. It deliberately does NOT mutate the
# filesystem: a SessionStart hook must not do risky junction surgery unattended.
# Only run inside the worldwideview repo or a worktree of it.
GIT_COMMON=$(git rev-parse --git-common-dir 2>/dev/null)
if ! echo "$GIT_COMMON" | grep -qi "worldwideview"; then
exit 0
fi
TOPLEVEL=$(git rev-parse --show-toplevel 2>/dev/null)
[ -z "$TOPLEVEL" ] && exit 0
PLANNING="$TOPLEVEL/.planning"
# No .planning yet: worktree-bootstrap Step 5 will create it. Stay quiet.
[ -e "$PLANNING" ] || exit 0
# Isolated worktrees have a WORKSPACE.md manifest. If present, all good.
if [ -f "$PLANNING/WORKSPACE.md" ]; then
exit 0
fi
# .planning exists but has no WORKSPACE.md: likely a leftover junction to the shared
# root (or an un-bootstrapped dir). Warn only.
echo "{\"continue\":true,\"suppressOutput\":false,\"message\":\"[planning-sync] This worktree's .planning is not isolated (no WORKSPACE.md). If it is a junction to the shared root, phases may bleed across worktrees. Isolate it via worktree-bootstrap Step 5.\"}"
exit 0
+27 -8
View File
@@ -22,17 +22,36 @@ paths:
## Auth Strategy
> [!NOTE]
> **See [ADR-003](../../docs/architecture/decisions/adr-0003-shared-identity-and-ecosystem-auth-host.md) for the canonical auth architecture.**
> The implementation described below (NextAuth as ecosystem IdP, `app.worldwideview.dev` as auth host) was superseded on 2026-05-22. Phase 2A-D (completed 2026-05-25) implemented the ADR-003 architecture:
> - **Supabase Auth** is the identity provider for the cloud edition
> - **`worldwideview.dev`** owns login, signup, and auth callback routes (`/login`, `/signup`, `/auth/callback`)
> - **NextAuth** (Credentials provider) remains active for the local and demo editions only
> - Session cookies use `@supabase/ssr` `createServerClient` scoped to `NEXT_PUBLIC_WWV_COOKIE_DOMAIN`
> - Both worldwideview and worldwideview-marketplace use the same `buildCookieOptions()` recipe
> [!WARNING]
> **This section is superseded by [ADR-003](../../docs/architecture/decisions/adr-0003-shared-identity-and-ecosystem-auth-host.md) (2026-05-22).**
> The decisions below describing `app.worldwideview.dev` as the auth owner and NextAuth as the ecosystem IdP are no longer current. The new architecture is:
> - **`worldwideview.dev`** (the `worldwideview-web` repo) owns all login, signup, and account management UI
> - **Supabase Auth** (single shared project) is the identity provider for all products
> - Session cookies are scoped to `.worldwideview.dev` and shared across all subdomains
> - The Marketplace never owns auth UI and never redirects to `app.worldwideview.dev`
>
> The multi-tenant RLS, tier matrix, license key, and CI/CD sections below remain valid.
**Auth.js (NextAuth)** is the single auth API across all editions. `app.worldwideview.dev` owns **all auth UI** — login, registration, password reset. No other subdomain ever shows a login or registration form.
```typescript
// src/lib/auth.ts
const providers = isCloud
? [SupabaseProvider({ /* delegates to Supabase Auth (GoTrue) */ })]
: [Credentials({ /* local bcrypt + PostgreSQL */ })];
export const { auth, signIn, signOut } = NextAuth({ providers });
```
Application code is **identical** across editions — always calls `auth()`, `signIn()`, `signOut()`.
**One account for the entire ecosystem:**
- Cloud instance (`[user].app.worldwideview.dev`)
- Marketplace (browse, install, publish) — via SSO redirect
- Local instance Bridge API connection
**Marketplace never owns auth UI** — it redirects to `app.worldwideview.dev/login?redirect_to=marketplace...` for all sign-in.
---
## Decentralized Plugin Authentication (Token Exchange)
-110
View File
@@ -72,23 +72,6 @@ mapWebsocketPayload(payload: any, existingEntities: GeoEntity[]): GeoEntity[]
- `existingEntities` — current entities in the store for this plugin (use for incremental updates)
- Must return a **complete replacement** `GeoEntity[]` (not a diff)
### Property Tag Helpers
Always wrap typed property values with the SDK helpers so the Intel panel renders them richly:
```ts
import { dtProp, urlProp, imageProp, videoProp } from "@worldwideview/wwv-plugin-sdk";
properties: {
updated: dtProp(item.updated_at ?? null), // renders as expandable datetime row
source: urlProp(item.source_url ?? null), // renders as clickable link
preview: imageProp(item.image ?? null), // renders as inline thumbnail
stream: videoProp(item.video_url ?? null), // renders as "Watch" link
}
```
All helpers are null-safe and return `null` for empty input. Plain string/number values that don't need rich rendering need no wrapper.
---
## Plugin Architectures (Manifest Formats)
@@ -503,96 +486,3 @@ Seeders in `wwv-data-engine/src/seeders/` provide the backend data.
- **Cron Seeder**: Uses `registerSeeder({ name: "plugin-id", cron: "..." })` and pushes to Redis via `setLiveSnapshot()`.
- **Init Seeder**: High-frequency or persistent websockets via `init: () => void`.
- **Constraint**: Seeder `name` MUST exactly match the frontend plugin `id`. Do not bundle workspace dependencies like `ws` or `zod` into the seeder `dist`.
---
## Plugin Bundle Footguns
### 1. CSP blocks CDN stylesheets -- use `?inline` + `<style>` injection
The host app's `style-src` policy includes `'unsafe-inline'` but does **not** include external CDN origins like `unpkg.com` or `cdn.jsdelivr.net`. Any plugin that loads CSS via a runtime `<link href="https://unpkg.com/...">` will have that stylesheet **silently blocked** by the browser -- no console error, just broken styling and broken layout (elements fall back to `position: static`).
**Wrong:**
```typescript
// In useEffect:
link.href = "https://unpkg.com/gridstack@11/dist/gridstack.min.css";
document.head.appendChild(link);
```
**Correct -- bundle CSS as inline string and inject via `<style>` tag:**
```typescript
import gridstackCSS from "gridstack/dist/gridstack.min.css?inline";
// In useEffect:
const style = document.createElement("style");
style.textContent = gridstackCSS;
document.head.appendChild(style);
```
Vite's `?inline` suffix imports CSS as a plain string. The `<style>` tag is allowed by `'unsafe-inline'`. This is the only CSP-safe pattern for third-party CSS in plugin bundles.
---
### 2. `inlineDynamicImports: true` -- not `codeSplitting: false`
`codeSplitting: false` is **not a valid Rollup option**. It is silently ignored, and Rollup will still produce code-split chunks. The host can only load a single `frontend.mjs` -- any sibling chunk (e.g. `gridstack-XXXXX.js`) will 404.
**Wrong:**
```js
rollupOptions: {
output: { codeSplitting: false } // silently ignored
}
```
**Correct:**
```js
rollupOptions: {
output: { inlineDynamicImports: true } // prevents all code splitting
}
```
Both `vite.config.ts` (for manual `pnpm build`) and `scripts/sync-local-plugins.mjs` must use `inlineDynamicImports: true`.
---
### 3. `process.env.NODE_ENV` in CJS deps -- needs inline Rollup transform, not Vite `define`
CJS dependencies like `recharts`, `prop-types`, and `react-is` reference `process.env.NODE_ENV` inside their CommonJS module body. Vite's top-level `define: { 'process.env.NODE_ENV': '"production"' }` does NOT reach inside Rollup's `__commonJSMin()` wrappers -- those run after the `define` substitution phase.
**Wrong:**
```js
// In vite.config.ts or sync-local-plugins.mjs:
define: { 'process.env.NODE_ENV': '"production"' } // doesn't reach CJS wrappers
```
**Correct -- add an inline Rollup transform plugin inside `rollupOptions.plugins`:**
```js
rollupOptions: {
plugins: [
{
name: 'replace-process-env',
transform(code) {
if (!code.includes('process.env.NODE_ENV')) return null;
return {
code: code.replace(/process\.env\.NODE_ENV/g, '"production"'),
map: null,
};
},
},
// ... other plugins
]
}
```
This plugin runs on each module source before Rollup wraps it, so it correctly replaces `process.env.NODE_ENV` in CJS sub-dependencies.
---
### 4. Yahoo Finance and third-party data APIs -- CORS from the browser
Yahoo Finance's chart API (`query1.finance.yahoo.com`) does **not** return `Access-Control-Allow-Origin` headers. Direct `fetch()` from a plugin bundle running in the browser will be CORS-blocked.
**Architectural preference (in order):**
1. **Best:** Have the wwv-data-engine seeder fetch the data server-side and stream it to the plugin via WebSocket -- no CORS issue.
2. **Acceptable stopgap:** Use `corsproxy.io` relay: `fetch('https://corsproxy.io/?url=' + encodeURIComponent(apiUrl))`. This adds a third-party relay dependency.
3. **Do not do:** Assume browser `fetch()` to third-party financial APIs will work -- most do not send CORS headers.
-161
View File
@@ -1,161 +0,0 @@
param([string]$worktreePath)
# git-wt (Worktrunk) pre-start hook.
# Per the per-worktree isolation model, each worktree gets its OWN real, isolated
# .planning directory -- NEVER a junction to the shared root. A shared junction caused
# GSD phase bleed across simultaneous worktree sessions. See worktree-bootstrap SKILL Step 5
# and C:\dev\wwv\.planning\WORKSPACE-OPTIMIZATION-BACKLOG.md.
#
# This hook is idempotent:
# - existing REAL .planning with a WORKSPACE.md -> left untouched (preserve phase history)
# - existing REAL .planning missing manifests -> manifests seeded, contents preserved
# - existing JUNCTION (old-model leftover) -> junction removed, real dir created
# - nothing there -> real isolated dir created
$ErrorActionPreference = 'Stop'
# Convert Unix-style path (/c/dev/...) to Windows path (C:\dev\...) if needed
if ($worktreePath -match '^/([a-zA-Z])/(.+)$') {
$worktreePath = $matches[1].ToUpper() + ':/' + $matches[2]
}
$sharedRoot = 'C:\dev\wwv\.planning'
$target = Join-Path $worktreePath '.planning'
# If a junction/symlink leftover from the old model exists, remove only the reparse
# point (never recurse into the shared root it points at).
if (Test-Path $target) {
$item = Get-Item -LiteralPath $target -Force
$isReparse = ($item.Attributes -band [IO.FileAttributes]::ReparsePoint) -eq [IO.FileAttributes]::ReparsePoint
if ($isReparse) {
[System.IO.Directory]::Delete($target, $false)
} elseif (Test-Path (Join-Path $target 'WORKSPACE.md')) {
# Already a fully set-up real isolated dir -- preserve it, do nothing.
exit 0
}
}
# Resolve the branch name (fall back to the worktree folder leaf).
$branch = $null
try { $branch = (& git -C $worktreePath rev-parse --abbrev-ref HEAD 2>$null) } catch {}
if (-not $branch) { $branch = Split-Path $worktreePath -Leaf }
$branch = $branch.Trim()
$date = (Get-Date).ToString('yyyy-MM-dd')
$isoDate = (Get-Date).ToUniversalTime().ToString('yyyy-MM-ddTHH:mm:ss.000Z')
# Create the isolated structure.
foreach ($d in @('phases', 'debug', 'surveys')) {
New-Item -ItemType Directory -Force -Path (Join-Path $target $d) | Out-Null
}
# Copy the GSD config from the shared root if present.
$sharedConfig = Join-Path $sharedRoot 'config.json'
$localConfig = Join-Path $target 'config.json'
if ((Test-Path $sharedConfig) -and -not (Test-Path $localConfig)) {
Copy-Item $sharedConfig $localConfig -Force
}
# Seed the three manifest files (only if missing -- never overwrite existing).
$workspaceFile = Join-Path $target 'WORKSPACE.md'
if (-not (Test-Path $workspaceFile)) {
$workspace = @"
# Workspace: $branch
Created: $date
Strategy: worktree (git-wt)
Branch: $branch
Repo: $worktreePath
## Isolation
This ``.planning`` is a REAL, isolated directory (not a junction to the shared root).
Phases and STATE.md here are private to this worktree. Changes do NOT bleed into
other worktrees or the shared root. Created automatically by the git-wt pre-start
hook (link-planning.ps1), per the per-worktree isolation model.
## Member Repos
| Repo | Source | Branch | Strategy |
|------|--------|--------|----------|
| worldwideview | C:\dev\wwv\worldwideview | $branch | worktree |
## Notes
Cross-feature docs are NOT duplicated here. See SHARED-DOCS.md.
"@
Set-Content -LiteralPath $workspaceFile -Value $workspace -Encoding UTF8
}
$stateFile = Join-Path $target 'STATE.md'
if (-not (Test-Path $stateFile)) {
$state = @"
---
gsd_state_version: 1.0
milestone: none
milestone_name: $branch
status: idle
last_updated: "$isoDate"
last_activity: $date -- isolated .planning created (git-wt pre-start hook)
progress:
total_phases: 0
completed_phases: 0
total_plans: 0
completed_plans: 0
percent: 0
---
# Project State
## Project Reference
This is an ISOLATED worktree workspace. Its planning state is private to the
``$branch`` worktree and intentionally distinct from the shared root
(C:\dev\wwv\.planning).
Cross-feature docs (ROADMAP, MILESTONES, research) are NOT duplicated here.
See SHARED-DOCS.md for how to reach them.
## Current Position
No active milestone. Workspace is idle.
"@
Set-Content -LiteralPath $stateFile -Value $state -Encoding UTF8
}
$sharedDocsFile = Join-Path $target 'SHARED-DOCS.md'
if (-not (Test-Path $sharedDocsFile)) {
$sharedDocs = @"
# Shared Cross-Feature Docs
This worktree's ``.planning`` is isolated (phases + STATE are private). But some
documents are cross-feature and must stay authoritative in ONE place rather than
being copied per worktree (copies drift out of sync).
## Where they live
Shared root: ``C:\dev\wwv\.planning``
| Doc | Purpose |
|-----|---------|
| ROADMAP.md | Cross-milestone roadmap |
| MILESTONES.md | Milestone definitions and ship status |
| research/ | ARCHITECTURE / FEATURES / PITFALLS / STACK |
| PROJECT.md | Project identity |
| REQUIREMENTS.md | Cross-cutting requirements |
## How to reach them
A ``WWV_SHARED_PLANNING`` env var in each worktree's ``.env.local`` points at the
shared root, and GSD skills that need a cross-feature doc read
``\$WWV_SHARED_PLANNING/<doc>`` instead of the local ``.planning/<doc>``.
Do NOT copy these docs into this worktree. Single source of truth lives at the
shared root.
"@
Set-Content -LiteralPath $sharedDocsFile -Value $sharedDocs -Encoding UTF8
}
# Always report success: the internal git probe above may set a non-zero $LASTEXITCODE,
# which would otherwise be propagated as this hook's exit code and read as a failure.
exit 0
-29
View File
@@ -1,29 +0,0 @@
param([string]$worktreePath)
# git-wt (Worktrunk) pre-remove hook.
# Per the per-worktree isolation model, a worktree's .planning is a REAL isolated
# directory (its phase history is archived by /branch-cleanup before removal, and the
# directory itself is deleted along with the worktree by `git-wt remove`).
#
# This hook therefore ONLY cleans up a leftover JUNCTION from the old shared-link model.
# It must NEVER recursively delete a real .planning directory -- doing so would destroy
# phase history that /branch-cleanup is responsible for archiving.
# Convert Unix-style path (/c/dev/...) to Windows path (C:\dev\...) if needed
if ($worktreePath -match '^/([a-zA-Z])/(.+)$') {
$worktreePath = $matches[1].ToUpper() + ':/' + $matches[2]
}
$target = Join-Path $worktreePath '.planning'
if (Test-Path $target) {
$item = Get-Item -LiteralPath $target -Force
$isReparse = ($item.Attributes -band [IO.FileAttributes]::ReparsePoint) -eq [IO.FileAttributes]::ReparsePoint
if ($isReparse) {
# Old-model junction: remove only the reparse point (do not follow into the target).
[System.IO.Directory]::Delete($target, $false)
}
# Real directory: leave it. git-wt remove deletes the worktree (and this dir) anyway;
# /branch-cleanup is responsible for archiving its phase history first.
}
exit 0
-148
View File
@@ -1,148 +0,0 @@
{
"_comment": "PORTABLE TEMPLATE. Copy to .agents/settings.json (gitignored, machine-local) and adjust if needed. Hook script paths use $CLAUDE_PROJECT_DIR so they resolve per-worktree on any machine/repo location. The bash interpreter path assumes a standard Windows Git install (C:/Program Files/Git/bin/bash.exe); if yours differs, override it in your local .agents/settings.json. The referenced gsd-* hooks are provided by the GSD install (gitignored), not this repo, so run the GSD installer first.",
"permissions": {
"allow": [
"Bash(cymbal *)",
"Bash(gsd-sdk query *)"
]
},
"hooks": {
"SessionStart": [
{
"hooks": [
{
"type": "command",
"command": "node",
"args": [
"$CLAUDE_PROJECT_DIR/.claude/hooks/gsd-check-update.js"
]
}
]
},
{
"hooks": [
{
"type": "command",
"command": "C:/Program Files/Git/bin/bash.exe",
"args": [
"$CLAUDE_PROJECT_DIR/.claude/hooks/gsd-session-state.sh"
]
}
]
},
{
"hooks": [
{
"type": "command",
"command": "C:/Program Files/Git/bin/bash.exe",
"args": [
"$CLAUDE_PROJECT_DIR/.claude/hooks/planning-sync.sh"
]
}
]
}
],
"PostToolUse": [
{
"matcher": "Bash|Edit|Write|MultiEdit|Agent|Task",
"hooks": [
{
"type": "command",
"command": "node",
"args": [
"$CLAUDE_PROJECT_DIR/.claude/hooks/gsd-context-monitor.js"
],
"timeout": 10
}
]
},
{
"matcher": "Read",
"hooks": [
{
"type": "command",
"command": "node",
"args": [
"$CLAUDE_PROJECT_DIR/.claude/hooks/gsd-read-injection-scanner.js"
],
"timeout": 5
}
]
},
{
"matcher": "Write|Edit",
"hooks": [
{
"type": "command",
"command": "C:/Program Files/Git/bin/bash.exe",
"args": [
"$CLAUDE_PROJECT_DIR/.claude/hooks/gsd-phase-boundary.sh"
],
"timeout": 5
}
]
}
],
"PreToolUse": [
{
"matcher": "Write|Edit",
"hooks": [
{
"type": "command",
"command": "node",
"args": [
"$CLAUDE_PROJECT_DIR/.claude/hooks/gsd-prompt-guard.js"
],
"timeout": 5
}
]
},
{
"matcher": "Write|Edit",
"hooks": [
{
"type": "command",
"command": "node",
"args": [
"$CLAUDE_PROJECT_DIR/.claude/hooks/gsd-read-guard.js"
],
"timeout": 5
}
]
},
{
"matcher": "Write|Edit",
"hooks": [
{
"type": "command",
"command": "node",
"args": [
"$CLAUDE_PROJECT_DIR/.claude/hooks/gsd-workflow-guard.js"
],
"timeout": 5
}
]
},
{
"matcher": "Bash",
"hooks": [
{
"type": "command",
"command": "C:/Program Files/Git/bin/bash.exe",
"args": [
"$CLAUDE_PROJECT_DIR/.claude/hooks/gsd-validate-commit.sh"
],
"timeout": 5
}
]
}
]
},
"enabledPlugins": {
"stripe@claude-plugins-official": false,
"frontend-design@claude-code-plugins": false
},
"worktree": {
"bgIsolation": "none"
}
}
+1 -2
View File
@@ -5,8 +5,7 @@
"permissions": {
"allow": [
"Read(src/**)",
"Bash(npm test)",
"Bash(node \"C:/Users/silve/.claude/hooks/claude-dir-guard.js\")"
"Bash(npm test)"
],
"deny": [
"Read(.env)",
-154
View File
@@ -1,154 +0,0 @@
---
name: branch-cleanup
description: Post-merge lifecycle cleanup. Investigates the full state first, presents one decision summary, then executes after user approval. Commits leftover artifacts, deletes the session plan file, archives the worktree's isolated .planning, and delegates worktree removal to the worktree-manager agent. Pairs with branch-finisher as the closing bookend of a feature branch.
---
# Branch Cleanup
## Overview
The post-merge companion to `branch-finisher`. Where `branch-finisher` ships a branch (commit -> push -> PR), `branch-cleanup` closes the loop after the PR merges: commit leftovers, discard the session plan, and tear down the worktree cleanly.
Invoke with `/branch-cleanup` when a PR has been merged and the worktree is no longer needed.
## When to Use
- The PR for this branch has been merged to main
- Untracked artifacts exist that need committing or discarding
- A session plan file exists in `.claude/plans/` and the work is fully shipped
- User says "clean up", "we're done", "remove this worktree", "tear down this branch"
---
## Procedure
### Phase 1 - Investigate everything first (no actions yet)
Run all of these in parallel before asking the user anything:
**1a. PR status**
```bash
gh pr list --head <current-branch> --state all
```
Note: merged PR numbers, titles, and merge dates.
**1b. Untracked files**
```bash
git status
```
For each untracked item, classify it:
| Item | Classification |
|---|---|
| `.planning/` | **Real isolated dir** - archive before removal. Each worktree now owns a real, isolated `.planning` (phases + `STATE.md` private to this branch). Before tearing the worktree down, archive it to the shared archive so the phase history survives (Phase 3e). It is gitignored, so never commit it to the feature branch and never `git clean` it. |
| `local-scripts/` | Scratch scripts - likely discard |
| Any other untracked path | Unknown - needs user decision |
**1c. Modified tracked files**
Check `git diff` and `git diff --staged` for any uncommitted changes to tracked files.
**1d. Plan files**
```powershell
Get-ChildItem $env:USERPROFILE\.claude\plans\ -Name
```
Identify which plan files belong to this branch's work (by name or content). A plan is safe to delete if its PR is merged and no "Remaining work" section has open items.
---
### Phase 2 - Present one decision summary
After gathering all of the above, show the user a single block:
```
Branch: <branch-name>
PRs merged: #<n> — <title>, #<n> — <title>
Untracked artifacts:
.planning/ → real isolated dir; archive to shared archive before removal
<other-path> → [commit as docs: / discard] ← your choice
Modified tracked files:
<file> → [commit / discard] ← your choice
(none)
Plan files to delete:
<plan-file>.md → [delete / keep] ← your choice
After your approval:
1. Commit / discard artifacts as decided above
2. Delete selected plan files
3. Archive .planning, then remove worktree via worktree-manager (destroys Docker volume, irreversible)
Proceed?
```
**Wait for the user's typed response before doing anything.**
---
### Phase 3 - Execute after approval
Execute the approved actions in order:
**3a. Handle modified tracked files first** (if any)
Use the `/commit` skill — dirty working tree must be clean before worktree removal.
**3b. Commit approved artifacts**
Use the `/commit` skill with type `docs:` (patch bump). Stage only the specific approved paths — never `git add .` or `git add -A`.
**3c. Discard rejected artifacts**
```bash
git clean -fd <specific-path>
```
Never run `git clean -fd` without a specific path.
**3d. Delete approved plan files**
```powershell
Remove-Item "$env:USERPROFILE\.claude\plans\<plan-file>.md"
```
**3e. Archive this worktree's `.planning`**
The worktree's `.planning` is a real, isolated directory. Preserve its phase history before removal:
```powershell
$src = "C:\dev\wwv\worldwideview.<branch>\.planning"
$dst = "C:\dev\wwv\.planning\archive\<branch>"
New-Item -ItemType Directory -Force -Path $dst | Out-Null
Copy-Item "$src\*" $dst -Recurse -Force
```
Skip if `.planning` has no phases worth keeping (fresh `milestone: none` state). The shared archive lives at `C:\dev\wwv\.planning\archive\`.
**3f. Remove the worktree**
Delegate to the `worktree-manager` agent:
> "Remove the worktree for branch `<branch-name>`."
The agent runs `git-wt remove --force --yes <branch>` from the main repo, tears down Docker volumes, and verifies cleanup. (The worktree's real `.planning` is removed along with the worktree directory, which is why Phase 3e archives it first.)
**Do not run `git-wt remove` directly** - use the agent so verification and troubleshooting are handled correctly.
---
### Phase 4 - Summary
Report:
- PRs merged: `#<n> — <title>` (list)
- Artifacts: committed / discarded (list)
- Plan files: deleted / kept
- Worktree: removed from `C:\dev\wwv\worldwideview.<branch>\`
- Next step:
```bash
cd C:\dev\wwv\worldwideview && git pull
```
---
## Quick reference
| Phase | Action | Rule |
|---|---|---|
| Investigate | PR status, git status, plan files | Parallel - no actions yet |
| Decide | One summary block | Wait for user approval |
| Execute | Commit, delete plans, remove worktree | In order, after approval |
| `.planning/` | Archive then remove | Real isolated dir; preserve phase history first |
-225
View File
@@ -1,225 +0,0 @@
---
name: full-manual-e2e
description: User-triggered only. Invoke when the user explicitly runs /full-manual-e2e, or when /gsd:progress routes to manual verification. Do NOT auto-activate. Full WWV ecosystem E2E test script covering 3-server startup on https://wwv.local, chrome-devtools MCP browser driving, and 7-step auth+install validation flow.
---
# WWV Full Manual E2E
## Overview
The WorldWideView ecosystem is three Next.js apps that share a Supabase session via cookies scoped to `.wwv.local`. Testing the full install flow requires all three running simultaneously on HTTPS at sibling hostnames so the cookie domain trick works. This skill covers the exact commands, the gotchas that bite every time, and how to drive the browser via chrome-devtools MCP to validate each step.
## Architecture quick reference
| App | Local URL | Auth | Notes |
|---|---|---|---|
| `worldwideview-web` | `https://wwv.local:3001` | Supabase (`@supabase/ssr`) | Apex auth host. Owns `/login`, `/signup`, `/auth/callback`, `/accounts`. Refreshes session via `proxy.ts`. |
| `worldwideview-marketplace` | `https://marketplace.wwv.local:3002` | Supabase (cookie inherited) | Owns `/api/install/start` gate, `/api/instances*` routes, `InstanceCapture` + `InstanceHydrator`. |
| `worldwideview` (instance) | `https://wwv.local:3000` | NextAuth (local edition) | Local edition uses NextAuth credentials. Cloud edition delegates to the auth host. |
Shared Supabase session cookie: `sb-<project-ref>-auth-token.0/.1`, `Domain=.wwv.local`, `Secure=true`, `SameSite=Lax`. Different ports do NOT break cookie sharing (cookies are domain-based, not origin-based).
## Prerequisites (verify each — they are non-obvious)
1. **Hosts file** (Windows: `C:\Windows\System32\drivers\etc\hosts`, requires admin):
```
127.0.0.1 wwv.local
127.0.0.1 marketplace.wwv.local
```
2. **HTTPS in dev.** Either:
- Run each server with `--experimental-https` (Next.js auto-generates a self-signed cert — browser warns once per host, click through)
- OR run `mkcert -install` once + `mkcert "*.wwv.local" wwv.local` to get a trusted local CA
3. **Docker running** — only needed for the `worldwideview` instance (its `predev` script runs `prisma db push` against a Postgres container).
4. **Supabase project provisioned.** All three apps must point at the same `NEXT_PUBLIC_SUPABASE_URL` in their respective `.env.local`. The marketplace and worldwideview-web also need `NEXT_PUBLIC_SUPABASE_PUBLISHABLE_KEY` and `NEXT_PUBLIC_WWV_COOKIE_DOMAIN=.wwv.local`.
5. **`NEXT_PUBLIC_AUTH_HOST_URL=https://wwv.local:3001`** set in marketplace `.env.local` so the install gate knows where to send unauthenticated users.
## Server startup (exact commands that work)
Each app's `pnpm dev` script varies — don't assume one command fits all. These are what actually work:
```powershell
# Terminal 1 — Auth host (worldwideview-web)
cd C:/dev/wwv/worldwideview-web
pnpm exec next dev --experimental-https --port 3001 -H wwv.local
# Terminal 2 — Marketplace
cd C:/dev/wwv/worldwideview-marketplace
pnpm exec next dev --webpack --experimental-https --port 3002 -H marketplace.wwv.local
# Terminal 3 — WWV instance (requires Docker)
cd C:/dev/wwv/worldwideview
pnpm dev
```
**Why `pnpm exec next dev` and not `pnpm dev` for the first two:** the marketplace `dev` script uses `--webpack` (mandatory — Turbopack breaks middleware), and we need to pass `-H` + `--port` which the npm script doesn't forward cleanly. WWV's `dev` script (`scripts/dev.mjs`) runs `concurrently` + handles predev (Docker boot, Prisma push, Cesium copy) — don't bypass it.
**Watching servers from inside an agent session:** use `Bash` with `run_in_background: true` for each, then `Monitor` with `tail -f <output> | grep --line-buffered -E "⨯|FAILED|Error:|TypeError|EADDR|ECONN|500"` for ongoing error notifications. Persistent monitors stay armed for the session.
## Common server-startup failures
| Symptom | Cause | Fix |
|---|---|---|
| `prisma db push --accept-data-loss` exits non-zero | Docker not running, Postgres unreachable | Start Docker Desktop, then `docker compose up -d wwv-db` from `worldwideview/`, then relaunch |
| `ENOENT: ...middleware.js.nft.json` during build | Next.js 16 renamed `middleware.ts` → `proxy.ts` | Rename the file and the exported function (`middleware` → `proxy`). Same APIs otherwise. |
| `PrismaConfigEnvError: Cannot resolve environment variable: DATABASE_URL` | Prisma 7 config file doesn't read `.env.local` automatically | `export DATABASE_URL="file:./prisma/registry.db"` before running prisma commands, OR run via the `pnpm dev` script which loads env |
| Cookie not visible on sibling subdomain | Wrong cookie domain or browser blocking insecure cookies | Verify `NEXT_PUBLIC_WWV_COOKIE_DOMAIN=.wwv.local` (leading dot mandatory). Cookies need `secure: true`, which requires HTTPS. |
| Browser keeps redirecting `?next=` to `/accounts` | `safeNext` rejecting cross-subdomain return URL | Confirm `worldwideview-web/src/lib/safeNext.ts` allows hostnames ending in `NEXT_PUBLIC_WWV_COOKIE_DOMAIN` |
## Driving the browser via chrome-devtools MCP
The MCP **drives the user's actual browser** — cookies, localStorage, and active sessions from prior testing persist across navigation. Treat it as live shared state, not a fresh fixture.
Tools loaded on first use via `ToolSearch`:
- `mcp__chrome-devtools__list_pages` — see open tabs
- `mcp__chrome-devtools__navigate_page` — `type=url|reload|back|forward`
- `mcp__chrome-devtools__new_page` — opens a new tab
- `mcp__chrome-devtools__take_snapshot` — a11y tree (preferred over screenshots for finding `uid=`s)
- `mcp__chrome-devtools__fill_form` — batch fill (use over individual `fill` calls)
- `mcp__chrome-devtools__click` — by `uid` from snapshot
- `mcp__chrome-devtools__wait_for` — by text content
- `mcp__chrome-devtools__evaluate_script` — arbitrary page JS (must be a function declaration, returns JSON)
- `mcp__chrome-devtools__list_network_requests` — `includePreservedRequests: true` to see across navigations
- `mcp__chrome-devtools__take_screenshot` — saves to a file path you specify
## The 7-step E2E test script
Run in this order. Each builds on the previous. After Step 3 the database has a `LinkedInstance` row that subsequent steps depend on.
### Step 1 — Sign up at the auth host
Navigate to `https://wwv.local:3001/signup`. Fill email + password, click Sign Up. Verify the email link. Expected landing: `/accounts` showing "Signed in as <email>".
Use Supabase dashboard → Auth → Users to confirm `email_confirmed_at` is set.
### Step 2 — Cookie cross-subdomain check
Same browser session. Open `https://marketplace.wwv.local:3002` in a new tab. The marketplace should recognise you as signed in.
Programmatic verification via `evaluate_script`:
```js
async () => {
const probe = await fetch('/api/instances');
return { status: probe.status, body: probe.status === 200 ? await probe.json() : null };
}
```
Expected: `status: 200`. (401 means the cookie didn't cross — check Domain attribute.)
### Step 3 — `?from_instance=` capture writes server-side
Navigate to `https://marketplace.wwv.local:3002/?from_instance=https%3A%2F%2Fwwv.local%3A3000`. Then:
```js
async () => {
await new Promise(r => setTimeout(r, 1500)); // let InstanceCapture useEffect run
const list = await fetch('/api/instances').then(r => r.json());
return {
ls: localStorage.getItem('wwv_instance_url'),
urlStripped: !window.location.search.includes('from_instance'),
list,
};
}
```
Expected: `ls = "https://wwv.local:3000"`, URL param stripped, `list.instances` has one row matching the captured URL.
### Step 4 — Hydrator restores localStorage from server
```js
() => { localStorage.removeItem('wwv_instance_url'); return { cleared: true }; }
```
Then reload the page. Then:
```js
async () => {
await new Promise(r => setTimeout(r, 2000)); // let Hydrator's fetch resolve
return { ls: localStorage.getItem('wwv_instance_url') };
}
```
Expected: `ls` is repopulated with the most-recently-used instance URL. Network log should show one `GET /api/instances` (no `POST` — Hydrator only reads).
### Step 5 — Install gate redirects anonymous users to `?next=`
Delete the Supabase cookie (DevTools Application → Cookies → `.wwv.local` → delete `sb-...-auth-token.0/.1`). Click Install on any plugin browse page. Expected: 307 redirect to `https://wwv.local:3001/login?next=<full encoded install/start URL>`. After signing in, browser bounces back through `/api/install/start` to the WWV instance.
Direct test (no UI needed):
```js
async () => {
const url = new URL('/api/install/start', location.origin);
url.searchParams.set('pluginId', 'aviation');
url.searchParams.set('version', '0.0.0');
url.searchParams.set('manifest', 'eyJpZCI6ImF2aWF0aW9uIn0=');
url.searchParams.set('instanceUrl', 'https://wwv.local:3000');
url.searchParams.set('redirectTo', location.href);
const res = await fetch(url.toString(), { redirect: 'manual' });
return { status: res.status, location: res.headers.get('location') };
}
```
### Step 6 — Picker appears for multi-instance users
Seed a second instance:
```js
await fetch('/api/instances/link', {
method: 'POST', headers: {'content-type':'application/json'},
body: JSON.stringify({ url: 'https://demo.wwv.local:9999' }),
});
```
**Gotcha:** the InstanceHydrator runs once on layout mount. If `localStorage.wwv_instance_url` is already set when the page loads, the Hydrator early-returns and the picker won't appear. To force the picker:
```js
() => {
localStorage.removeItem('wwv_instance_url');
// Click in the SAME tick so nothing async re-populates between clear and click.
const btn = Array.from(document.querySelectorAll('button')).find(b => b.textContent?.includes('Install Plugin'));
btn.click();
return { cleared: localStorage.getItem('wwv_instance_url') === null, clicked: !!btn };
}
```
Then `wait_for({ text: ["Choose your instance"] })`. Expected: modal lists both instances ordered by `lastUsedAt DESC` (most recent first), each labelled by nickname or `new URL(url).host`.
### Step 7 — Manage page rename / delete
Navigate to `https://marketplace.wwv.local:3002/manage`. The "Your linked instances" section lists rows. Click a nickname → type → blur → `PATCH /api/instances/<id>` fires. Click Remove → confirm → `DELETE /api/instances/<id>` fires.
## Critical gotchas (will burn you again)
1. **Chrome MCP drives the user's real browser**, not a sandboxed instance. Cookies, localStorage, and tabs from earlier testing persist. Always clear state explicitly when verifying a "cold load" scenario.
2. **The Hydrator only runs once per layout mount.** Full navigation re-mounts and re-runs it. Same-page state changes don't. To test the "cold device" flow, do `clear localStorage → click` in the same `evaluate_script` call so no async useEffect can re-populate between them.
3. **`secure: true` on the cookie requires HTTPS even in dev.** Plain `localhost:3000` cannot share cookies with `marketplace.localhost:3002`. The `wwv.local` hostname + HTTPS is mandatory.
4. **Cookies are domain-based, not origin-based.** `wwv.local:3000` and `wwv.local:3001` share cookies because the cookie's `domain=.wwv.local` ignores port. This is the only reason the auth host and the WWV instance can both run at `wwv.local` on different ports without conflict.
5. **Supabase session cookies are `httpOnly: true` ecosystem-wide (ADR-0004).** `document.cookie` will NOT show `sb-...-auth-token.0/.1`. All auth reads MUST be server-side (`createServerClient` → `auth.getClaims()` / `auth.getUser()`). Adding a client-side `supabase.auth.*` call is a regression — code review will flag it.
6. **WWV's `?from_instance=` originates from the local instance's "Browse plugins" button** (`worldwideview/src/components/panels/PluginsTab.tsx:70`). When testing the capture without WWV running, you can simulate by hand-crafting the URL.
7. **The marketplace install gate uses `?next=` not `?callbackUrl=`.** Standardized to match `worldwideview-web`'s `safeNext`. Don't mix them — `safeNext` ignores `?callbackUrl=` and silently redirects to `/accounts`.
8. **Turbopack + `@supabase/ssr` is broken in Next.js 16.2.3.** After HMR cycles you get `TypeError: adapterFn is not a function` from inside the proxy/middleware stack. Run worldwideview-web with `--webpack` until a known-good Turbopack version is verified. Marketplace `pnpm dev` already passes `--webpack` mandatorily — keep it. Full command: `pnpm exec next dev --experimental-https --webpack --port 3001 -H wwv.local`.
9. **Cookie-option changes only apply to newly-issued cookies.** If you edit `cookieOptions.ts` (e.g. flipping `httpOnly`) and reload, the browser still presents the OLD cookie until a token refresh or a fresh `signInWithPassword`. To verify a flag change, force a re-issue: sign out → sign in again. Don't trust a stale cookie inspection — it shows the old state.
10. **Next.js dev server OOMs after ~60 minutes of heavy use.** Symptom: V8 mark-compact failures around 16 GB heap (`FATAL ERROR: Ineffective mark-compacts near heap limit`), process exits with code 134 (SIGABRT). Restart the dev server every ~45 min during long E2E sessions. If the port stays bound after crash (`EADDRINUSE 3001`), kill the orphan: `Get-Process -Id (Get-NetTCPConnection -LocalPort 3001 -State Listen).OwningProcess | Stop-Process -Force` (PowerShell). The Git-Bash `taskkill /PID <n> /F` does NOT work — it mangles the flag.
11. **Prisma 7 + TypeScript config doesn't auto-load `.env.local`.** `pnpm prisma db push` (or any prisma CLI) fails with `PrismaConfigEnvError: Cannot resolve environment variable: DATABASE_URL` unless the variable is in the shell environment. Fix: `export DATABASE_URL="file:./prisma/registry.db"` before the prisma command, OR run prisma via the `pnpm dev` script which loads env first.
12. **Hydrator overwrites stale localStorage on every cold mount (ADR/UX choice).** Don't assume `localStorage.wwv_instance_url` survives across page loads — the `InstanceHydrator` always re-syncs to the server's most-recent `LinkedInstance`. To test the picker, you must clear localStorage AND click Install in the same `evaluate_script` tick, or the Hydrator's `useEffect` will re-populate before you can interact.
## Verification checklist before closing the session
- [ ] All three servers still running (check `Monitor` task IDs aren't dead)
- [ ] No unhandled errors fired in any server's log monitor during your session
- [ ] DB rows you created (`LinkedInstance` etc.) are cleaned up or you've noted them for the next session
- [ ] Cookie state in browser matches what your test expects (don't leave half-deleted cookies for the next session)
- [ ] Stop background server tasks if you don't need them: `TaskStop` with each `task_id`
## What this skill does NOT cover
- The cloud edition WWV flow (not implemented yet — local edition only).
- The full install completion at the WWV instance (requires Docker + Postgres + plugin DB seed). Stop after the install-redirect 302 is sufficient for verifying the marketplace gate; the WWV-side install is its own scope.
- E2E for the Playwright suite (those run headless in CI; see `worldwideview/tests/marketplace-from-instance.spec.ts` and the `e2e-testing.md` rule).
- mkcert setup details — covered well by `mkcert -install` docs.
-93
View File
@@ -1,93 +0,0 @@
---
name: plugin-new
description: End-to-end pipeline that takes a WorldWideView plugin from idea to npm-published and cleaned up. Conducts an isolated worktree + scaffold, GSD research/plan/build/UAT, dual-repo PRs (plugin + seeder), npm publish, and teardown, with user gates only at plan approval, UAT sign-off, npm publish, and teardown. Use when the user says "/plugin-new", "new plugin", "build a plugin end to end", "ship a plugin", or "run the plugin pipeline".
---
# Plugin Pipeline (`/plugin-new`)
A **conductor** skill: it runs the staged pipeline in the main conversation and **delegates** the real work to existing agents and GSD. It does not reimplement research, planning, execution, or worktree mechanics; it sequences them and enforces the gates.
## Principles
- **Don't reinvent GSD.** Research, spec, plan, execute, verify, and UAT are GSD phases (running in the plugin's own isolated `.planning`). The pipeline adds only the plugin-specific bookends: isolated-worktree scaffolding at the front, and dual-repo PR + npm publish + teardown at the back.
- **Compose existing agents.** Reuse `plugin-researcher`, `plugin-implementer`, `worktree-manager`, `branch-finisher`, `branch-cleanup`. Never write logic these already provide.
- **Gate only the irreversible.** Autonomous within a stage; stop for the user only at the 4 gates below.
- **One isolated worktree per plugin**, with a project-per-plugin GSD project in its own real `.planning` (per the per-worktree isolation model). This keeps each plugin's planning private and teardown clean.
- **Stop on red.** A failed stage halts the pipeline with a clear report. Never continue past a failed gate. Resume with `/gsd:resume-work`.
## Gates (the only stops)
| Stage | Gate | Why |
|---|---|---|
| 2 | Plan approval | Don't build the wrong thing |
| 4 | UAT sign-off | Human judges the live result |
| 6 | npm publish | Public, irreversible |
| 7 | Teardown | Destroys worktree + Docker volume |
---
## Stage 0: Kickoff
1. Read the plugin name from the invocation (kebab-case). If missing, ask for it.
2. Validate: name is kebab-case and not already present in `local-plugins/` or on npm. If taken, stop and ask.
3. Capture the one-line idea (ask if not given).
## Stage 1: Isolated worktree + scaffold
1. Delegate worktree creation to the **worktree-manager** agent: a sibling worktree `feat/plugin-<name>` based off `origin/main`. (Delegating keeps the verbose git-wt/hook output out of the main session.)
2. Bootstrap it with the **worktree-bootstrap** skill, ensuring its `.planning` is a REAL isolated directory (Step 5 of that skill), NOT a junction. If git-wt's `link_planning` hook left a junction, convert it to a real dir before proceeding (see worktree-bootstrap Step 5).
3. Scaffold the plugin skeleton (`WorldPlugin` entry + seeder stub + `package.json`) using the **worldwideview-plugin-creation** skill / **plugin-implementer** agent.
4. **Label Docker resources** for this plugin so teardown is exact: when the seeder/data-engine containers and volumes are created for this plugin, tag them with `wwv-plugin=<name>` (compose label or naming convention `wwv-<name>-*`). Record the labels in the worktree's `.planning/WORKSPACE.md`.
5. Verify the scaffold builds clean. If it errors, STOP and report (do not limp forward).
- **Gate:** none.
## Stage 2: Idea -> research -> plan (GSD, project-per-plugin)
1. In the worktree, run `/gsd:new-project` scoped to this plugin (its own GSD project in the isolated `.planning`).
2. Run the **plugin-researcher** agent: find the data API, rate limits, auth, endpoints, update cadence; produce an implementation-ready research brief.
3. Run GSD `discuss-phase` then `plan-phase` to produce the phase plan, weaving in the research.
- **Gate: user approves the plan** (GSD plan-review). Do not proceed without it.
## Stage 3: Autonomous build + e2e test (handoff)
1. Run GSD `execute-phase`, which drives the **plugin-implementer** agent to build per the approved plan.
2. Test end to end: `pnpm dev` (frontend), Docker (data engine + Redis + seeder), and Playwright (**playwright-testing** skill) for on-globe rendering.
3. Loop on failures via GSD (`tdd-loop` / `code-review` / `autofix`) until green. No user intervention in this stage.
- **Gate:** none (autonomous until all issues are fixed).
## Stage 4: UAT (servers already up)
1. Spin up the relevant servers (frontend + data engine + seeder) and run GSD `audit-uat`.
2. Present the plugin live on the globe and ask the user to try it.
- **Gate: user UAT sign-off.** If the user reports issues, loop back to Stage 3 (standard GSD fix cycle), then re-run UAT.
## Stage 5: Ship (dual PRs, independent versions)
1. The plugin (`local-plugins/`) and seeder (`local-seeders/`) are independent git repos with their own remotes. Bump each repo's semver **independently** (only bump the side that changed) and open a PR for each via the **branch-finisher** agent, once per repo.
2. Wait for BOTH PRs' CI to pass. If either fails, STOP and report (do not publish a half-shipped pair).
- **Gate:** none to open; the pipeline waits for both green.
## Stage 6: npm publish (user-confirmed)
1. After both PRs succeed, ASK the user whether to publish the plugin to npm.
2. On confirm, publish with bundled assets for CDN distribution (reuse the existing publish workflow / the market-tracker v1.0.x package.json bundling config + `syncToPublic`).
- **Gate: user confirms publish.** Never publish without it.
## Stage 7: Teardown (user-confirmed)
1. ASK the user if they are ready to clean up.
2. On confirm, run **`/branch-cleanup`** (it archives the worktree's isolated `.planning` to the shared archive, then removes the worktree via worktree-manager).
3. Tear down this plugin's Docker resources using the `wwv-plugin=<name>` labels from Stage 1 (containers + volumes). Verify nothing for this plugin remains.
- **Gate: user confirms teardown.**
---
## State, resume, and failure handling
- State lives in the worktree's isolated `.planning` (GSD STATE.md + phases). After any interruption, resume with `/gsd:resume-work` without re-paying for completed stages.
- A failed stage stops the pipeline with a clear report. The user fixes and resumes; nothing continues silently past a red gate.
- Dual-PR coordination (Stage 5): if one PR's CI fails, report and pause; do not publish (Stage 6) until both are green.
## Reused components (must exist)
`worktree-manager`, `worktree-bootstrap`, `worldwideview-plugin-creation`, `plugin-researcher`, `plugin-implementer`, `branch-finisher`, `branch-cleanup`, GSD phase skills (`new-project`, `discuss-phase`, `plan-phase`, `execute-phase`, `tdd-loop`, `audit-uat`, `resume-work`), `playwright-testing`, and the npm publish workflow.
-85
View File
@@ -1,85 +0,0 @@
---
name: recall-context
description: >
Invoke when the user asks to "check history", "recall context", "what did we work on",
"load prior context", "what's the background on X", "remind me where we left off", or any
similar request to surface prior session work. Also invoke proactively at the start of a
session continuation when the user references past work that isn't fully described in the
current conversation. Uses observation history, context-mode index, and memory files to
reconstruct relevant prior state.
---
# Recall Context
Reconstruct what happened in prior sessions so you can continue intelligently.
Use all three layers in order — each covers different ground.
---
## Layer 1: Session timeline (already in context)
The `<system-reminder>` injected at conversation start contains a pre-rendered timeline of
recent observations. Scan it now for entries relevant to the current topic and note their IDs.
This is free — no tool call needed.
---
## Layer 2: Observation store (semantic history)
Fetch detailed records for anything relevant. Use the IDs spotted in Layer 1, plus a semantic
search to catch things the timeline summary might have omitted.
**Semantic search** (cast wide first):
```
mcp__mcp-search__observation_search(query: "<topic>", limit: 10)
```
**Fetch by ID** (for specific timeline entries):
```
mcp__mcp-search__get_observations(ids: [ID1, ID2, ...])
```
Run both if you have IDs. Use 2-3 different query phrasings if the topic is broad.
---
## Layer 3: Context-mode index (indexed session content)
Search the context-mode knowledge base for anything that was indexed during prior sessions
(code excerpts, command output, research, etc.):
```
mcp__plugin_context-mode_context-mode__ctx_search(
queries: ["<angle 1>", "<angle 2>", "<angle 3>"]
)
```
Use 3-5 queries covering different aspects of the topic. For example, if the topic is
"StreamProxy fix", try queries like "streamProxy HTTP SSRF", "safeFetch protocol", and
"camera proxy endpoint".
---
## Layer 4: Persistent memory files
Read `C:\Users\silve\.claude\projects\C--dev-wwv\memory\MEMORY.md` for the index, then read
any files whose one-line description is relevant to the topic. Memory files capture user
preferences, project decisions, and recurring feedback that spans sessions.
---
## Synthesize and report
After gathering, present a focused summary:
| Section | Content |
|---|---|
| **Prior work** | What sessions touched this area and what was done |
| **Current state** | What's complete, what's in progress, what's pending |
| **Key decisions** | Architectural choices, root causes identified, patterns established |
| **Known blockers** | Any flags, gotchas, or unresolved issues noted |
Keep it tight. The goal is a mental model to continue from, not an exhaustive replay.
If nothing relevant is found across all three layers, say so explicitly so the user knows
the search was complete.
-129
View File
@@ -1,129 +0,0 @@
---
name: worktree-bootstrap
description: Create and fully bootstrap a WWV git worktree for isolated feature/debug work. Copies env vars, installs deps, generates Prisma client, and verifies the dev server boots before reporting ready.
---
# Worktree Bootstrap
Use this skill when creating a new worktree for isolated debugging or feature work. Eliminates the repeated "missing .env.local" / "Prisma client not generated" failures.
## Steps
### 1. Create the worktree
From `C:\dev\wwv\worldwideview\`:
```powershell
git-wt switch --create <branch-name> --yes
```
This places the worktree at `C:\dev\wwv\worldwideview.<branch-name>\`.
### 2. Copy environment file
`.env` contains shared secrets (AUTH_SECRET, DATABASE_URL, etc.) that are gitignored and not present in new worktrees. Copy it:
```powershell
Copy-Item C:\dev\wwv\worldwideview\.env C:\dev\wwv\worldwideview.<branch-name>\.env
```
Verify required vars are present:
```powershell
Select-String -Path C:\dev\wwv\worldwideview.<branch-name>\.env -Pattern "^AUTH_SECRET=|^DATABASE_URL="
```
If you need worktree-specific overrides (different port, DB name, feature flags), create a `.env.local` alongside it - Next.js loads `.env.local` on top of `.env`.
Both must match. If either is missing, copy from a teammate or the main `.env.local`.
### 3. Install dependencies
```powershell
Set-Location C:\dev\wwv\worldwideview.<branch-name>
pnpm install
```
### 4. Generate Prisma client
```powershell
npx prisma generate
```
### 5. Initialize isolated `.planning`
Each worktree gets its OWN real `.planning` directory, never a junction to the shared root. This keeps phases and STATE private per worktree so simultaneous sessions never bleed into each other.
```powershell
$wt = "C:\dev\wwv\worldwideview.<branch-name>"
New-Item -ItemType Directory -Force -Path "$wt\.planning\phases", "$wt\.planning\debug", "$wt\.planning\surveys" | Out-Null
Copy-Item C:\dev\wwv\.planning\config.json "$wt\.planning\config.json" -Force
```
Then add three files in `$wt\.planning\`:
- `WORKSPACE.md` (manifest: branch, repo, "real isolated dir" note)
- `STATE.md` (fresh GSD state: `milestone: none`, `status: idle`)
- `SHARED-DOCS.md` (pointer to the shared cross-feature docs)
Quickest way is to copy these three from an existing isolated worktree and edit the branch/feature name:
```powershell
Copy-Item C:\dev\wwv\worldwideview.test-sandbox\.planning\WORKSPACE.md, C:\dev\wwv\worldwideview.test-sandbox\.planning\STATE.md, C:\dev\wwv\worldwideview.test-sandbox\.planning\SHARED-DOCS.md "$wt\.planning\"
```
NEVER create `.planning` as a junction. The shared root `C:\dev\wwv\.planning` holds ONLY cross-feature docs (ROADMAP, MILESTONES, research, PROJECT, REQUIREMENTS).
### 6. Wire the shared-docs env var
So GSD skills can find cross-feature docs without duplicating them, add to the worktree's `.env.local`:
```
WWV_SHARED_PLANNING=C:/dev/wwv/.planning
```
### 7. (Optional) Sync with main
If the branch is behind main:
```powershell
git merge main
# or
git rebase main
```
### 8. Verify the dev server boots
```powershell
pnpm dev
```
Watch for the "Ready" line. If it fails, check:
- Docker is running (needed for PostgreSQL): `docker compose up -d`
- No port conflict on 3000: `netstat -ano | Select-String ":3000"`
### 9. Report ready
Once the dev server returns HTTP 200 on `http://localhost:3000`, the worktree is ready for use.
## .planning Isolation (per-worktree)
Each worktree owns a **real, isolated** `.planning` directory (created in Step 5). It is NOT a junction and NOT shared. Phases and `STATE.md` are private to the worktree, so two simultaneous sessions on different features never see each other's phases.
This replaces the older "one `.planning` branch per worktree" model (which relied on `.planning` being a git repo with a SessionStart/post-checkout hook switching its branch). That model could not isolate simultaneous sessions: all worktrees shared one physical `.planning` HEAD, so whichever session ran its hook last won and the others silently read the wrong feature's phases. The per-worktree real-directory model fixes that.
### Cross-feature docs
ROADMAP, MILESTONES, `research/`, PROJECT, REQUIREMENTS stay authoritative at the shared root `C:\dev\wwv\.planning`. Each worktree references them via the `WWV_SHARED_PLANNING` env var (Step 6) instead of duplicating them. See `SHARED-DOCS.md` in any worktree's `.planning`.
### Self-healing
If you find a worktree whose `.planning` is still a junction (created before the migration), the `planning-sync.sh` SessionStart hook converts it into a real isolated directory automatically on the next session open.
## Teardown
When done, use `/branch-cleanup`. It archives this worktree's real `.planning` to the shared archive first, then removes the worktree via the worktree-manager agent. Never `rm -rf` a worktree (it orphans the Docker volume). For a manual teardown:
```powershell
Set-Location C:\dev\wwv\worldwideview
git-wt remove
```
@@ -202,40 +202,7 @@ import { Plane } from "lucide-react";
const iconUrl = createSvgIconUrl(Plane, { color: "#3b82f6" });
```
### 1.5 Property Tag Helpers
Use these SDK helpers to wrap property values so the Intel panel renders them richly. Without the tag prefix, the panel falls back to plain text.
```ts
import { dtProp, urlProp, imageProp, videoProp } from "@worldwideview/wwv-plugin-sdk";
```
| Helper | Tag format stored | Panel renders as |
|---|---|---|
| `dtProp(iso)` | `"datetime:2026-06-01T05:00:00Z"` | Expandable row: local time (collapsed), UTC + relative (expanded) |
| `urlProp(href)` | `"url:https://..."` | Clickable link with external icon |
| `imageProp(src)` | `"image:https://..."` | Inline thumbnail |
| `videoProp(href)` | `"video:https://..."` | "Watch" link with play icon |
All four helpers return `null` for empty, `null`, or `undefined` input — safe to call unconditionally.
**Usage pattern:**
```ts
properties: {
// dates — pass ISO 8601 strings
updated_at: dtProp(item.updated_at ?? null),
// URLs
source_url: urlProp(item.url ?? null),
// images and video
preview: imageProp(item.image_url ?? null),
stream: videoProp(item.video_url ?? null),
// plain values — no helper needed
name: item.name,
count: item.count,
}
```
### 1.6 Registration (3 Paths)
### 1.5 Registration (3 Paths)
**A. Built-in (monorepo package):**
@@ -257,7 +224,7 @@ await pluginManager.registerPlugin(plugin);
**C. Dynamic import (runtime):**
For user-imported GeoJSON layers, call `pluginManager.loadFromManifest(manifest)` directly.
### 1.7 Build Configuration (Core Monorepo Packages Only)
### 1.6 Build Configuration (Core Monorepo Packages Only)
If you decide to manually move a plugin from the local sandbox into the core `packages/` directory, you must manually add it to the build pipeline:
@@ -685,10 +652,6 @@ Categories are **lowercase**: `"aviation"` not `"Aviation"`, `"natural-disaster"
- [ ] **Frontend:** Added path alias in `tsconfig.json`
- [ ] **Frontend:** Registered via `pluginRegistry` + `pluginManager` in `AppShell.tsx`
- [ ] **Frontend:** `renderEntity()` doesn't mix point/billboard properties
- [ ] **Frontend:** Date/time fields wrapped with `dtProp()`
- [ ] **Frontend:** External links wrapped with `urlProp()`
- [ ] **Frontend:** Image URLs wrapped with `imageProp()`
- [ ] **Frontend:** Video/stream URLs wrapped with `videoProp()`
- [ ] **Engine:** Seeder file in `wwv-seeders/src/seeders/<name>.ts` (or private repo)
- [ ] **Engine:** `registerSeeder({ name })` matches frontend plugin `id`
- [ ] **Engine:** Calls `setLiveSnapshot()` with correct plugin ID key
+3 -9
View File
@@ -2,19 +2,13 @@
install_deps = "pnpm install"
[[pre-start]]
copy_env = """powershell -Command "if (Test-Path '../worldwideview/.env') { Copy-Item '../worldwideview/.env' '.env' -Force }; if (Test-Path '../worldwideview/.env.local') { Copy-Item '../worldwideview/.env.local' '.env.local' -Force }""""
copy_env = """powershell -Command "if (Test-Path '../worldwideview/.env') { Copy-Item '../worldwideview/.env' '.env' -Force }""""
[[pre-start]]
copy_local_plugins = """powershell -Command "if (Test-Path '../worldwideview/local-plugins') { robocopy '../worldwideview/local-plugins' 'local-plugins' /E /XD node_modules dist .next coverage /XF '*.log' /NFL /NDL /NJH /NJS | Out-Null }; exit 0""""
copy_local_plugins = """powershell -Command "if (Test-Path '../worldwideview/local-plugins') { robocopy '../worldwideview/local-plugins' 'local-plugins' /E /XD node_modules dist .next coverage /XF '*.log' /NFL /NDL /NJH /NJS | Out-Null }""""
[[pre-start]]
copy_local_seeders = """powershell -Command "if (Test-Path '../worldwideview/local-seeders') { robocopy '../worldwideview/local-seeders' 'local-seeders' /E /XD .pnpm-store node_modules /NFL /NDL /NJH /NJS | Out-Null }; exit 0""""
[[pre-start]]
link_planning = "powershell -File 'C:/dev/wwv/worldwideview/.agents/scripts/link-planning.ps1' '{{ worktree_path }}'"
[[pre-remove]]
unlink_planning = "powershell -File 'C:/dev/wwv/worldwideview/.agents/scripts/unlink-planning.ps1' '{{ worktree_path }}'"
copy_local_seeders = """powershell -Command "if (Test-Path '../worldwideview/local-seeders') { robocopy '../worldwideview/local-seeders' 'local-seeders' /E /XD .pnpm-store node_modules /NFL /NDL /NJH /NJS | Out-Null }""""
[[pre-remove]]
docker_clean = "docker compose down -v"
-25
View File
@@ -35,34 +35,15 @@ NEXT_PUBLIC_BING_MAPS_KEY=
# Supabase Configuration (Optional — cloud-only features)
# Found in your Supabase Project Settings > API
NEXT_PUBLIC_SUPABASE_URL=
# Preferred name for the Supabase anon/publishable key (matches worldwideview-marketplace).
# NEXT_PUBLIC_SUPABASE_ANON_KEY is still accepted as a backwards-compat alias.
NEXT_PUBLIC_SUPABASE_PUBLISHABLE_KEY=
NEXT_PUBLIC_SUPABASE_ANON_KEY=
SUPABASE_SERVICE_ROLE_KEY=
SUPABASE_DATABASE_URL=
# Cross-subdomain cookie domain (ADR-003).
# Local dev with hosts-file trick: .wwv.local
# Production: .worldwideview.dev
# Leave empty when running on plain localhost (cookies cannot share across ports).
NEXT_PUBLIC_WWV_COOKIE_DOMAIN=
# Auth host URL — the public URL of this WorldWideView instance when acting as auth host.
# Used to build emailRedirectTo in signup and OAuth flows.
# Local dev: http://wwv.local:3000
# Production: https://worldwideview.dev
NEXT_PUBLIC_AUTH_HOST_URL=
# OpenSky Network — multi-account credential rotation
# Comma-separated clientId:clientSecret pairs.
# Example: OPENSKY_CREDENTIALS=acct1-id:acct1-secret,acct2-id:acct2-secret
OPENSKY_CREDENTIALS=
# OpenWeatherMap — Weather radar tile overlays (free tier: 1,000 calls/day)
# Sign up at https://openweathermap.org/api and copy your API key.
OPENWEATHERMAP_API_KEY=
# ACLED API — Civil Unrest & Conflict data (free for non-commercial use)
# Register at https://developer.acleddata.com/ to obtain credentials.
# Data is updated weekly (Mondays) with global protest, riot, and violence events.
@@ -108,9 +89,3 @@ WWV_TEARDOWN_DB_ON_EXIT=false
# Allows operator plugin management on the demo instance via x-wwv-admin-secret header.
# Generate any random string. Only set this on the demo deployment.
# WWV_DEMO_ADMIN_SECRET=
# Shared cross-feature GSD planning docs (ROADMAP, MILESTONES, research).
# Each worktree has its OWN isolated .planning; this points at the shared root so
# GSD skills can read cross-feature docs without duplicating them per worktree.
# Set in each worktree's .env.local (path is machine-specific).
# WWV_SHARED_PLANNING=C:/dev/wwv/.planning
-9
View File
@@ -1,9 +0,0 @@
#!/usr/bin/env bash
# post-checkout hook: $1=prev HEAD, $2=new HEAD, $3=1 if branch checkout, 0 if file checkout
# Only sync .planning on branch switches, not file restores.
if [ "$3" = "1" ]; then
SYNC_SCRIPT="C:/dev/wwv/.claude/hooks/planning-sync.sh"
if [ -f "$SYNC_SCRIPT" ]; then
bash "$SYNC_SCRIPT"
fi
fi
-16
View File
@@ -1,26 +1,10 @@
name: CI
# CLI-only changes are covered by cli.yml; skip the full app suite for them.
# Docs/agent-rules/issue-template changes also skip the suite.
on:
push:
branches: [main]
paths-ignore:
- "packages/wwv-cli/**"
- ".agents/**"
- "**/*.md"
- ".github/ISSUE_TEMPLATE/**"
- ".github/PULL_REQUEST_TEMPLATE.md"
- ".github/gc/**"
pull_request:
branches: [main]
paths-ignore:
- "packages/wwv-cli/**"
- ".agents/**"
- "**/*.md"
- ".github/ISSUE_TEMPLATE/**"
- ".github/PULL_REQUEST_TEMPLATE.md"
- ".github/gc/**"
jobs:
# Non-blocking dependency audit. Surfaces high / critical vulnerabilities
-36
View File
@@ -1,36 +0,0 @@
name: CLI
# Scoped to the plugin CLI package only. Heavy app workflows (ci.yml,
# playwright.yml, docker-publish.yml, ...) exclude this path, so a CLI-only
# change runs just this fast build + test instead of the full suite.
on:
push:
branches: [main]
paths:
- "packages/wwv-cli/**"
- ".github/workflows/cli.yml"
pull_request:
branches: [main]
paths:
- "packages/wwv-cli/**"
- ".github/workflows/cli.yml"
permissions:
contents: read
jobs:
build-and-test:
name: Build & Test CLI
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v6
- uses: pnpm/action-setup@v6
- uses: actions/setup-node@v6
with:
node-version: 22
cache: pnpm
- run: pnpm install --frozen-lockfile
# `build` runs tsc — this also type-checks the CLI source.
- run: pnpm --filter @worldwideview/wwv-cli build
# Integration tests spawn the built CLI and assert the scaffolded tree.
- run: pnpm --filter @worldwideview/wwv-cli test
-23
View File
@@ -1,33 +1,10 @@
name: Docker Image CI
# Only rebuild the image when files that affect the final container change.
# Doc/agent-rules/test-only changes skip the expensive multi-stage build.
# Skipped workflow runs satisfy GitHub required status checks (neutral pass).
on:
push:
branches: [ "main" ]
paths:
- "src/**"
- "prisma/**"
- "packages/**"
- "scripts/**"
- "Dockerfile"
- "package.json"
- "pnpm-lock.yaml"
- "next.config.ts"
- ".github/workflows/docker-publish.yml"
pull_request:
branches: [ "main" ]
paths:
- "src/**"
- "prisma/**"
- "packages/**"
- "scripts/**"
- "Dockerfile"
- "package.json"
- "pnpm-lock.yaml"
- "next.config.ts"
- ".github/workflows/docker-publish.yml"
env:
REGISTRY: ghcr.io
-23
View File
@@ -1,32 +1,9 @@
name: Playwright Tests
# Only run E2E tests when application code actually changes.
# Doc/agent-rules/config-only changes skip the full 3-browser matrix.
# Skipped workflow runs satisfy GitHub required status checks (neutral pass).
on:
push:
branches: [ main, master ]
paths:
- "src/**"
- "tests/**"
- "prisma/**"
- "public/**"
- "package.json"
- "pnpm-lock.yaml"
- "next.config.ts"
- "playwright.config.ts"
- ".github/workflows/playwright.yml"
pull_request:
branches: [ main, master ]
paths:
- "src/**"
- "tests/**"
- "prisma/**"
- "public/**"
- "package.json"
- "pnpm-lock.yaml"
- "next.config.ts"
- "playwright.config.ts"
- ".github/workflows/playwright.yml"
permissions:
contents: read
jobs:
+36 -102
View File
@@ -3,128 +3,62 @@ name: Test Local Dev Environment
on:
push:
branches: [main]
paths-ignore:
- "packages/wwv-cli/**"
pull_request:
branches: [main]
paths:
- "scripts/setup.mjs"
- ".env.example"
- "package.json"
- ".github/workflows/test-local-dev.yml"
permissions:
contents: read
jobs:
# Runs on all 3 OS — tests the actual pnpm setup command users run
test-setup:
name: pnpm setup (${{ matrix.os }})
runs-on: ${{ matrix.os }}
strategy:
fail-fast: false
matrix:
os: [ubuntu-latest, windows-latest, macos-latest]
steps:
- uses: actions/checkout@v4
- uses: pnpm/action-setup@v4
- uses: actions/setup-node@v4
with:
node-version: 22
cache: pnpm
- name: Install dependencies
run: pnpm install --frozen-lockfile
- name: Run pnpm setup
run: pnpm run setup
- name: Verify .env.local was created (not .env)
shell: bash
run: |
if [ ! -f .env.local ]; then
echo "FAIL: .env.local was not created by pnpm setup"
exit 1
fi
echo "PASS: .env.local exists"
# Regression guard: AUTH_SECRET must NOT be in the git-tracked .env
if [ -f .env ] && grep -qE "^AUTH_SECRET=[a-f0-9]{10}" .env; then
echo "FAIL: AUTH_SECRET was written to .env (git-tracked) instead of .env.local"
exit 1
fi
echo "PASS: .env does not contain a generated AUTH_SECRET"
- name: Verify AUTH_SECRET generated correctly
shell: bash
run: |
if ! grep -qE "^AUTH_SECRET=[a-f0-9]{64}$" .env.local; then
echo "FAIL: AUTH_SECRET missing or wrong format in .env.local"
grep "AUTH_SECRET" .env.local || echo "(AUTH_SECRET line not found)"
exit 1
fi
echo "PASS: AUTH_SECRET is 64-char hex"
- name: Verify ENCRYPTION_MASTER_KEY generated correctly
shell: bash
run: |
if ! grep -qE "^ENCRYPTION_MASTER_KEY=[a-f0-9]{64}$" .env.local; then
echo "FAIL: ENCRYPTION_MASTER_KEY missing or wrong format in .env.local"
exit 1
fi
echo "PASS: ENCRYPTION_MASTER_KEY is 64-char hex"
- name: Verify idempotency (re-run must skip gracefully)
shell: bash
run: |
output=$(pnpm run setup 2>&1)
echo "$output"
if echo "$output" | grep -q "already exists"; then
echo "PASS: second run skipped gracefully"
else
echo "FAIL: re-run did not skip"
exit 1
fi
# Linux-only: full predev + dev server smoke test (requires Docker)
test-dev:
name: pnpm dev smoke test (ubuntu)
name: Test Local Dev Startup
runs-on: ubuntu-latest
env:
NEXT_PUBLIC_WWV_EDITION: local
AUTH_SECRET: ci-placeholder-secret-not-used-in-tests
ENCRYPTION_MASTER_KEY: ci-placeholder-key-not-used-in-tests
steps:
- uses: actions/checkout@v4
- uses: pnpm/action-setup@v4
- uses: actions/setup-node@v4
- uses: actions/checkout@v6
- uses: pnpm/action-setup@v6
- uses: actions/setup-node@v6
with:
node-version: 22
cache: pnpm
- name: Create coolify docker network
# Needed because docker-compose.yml references external coolify network
run: docker network create coolify || true
- name: Install dependencies
run: pnpm install --frozen-lockfile
- name: Run pnpm setup
run: pnpm run setup
- name: Run predev (Docker DB + Prisma + Cesium + plugins)
run: pnpm run predev
- name: Start dev server and wait for ready
- name: Setup environment
run: |
cp .env.example .env
sed -i 's/^ENCRYPTION_MASTER_KEY=.*/ENCRYPTION_MASTER_KEY=ci-placeholder-key-not-used-in-tests/' .env
sed -i 's/^AUTH_SECRET=.*/AUTH_SECRET=ci-placeholder-secret-not-used-in-tests/' .env
- name: Test predev setup
# Tests that Docker DB boots, port mappings work, Prisma pushes schema, and plugins sync
run: pnpm run predev
- name: Test dev server startup
run: |
# Start the dev server in the background
pnpm run dev &
DEV_PID=$!
echo "Waiting for Next.js on :3000 (up to 120s)..."
timeout 120 bash -c 'until curl -sf http://localhost:3000 > /dev/null; do sleep 5; done'
echo "Waiting for Next.js to respond on port 3000..."
# Wait up to 120 seconds for the dev server to return a response
timeout 120 bash -c 'until curl -s http://localhost:3000 > /dev/null; do sleep 5; done'
# Check if the process is still running
if ! kill -0 $DEV_PID 2>/dev/null; then
echo "FAIL: dev server crashed before responding"
echo "Dev server crashed or failed to start within timeout!"
exit 1
fi
echo "PASS: dev server responded on :3000"
echo "Dev server started successfully!"
# Graceful shutdown
kill $DEV_PID
-5
View File
@@ -1,15 +1,10 @@
name: Test Setup Scripts
# CLI-only changes are covered by cli.yml; skip setup-script tests for them.
on:
push:
branches: [ "main" ]
paths-ignore:
- "packages/wwv-cli/**"
pull_request:
branches: [ "main" ]
paths-ignore:
- "packages/wwv-cli/**"
permissions:
contents: read
-32
View File
@@ -23,27 +23,6 @@ node_modules/
/.agents/context/
/.agents/settings.local.json
# GSD tool install (vendored global tool, reinstall via the GSD installer; not committed)
/.agents/get-shit-done/
/.agents/commands/gsd/
/.agents/agents/gsd-*.md
/.agents/hooks/gsd-*
/.agents/worktrees/
/.agents/.gsd-profile
/.agents/gsd-file-manifest.json
/.agents/gsd-install-state.json
/.agents/gsd-migration-journal/
/.agents/scheduled_tasks.lock
/.agents/VERSION
/.agents/package.json
# machine-specific (absolute hook paths); portable version TODO before open-source handoff
/.agents/settings.json
# stray Windows redirect artifact
/nul
# planning (private repo: github.com/silvertakana/wwv-planning)
.planning/
# performance
/performance/
@@ -87,7 +66,6 @@ yarn-error.log*
# env files (can opt-in for committing if needed)
.env
.env.local
.env.test
# vercel
.vercel
@@ -157,13 +135,3 @@ worldwideview.code-workspace
/playwright-report/
/test-results/
/playwright/.auth/
/.playwright-mcp/
# MCP client config -- contains a live API key, never commit
.mcp.json
# Root-level UAT / debug scratch artifacts
/uat-*.png
/step*.png
/mcp-test-*.png
/api-keys-response.json
-130
View File
@@ -1,130 +0,0 @@
# Roadmap: Full MCP Support (v1.2)
> **Planning home:** This is the isolated planning for the **v1.2 Full MCP Support** milestone,
> living in the `feat/mcp-support` worktree of `worldwideview`. It was split out of the shared
> root `C:\dev\wwv\.planning\ROADMAP.md` on 2026-05-29 (that file now holds v1.1 only).
> Combined pre-split original: `C:\dev\wwv\.planning\archive\pre-split-2026-05-29\ROADMAP.combined.md`.
>
> **QA ID note:** the v1.2 quality requirements were renamed `QA-01/02/03` -> `MCP-QA-01/02/03`
> during the split, to resolve a collision with v1.1's own `QA-01/02/03`.
## Milestone
- v1.2 **Full MCP Support** - Phases 16-21 (greenfield; no v1.2 code exists yet as of 2026-05-29)
**Primary repo:** `worldwideview` (this worktree).
**Cross-repo dependencies (declared, not separate planning roots):**
- **Phase 20** delegates entity queries to `wwv-data-engine` REST endpoints.
- **Phase 21** depends on `wwv-data-engine` REST (plugin tool delegation), the `@worldwideview/wwv-plugin-sdk` package (manifest `mcpTools` extension), and the existing `worldwideview-marketplace` JWT install bridge (for `list_plugins` / `install_plugin`).
## Phases
- [ ] **Phase 16: API Key Auth Foundation** - generic transport-agnostic API keys: UserApiKey DB model, bcrypt prefix-hash storage, reusable `authenticateApiKey()` middleware, "Keys & Access" modal key CRUD + reveal-once UI
- [ ] **Phase 17: MCP Server Foundation** - raw `@modelcontextprotocol/sdk` Streamable HTTP endpoint in a Next 16 route handler, Bearer auth gate, stateless operation, Coolify/Nginx runbook, + "Connect your agent" helper (per-client config + prompt template)
- [ ] **Phase 18: Globe State Sync + Sessions** - useGlobeStatePush hook, heartbeat, globe://state and globe://sessions resources
- [ ] **Phase 19: Globe Command Bridge** - WebSocket relay, useGlobeCommandBridge hook, per-session Redis pub/sub routing
- [ ] **Phase 20: Data Query Tools** - search_entities, get_entities_in_region, get_entity_details, get_plugin_data (parallel with Phase 19)
- [ ] **Phase 21: Plugin-Contributed Tools + Quality** - SDK manifest extension, auto-namespace, plugin bridge, dynamic tool list, tests, README
## Phase Details
### Phase 16: API Key Auth Foundation
**Goal**: Users can generate, view, and revoke generic personal API keys; keys are stored securely and a reusable `authenticateApiKey()` middleware is ready for Phase 17 (MCP) and future REST routes
**Depends on**: Nothing (first phase of v1.2)
**Target repo**: worldwideview ("Keys & Access" modal section + Prisma `UserApiKey` model + auth middleware)
**Requirements**: KEY-01, KEY-02, KEY-03, KEY-04, API-01 (middleware portion)
**Architecture note**: keys are **transport-agnostic** (not MCP-only); the auth middleware is reusable; capabilities are a shared service layer wrapped thinly by MCP/REST. The per-client "Connect your agent" helper is **Phase 17** (needs the live `/api/mcp` URL); Phase 16 ships only the key + the "API & MCP Access" section shell.
**Success Criteria** (what must be TRUE):
1. User clicks "Generate API Key" in the "Keys & Access" modal and sees the full combined `wwv_<prefix>.<secret>` token exactly once in a copyable field with a "won't see this again" warning
2. Key is stored as `{prefix}.{bcrypt(secret, cost=12)}`; lookup by prefix; a dummy bcrypt compare runs even when prefix is not found (timing oracle prevention)
3. User can view active keys (name, created, last used) and revoke any -- revoked (hard-deleted) key returns 401 immediately on next request
4. Attempting to generate a 4th key returns a clear error; max 3 enforced in the create route AND the UI
5. `authenticateApiKey(request)` exists as a standalone, transport-neutral helper (importable by any future route), with `lastUsedAt` updated best-effort/throttled on success
**Plans**: 4 plans (4 waves)
- [x] 16-01-PLAN.md — Wave 0 RED test scaffolds (helper + route tests) — COMPLETE 2026-05-29
- [x] 16-02-PLAN.md — UserApiKey schema + db push + timing-safe apiKeyAuth.ts helper
- [x] 16-03-PLAN.md — /api/api-keys CRUD routes (list, create max-3, ownership-scoped revoke)
- [ ] 16-04-PLAN.md — PersonalApiKeysSection UI + Header 'Keys & Access' wiring + semver bump
### Phase 17: MCP Server Foundation
**Goal**: A valid Streamable HTTP MCP endpoint is live, authenticated, stateless, confirmed streaming in production, AND users can connect their AI client in one copy-paste step
**Depends on**: Phase 16
**Target repo**: worldwideview (+ Coolify/Nginx infra config)
**Requirements**: MCP-01, MCP-02, MCP-03, MCP-04, CONNECT-01, CONNECT-02, CONNECT-03
**Success Criteria** (what must be TRUE):
1. Claude Code connects to `/api/mcp` with a valid Bearer key and receives a handshake with `protocolVersion`, `serverInfo`, and the full tools/resources list
2. An unauthenticated request returns HTTP 401 with a JSON-RPC 2.0 error shape
3. No `McpServer` instance is cached between requests -- each request creates a fresh one (stateless confirmed)
4. The Coolify Nginx config includes `X-Accel-Buffering: no` and extended `proxy_read_timeout`; a 60-second streaming connection holds open without disconnect
5. The "API & MCP Access" section renders a per-client connection helper (Claude Desktop / Cursor / VS Code `mcpServers` JSON with Bearer header; generic Manual block; Claude Code CLI deferred as "coming soon") + a copy-paste "prompt for your agent"; the secret never appears in a URL/deep-link query
**Library decision**: raw `@modelcontextprotocol/sdk@^1.29.0` (NOT mcp-handler) + the SDK's `WebStandardStreamableHTTPServerTransport` inside a Next 16 App Router route handler — confirmed against SDK 1.29.0 published types, NO custom server.ts needed in Phase 17 (per `BATCH-DECISIONS-17-21.md` / `17-CONTEXT.md`).
**Plans**: 3 plans (3 waves)
- [x] 17-01-PLAN.md — Wave 0: RED test scaffolds (route auth/edition/stateless/streaming-header tests) + raw-SDK transport runtime spike
- [ ] 17-02-PLAN.md — Wave 1: install SDK + createMcpServer() factory + gated stateless /api/mcp route handler (turns Wave 0 green)
- [ ] 17-03-PLAN.md — Wave 2: ConnectAgentHelper UI (per-client config + prompt) + Coolify/Nginx runbook + semver bump to 2.26.0
### Phase 18: Globe State Sync + Sessions
**Goal**: Claude can read what is on the globe and which browser tabs are open, with accurate live data
**Depends on**: Phase 17
**Target repo**: worldwideview (+ Redis)
**Requirements**: RSRC-01, RSRC-02, RSRC-03, RSRC-04
**Success Criteria** (what must be TRUE):
1. `globe://state/{sessionId}` returns the current viewport (lat/lng/altitude/heading/pitch), active layer list, selected entity, and timeline position
2. `globe://sessions` lists all open browser tabs with sessionId, viewport summary, active layers, and last-seen timestamp
3. A closed browser tab drops from `globe://sessions` within 30 seconds (heartbeat TTL expiry)
4. A viewport change in the browser is reflected in `globe://state` within 1 second (500ms debounce + network round trip)
**Plans**: TBD
### Phase 19: Globe Command Bridge
**Goal**: Claude can send pan, zoom, layer, and timeline commands that physically execute in the correct open browser tab
**Depends on**: Phase 18 (needs sessionIds from heartbeat registry)
**Target repo**: worldwideview (+ Redis)
**Requirements**: CTRL-01, CTRL-02, CTRL-03, CTRL-04, CTRL-05, CTRL-06
**Note**: Phases 19 and 20 can be executed in parallel once Phase 17 is complete
**Success Criteria** (what must be TRUE):
1. Calling `pan_globe({ sessionId, lat: 35.68, lng: 139.69 })` causes the CesiumJS camera to fly to Tokyo in the targeted browser tab
2. `toggle_layer` visibly enables or disables a plugin layer in the globe
3. With 2 tabs open, Claude reads `globe://sessions`, picks the contextually correct tab, and only that tab receives the command
4. Opening 10 concurrent MCP connections does not create 10 Redis subscriptions -- singleton subscriber confirmed
5. A Redis restart does not permanently break command delivery; the subscriber reconnects and commands resume within 5 seconds
**Plans**: TBD
### Phase 20: Data Query Tools
**Goal**: Claude can query live entity data across all active plugins without requiring an open browser tab
**Depends on**: Phase 17 (stateless reads from data engine REST; no browser session required)
**Target repo**: worldwideview MCP route + **wwv-data-engine** (REST query delegation) -- cross-repo
**Requirements**: QUERY-01, QUERY-02, QUERY-03, QUERY-04
**Note**: Can be executed in parallel with Phase 19
**Success Criteria** (what must be TRUE):
1. `search_entities({ query: "Singapore" })` returns entities from active plugins near Singapore with id, name, lat/lng, plugin
2. `get_entities_in_region` with a bounding box returns all entities within it, respecting the optional `pluginId` filter
3. `get_entity_details` returns the full properties object for a specific entity without error
4. `get_plugin_data` returns the current snapshot gracefully even when the plugin has no active data (empty array, not a crash)
**Plans**: TBD
### Phase 21: Plugin-Contributed Tools + Quality
**Goal**: Installed plugins expose their own namespaced MCP tools; the tool list updates dynamically; the feature is documented and tested
**Depends on**: Phases 19 and 20
**Target repo**: worldwideview MCP route + **`@worldwideview/wwv-plugin-sdk`** (manifest extension) + **wwv-data-engine** (plugin tool delegation) + **worldwideview-marketplace** JWT bridge (`list_plugins` / `install_plugin`) -- cross-repo, 3 external deps
**Requirements**: PLUG-01, PLUG-02, PLUG-03, PLUG-04, PLG-01, PLG-02, MCP-QA-01, MCP-QA-02, MCP-QA-03
**Success Criteria** (what must be TRUE):
1. A plugin with `mcpTools` in its manifest has those tools registered as `{pluginId}__toolName`; existing plugins without `mcpTools` are unaffected
2. Installing a new plugin fires `notifications/tools/list_changed`; Claude's next `tools/list` call includes the new plugin's tools
3. `list_plugins()` returns marketplace plugins with accurate install status; `install_plugin({ pluginSlug })` triggers the existing JWT bridge flow
4. The README contains a copy-pasteable Claude Code MCP config block that works end-to-end with a generated API key
5. Vitest tests pass: auth rejection (401), tool dispatch to correct handler, malformed input errors, Redis singleton reconnection
## Progress
**Execution Order:**
Phases execute in numeric order, with one parallel branch: 16 -> 17 -> { 18 -> 19, 20 } -> 21
(Phases 19 and 20 may run in parallel once 17 is done; 18 must precede 19.)
| Phase | Milestone | Plans Complete | Status | Completed |
|-------|-----------|----------------|--------|-----------|
| 16. API Key Auth Foundation | v1.2 | 3/4 | In Progress| |
| 17. MCP Server Foundation | v1.2 | 1/3 | In Progress| |
| 18. Globe State Sync + Sessions | v1.2 | 0/? | Not started | - |
| 19. Globe Command Bridge | v1.2 | 0/? | Not started | - |
| 20. Data Query Tools | v1.2 | 0/? | Not started | - |
| 21. Plugin-Contributed Tools + Quality | v1.2 | 0/? | Not started | - |
-216
View File
@@ -1,216 +0,0 @@
---
gsd_state_version: 1.0
milestone: v1.2
milestone_name: Full MCP Support
status: active
last_updated: "2026-05-30T09:54:58.389Z"
progress:
total_phases: 7
completed_phases: 3
total_plans: 23
completed_plans: 16
percent: 43
---
# Project State
## Project Reference
Isolated planning workspace for the **v1.2 Full MCP Support** milestone (Phases 16-21).
Split out of the shared root `C:\dev\wwv\.planning` on 2026-05-29 because v1.1 (marketplace
auth gate) and v1.2 (MCP server) were tangled in one ROADMAP/REQUIREMENTS with no shared
phases or dependency.
Primary repo: this worktree (`worldwideview`, branch `feat/mcp-support`).
Cross-repo deps declared in ROADMAP.md: Phase 20 -> wwv-data-engine REST; Phase 21 ->
wwv-data-engine + `@worldwideview/wwv-plugin-sdk` + marketplace JWT bridge.
Cross-feature docs (MILESTONES.md, the plugin-pipeline backlogs) stay at the shared root and
are NOT duplicated here.
## Current Position
**Phase 16:** code complete (5f3fe69 UI, 90dfb68 tsc fix). UAT run on :3005 (Playwright) passed steps
1-5 EXCEPT: (a) cosmetic - existing section header reads "Your API Keys" not spec's "Service Keys"
(fix in progress); (b) ANOMALY - after reload, GET /api/api-keys returned [] (all keys vanished). A
debugger agent is investigating whether (b) is a code bug or a shared-dev-DB-across-worktrees artifact,
and fixing (a). Phase 16 NOT marked verified until (b) is resolved.
**Phase 17:** COMPLETE (verified 2026-05-30). Wave 0 (fcb7a5e) + Wave 1 (e0ad832, 0053f1d) + Wave 2
(6f77495). Human-verify PASSED 6/6: token only in Authorization header, copy buttons flip, placeholder
after dismiss, 401 JSON-RPC body, 200 SSE with x-accel-buffering:no + protocolVersion 2025-06-18.
Version 2.26.0.
**Phase 18:** COMPLETE (2026-05-30). Wave 0 + 1 + 2 + 3 all done. 498/498 tests GREEN. Delivered:
ioredis singleton, GlobeStateSnapshot, globeStateStore (ZSET sessions), POST /api/globe/state (dual
auth per R-2), useSessionId + useGlobeStateSync hooks, GlobeView wiring, registerGlobeResources MCP
module (RSRC-02/03/04), route.ts first R-1 edit. Version TBD (bump in final commit).
**Phase 20:** COMPLETE (2026-05-30). All waves done. 498/498 tests GREEN. Delivered: types.ts +
service.ts transport-neutral service layer, REST routes (search + region), registerDataQueryTools
MCP module (TOOL-01/02/03/04), route.ts second R-1 edit. Commit a5b4b62 (combined with Phase 18).
Version 2.27.0.
**Phase 19:** SPLIT (2026-05-30, user-approved) into 19a + 19b. See phases/19-globe-command-bridge/19-SPLIT.md.
- **19a (Globe Command System Tools, poll-based):** COMPLETE (2026-05-30, commit a3c8f12). Command tools
(pan/focus/toggle/timeline) + Redis list queue + GET /api/globe/commands poll route + browser poll hook.
NO boot-path change. Covers CTRL-04 + CTRL-05. Version 2.28.0.
- All 3 waves GREEN. Full suite 573 passed / 0 failed.
- code-reviewer + security-reviewer DONE: no Critical/High. Security PASS on hard invariants (userId only
from auth, gate ordering rate-limit -> isDemo 403 -> auth, per-user/tenant queue namespacing, no secret in
URL, double-validation). Review findings folded in: timeWindow/currentTime validation, atomic enqueue
(multi.rpush.expire), UUID sessionId 400 guard, dedicated globeCommandsLimiter (120/60s), in-flight poll
guard, focusEntity entityId-only -> console.warn (no 0,0 fly), z.enum timeWindow.
- Committed: 15 files (excluded pnpm-lock.yaml, .planning, debug PNGs per constraints).
- **Phase 21 readiness (IMPORTANT — two conflicting plan generations on disk):**
- AUTHORITATIVE: 21-REPLAN.md + 21-DECISIONS.md (May 30 02:16) — PLUG-03 via streamUrl delegation,
worldwideview-only, NO engine endpoint, NO marketplace. 4 decisions LOCKED.
- STALE: 21-RESEARCH.md + 21-01..04-PLAN.md (May 29 22:31-23:28) — route PLUG-03 through a NEW
wwv-data-engine REST endpoint + marketplace JWT bridge. DO NOT EXECUTE these wave plans; they predate
the re-plan. Archive them and regenerate 21-01..N against 21-REPLAN.md before executing.
- OPEN DECISION (the one heavy item): InstalledPlugin model is keyed by tenantId (nullable), NOT userId.
The MCP auth chain produces userId. Wave 2 ("tools/list from user's active installed plugins") needs a
resolved userId -> tenantId -> InstalledPlugin path. Stale research's `findMany({ where: { userId } })`
will NOT compile. Decide enumeration strategy (likely tenantId null in local edition = single-tenant)
before planning Wave 2.
- **19b (WebSocket push transport):** DEFERRED out of milestone. server.ts + WS + Docker entrypoint + Nginx
+ pub/sub subscriber. Covers CTRL-01/02/03/06. Basis = original 19-01..19-05 plans. Not executed unless
user later requests. The risky boot-path change is parked here.
**Phase 21:** COMPLETE (2026-05-30). All 5 waves done. PR #215 open on feat/mcp-support. Re-planned as v3 frontend-relay (21-REPLAN.md). Delivered: McpToolDeclaration SDK type, mcpTools/mcpCapabilities manifest fields, executeMcpTool hook, per-session Redis catalog, catalog POST route, useMcpCatalogPublisher, composePluginToolsList, mcpRelay (enqueue/wait/post/drain), /api/mcp/invocations GET, /api/mcp/results POST, useMcpRelayBridge, namespaced-tool dispatch in route.ts, README MCP config block. Version 2.29.x committed.
**Phase 19b:** COMPLETE (2026-05-30, commit 7fd1d7e, version 2.30.0). SSE push transport replacing the 1500ms poll loop. Delivered: GET /api/globe/commands/stream (ReadableStream, 200ms drain, 15s keepalive, 16s max, dual-auth, isDemo 403, globeCommandsStreamLimiter 10/60s, cancel() disconnect cleanup), useGlobeCommandBridge rewritten to EventSource with envelope validation and onerror handler. Poll route kept. 654/654 tests GREEN. tsc + lint clean.
- **21-03 (Wave 2, catalog + tools/list):** COMPLETE (2026-05-30, commits 6779921 + cbd48e5 + 23b893c). publishSessionCatalog/readSessionCatalog (Redis SET+EX, {userId}:{sessionId} key, 120s TTL), POST /api/mcp/catalog (dual-auth, UUID guard, size cap, isDemo gate), useMcpCatalogPublisher hook (reads pluginManager manifests, POSTs on mount + 30s interval), composePluginToolsList (system tools + de-duplicated namespaced plugin tools + capability copy), registerPluginTools in route.ts (reads per-session catalog, registers stub handlers). CAT-01..06 + LIST-01..05 GREEN. Zod v4 z.record() fix applied.
- **21-02 (Wave 1, SDK):** COMPLETE (2026-05-30, commit bda235f). McpToolDeclaration (NO execution field), mcpTools? + mcpCapabilities? in PluginManifest, WorldPlugin.executeMcpTool? hook, validateManifest extended with MAN-01..08 rules, getNamespacedTools + validateToolArgs in src/lib/mcp/pluginTools.ts. 32/32 targeted tests GREEN. Pre-existing Wave 0 RED tests for future waves remain red (correct).
- **21-01 (Wave 0, RED):** COMPLETE (2026-05-30, commit 5930bc7). Six RED test files authored:
validateManifest.test.ts (extended + MAN-01..08), pluginTools.test.ts (PT-01..08),
mcpSessionCatalog.test.ts (CAT-01..06), pluginToolsList.test.ts (LIST-01..05),
mcpRelay.test.ts (RELAY-01 + SEC-01..05), pluginToolDispatch.test.ts (MCP-QA-01..04 + SEC-01/06).
tsc: only TS2307 "module not found" (expected). lint: zero errors on new files. vitest: 5 files FAIL
"Failed to resolve import" (correct RED), validateManifest.test.ts 7 RED + 9 GREEN.
**Execution order (per RECONCILIATION R-7):** 17 done -> 18 done -> 20 done -> 19 (PAUSED, awaiting
user discussion on WebSocket + custom server.ts) -> 21 (re-plan pause). route.ts serialization:
18-04 first, then 20-04 appended. Commit checkpoint: a5b4b62 (Phases 18 + 20 combined).
## Phase 17 Plan Map (planned 2026-05-29)
- **17-01 (Wave 0, autonomous):** RED test scaffolds for /api/mcp (401 JSON-RPC, 403 demo gate,
stateless per-request construction, X-Accel-Buffering header) + a raw-SDK transport runtime spike
that proves WebStandardStreamableHTTPServerTransport + McpServer instantiate/connect outside a
custom server (or flags a blocker). Covers MCP-01..04.
- **17-02 (Wave 1, autonomous + blocking-human SDK-legitimacy checkpoint):** `pnpm add
@modelcontextprotocol/sdk@^1.29.0`, `src/lib/mcp/server.ts` createMcpServer() factory (empty
capabilities), and the gated stateless `src/app/api/mcp/route.ts` (isDemo-first 403, Bearer 401
via Phase 16 authenticateApiKey, fresh server+transport per request, X-Accel-Buffering: no). Turns
Wave 0 green. Covers MCP-01..04.
- **17-03 (Wave 2, autonomous + human-verify checkpoint):** `ConnectAgentHelper.tsx` (per-edition
URL + mcpServers JSON with Bearer header for Desktop/Cursor/VS Code + Manual block + agent prompt;
Claude Code CLI deferred as "coming soon"), mounted where the Phase 16 TODO placeholder was;
`.agents/context/server-management.md` Coolify/Nginx streaming runbook (documented, not automated);
semver bump 2.25.0 -> 2.26.0. Covers CONNECT-01/02/03.
## Key Decisions
- **17 transport decision (LOCKED, BATCH-DECISIONS-17-21.md + 17-CONTEXT.md):** raw
`@modelcontextprotocol/sdk@^1.29.0` (NOT mcp-handler) + the SDK's
`WebStandardStreamableHTTPServerTransport` (`@modelcontextprotocol/sdk/server/webStandardStreamableHttp.js`)
inside a Next 16 App Router route handler at `src/app/api/mcp/route.ts`. **Confirmed against the
published SDK 1.29.0 types**: `handleRequest(req: Request): Promise<Response>` is Web-Standard and
documented for Hono/Cloudflare-Workers fetch handlers — identical Request/Response shape to a Next 16
route handler. NO custom `server.ts` is needed in Phase 17 (that is a Phase 19 / WebSocket artifact).
NO BLOCKER. A Wave 0 spike re-confirms at runtime and escalates only if the transport genuinely
cannot run in a route handler.
- **17 stateless (D-17-04):** fresh McpServer + transport per request (sessionIdGenerator: undefined),
never cached at module scope, with a guard-rail comment. No tools/resources registered this phase.
- **17 auth (D-17-03):** reuse Phase 16 `authenticateApiKey()` unchanged; 401 JSON-RPC body
`{ jsonrpc:"2.0", error:{ code:-32600, message:"Unauthorized" }, id:null }`. Demo -> 403, gate FIRST.
- **17 connect helper (D-17-08/09):** in "API & MCP Access"; mcpServers JSON carries the token in the
Authorization HEADER (raw-SDK Streamable HTTP form), never in a URL; Claude Code CLI deferred "coming
soon". Nginx/Coolify buffering documented (D-17-07), not automated.
- **17-03 (Wave 2):** mcpServers JSON shape uses `{ "url": "...", "headers": { "Authorization": "Bearer
<token>" } }` -- raw-SDK Streamable HTTP form (D-17-09). Local URL derived from window.location.origin
at runtime (not hardcoded :3000). CopyField sub-component pattern established for clipboard interactions.
.agents/context/ is gitignored in both worktree and main repo; server-management.md exists on disk as
a local operator runbook only (D-17-07 intent: document not automate).
- **17-02 (Wave 1):** vitest.config.ts clearMocks: true required to fix mock call count isolation
between describe blocks (MCP-04 demo-gate tests inherited stale call counts from MCP-03 without this).
Optional chaining on server?.connect?.() and transport.handleRequest?.() guards against vi.resetAllMocks()
clearing SDK mock implementations in happy-path tests. AuthInfo.token = "" (raw token not re-exposed).
createMcpServer() capabilities: { tools: { listChanged: true } } per RECONCILIATION R-1.
- **17-01 (Wave 0):** @vite-ignore on variable-based import specifier required in transport spike to
prevent Vite's import-analysis from crashing at transform time on absent SDK. Runtime try/catch alone
is insufficient because Vite resolves import() specifiers statically before test execution.
SDK vi.mock factories (no real module) work correctly for route.test.ts stateless-construction tests.
- **16-01 (Wave 0):** Three RED test files lock api-key contracts (token format, timing oracle T-16-01,
max-3, ownership-scoped delete) before any implementation. timing-oracle test tagged `// [slow]`.
- **16-02 (Wave 1):** UserApiKey Prisma model + apiKeyAuth.ts helper. DUMMY_HASH pre-baked literal
eliminates cold-start timing oracle gap; async bcrypt compare always runs on miss (T-16-01). All
helper tests GREEN.
- **16-03 (Wave 2):** isDemo gate runs FIRST (before auth()). deleteMany ownership-scoped delete
eliminates TOCTOU (T-16-06). fullToken returned once in POST 201; GET omits hashedSecret (T-16-07).
- **16-04 (Wave 3):** PersonalApiKeysSection UI shipped in the renamed "Keys & Access" modal; version
bumped to 2.25.0; Phase 17 connect-helper placeholder comment left at ~L347.
- v1.2 QA requirement IDs renamed `QA-01/02/03` -> `MCP-QA-01/02/03` to resolve a v1.1 collision.
- One primary worktree for the whole milestone; data-engine / SDK / marketplace are declared cross-repo
dependencies, not separate planning roots.
- **Scope expansion (2026-05-29, user-approved):** keys are generic transport-agnostic API keys (req
API-01); one reusable `authenticateApiKey()` middleware; capabilities as a shared service layer
wrapped by MCP + future REST (`/api/v1/*`, Phases 20/21). CLI deferred to a future milestone.
## Milestone Audit (2026-05-30)
v1.2 audited: status **tech_debt** (no critical blockers). Report: `.planning/v1.2-MILESTONE-AUDIT.md`.
Requirements: 31 Complete, 2 Partial (CTRL-02 entityId-only focus, MCP-QA-03 reconnect test),
2 Deferred (PLG-01/PLG-02 -- marketplace descoped to v1.3 per 21-REPLAN). Integration: 47/52,
all 4 browser hooks mounted, 9/12 E2E flows complete (2 broken = the deferred PLG tools).
**Tech debt closed (commit dc1e802, version 2.30.1, pushed):** mcpRelay enqueue+postResult made
atomic (multi/exec) + redis.ts RedisMultiChain.set(); v1 entity routes got dual-auth + isDemo 403
gate + NaN limit guard + global-read intent comment; useGlobeStateSync immediate initial push
(ran-once ref) closes the session-registration race; useMcpCatalogPublisher wired to
getNamespacedTools (DRY); deleted dead pluginToolsList.ts. Code + security review: no Critical/High;
all review warnings folded in. 643/643 tests, tsc clean, lint clean.
REQUIREMENTS.md traceability table + checkboxes updated to actual status. ROADMAP.md has a
"Deferred to v1.3 (Backlog)" section for PLG-01/02 (999.1/999.2).
**Gap-closure 18-05 (2026-05-30, commit ee16ec0):** POST /api/globe/state isDemo gate added.
Returns 403 { error: "Not available in demo edition" } before auth when isDemo is true. Closes
GAP-01 from PR #215 pre-merge audit. vi.hoisted() pattern used for mutable editionMock (deviation
from plan's simpler const suggestion -- Vitest hoisting requires it). 6/6 tests green, tsc clean.
useGlobeStateSync confirmed to handle 403 silently with no code change (fetch catches only thrown
exceptions, not HTTP responses). RSRC-01 fully satisfied.
## Remaining for milestone close
1. Merge PR #215 (feat/mcp-support -> main) once CI is green.
2. After merge: run milestone lifecycle (`/gsd:autonomous` no --only, or
`/gsd:complete-milestone v1.2` then `/gsd:cleanup`) to archive + tag + clean up the worktree.
## Blockers
None. (Phase 17 transport-in-route-handler question resolved: confirmed viable, no custom server.ts.)
@@ -1,94 +0,0 @@
---
phase: 16-api-key-auth-foundation
plan: "03"
subsystem: api-key-auth
tags: [api-keys, crud-routes, auth, ownership-scoping, reveal-once]
dependency_graph:
requires: ["16-01", "16-02"]
provides: ["16-04"]
affects: []
tech_stack:
added: []
patterns:
- "isDemo gate before session check on all handlers"
- "deleteMany for atomic ownership-scoped hard delete (BOLA prevention)"
- "P2002 prefix-collision retry on create"
- "reveal-once fullToken in 201 body only"
key_files:
created:
- src/app/api/api-keys/route.ts
- src/app/api/api-keys/[id]/route.ts
modified: []
decisions:
- "userId always taken from session.user.id, never from request body (T-16-09)"
- "isDemo gate runs FIRST before auth() in all three handlers (Pitfall 6 - demo has no DB row)"
- "createKeyWithRetry helper isolates P2002 catch so outer handler catch-all remains clean"
- "deleteMany (not delete) used to avoid Prisma NotFound throw on foreign key - returns count 0 instead (TOCTOU elimination)"
metrics:
duration: "~12 minutes"
completed_date: "2026-05-29"
tasks_completed: 2
tasks_total: 2
files_created: 2
files_modified: 0
---
# Phase 16 Plan 03: API Key CRUD Route Handlers Summary
## One-liner
GET/POST/DELETE route handlers for `/api/api-keys` with ownership-scoped revoke, max-3 enforcement, and reveal-once fullToken in the 201 response.
## What Was Built
Three HTTP handlers completing the Wave 2 server surface:
**`src/app/api/api-keys/route.ts`** (GET + POST)
- GET: lists the session user's keys selecting only `{ id, name, prefix, createdAt, lastUsedAt }` - `hashedSecret` never included in response
- POST: enforces max-3 (422 `max_keys_reached`), calls `generateApiKey()`, creates the row via Prisma, handles P2002 prefix collision with a single retry, returns 201 `{ key: { id, name, createdAt, fullToken } }` - fullToken returned exactly once here, never again
**`src/app/api/api-keys/[id]/route.ts`** (DELETE)
- `deleteMany({ where: { id, userId } })` - atomic ownership scoping; a foreign or missing key yields `count 0` -> 404, never a 500 or accidental deletion of another user's key
Both files apply the isDemo gate (403) before `auth()` on every handler, per Pitfall 6 (the demo virtual admin has no DB row).
## Tasks Completed
| Task | Name | Commit | Files |
|------|------|--------|-------|
| 1 | Implement GET + POST route.ts | 476befe | src/app/api/api-keys/route.ts |
| 2 | Implement DELETE [id]/route.ts | 4e6ff07 | src/app/api/api-keys/[id]/route.ts |
## Test Results
- `pnpm test -- --run src/app/api/api-keys/route.test.ts`: 6/6 GREEN
- `pnpm test -- --run "src/app/api/api-keys/[id]/route.test.ts"`: 4/4 GREEN
- Full suite `pnpm test -- --run`: 403/403 GREEN (56 test files, no regressions)
## Security Invariants Verified
| Threat ID | Disposition | Verification |
|-----------|-------------|--------------|
| T-16-06 (BOLA - foreign key delete) | mitigated | deleteMany where clause tested in [id]/route.test.ts "ownership enforcement" test |
| T-16-07 (hashedSecret/fullToken in list) | mitigated | "response never contains hashedSecret field" test asserts this; GET select projection omits it |
| T-16-08 (demo gate) | mitigated | isDemo check is first line of all three handlers |
| T-16-09 (userId from body) | mitigated | userId always `session.user.id`, never read from request body |
## Deviations from Plan
None - plan executed exactly as written.
## Known Stubs
None.
## Threat Flags
None - no new network surface beyond the planned routes.
## Self-Check: PASSED
- `src/app/api/api-keys/route.ts` exists: FOUND
- `src/app/api/api-keys/[id]/route.ts` exists: FOUND
- Commit 476befe exists: FOUND
- Commit 4e6ff07 exists: FOUND
@@ -1,99 +0,0 @@
---
phase: 18-globe-state-sync
plan: "05"
subsystem: api-route-security
tags: [security, demo-gate, gap-closure, rsrc-01]
dependency_graph:
requires: []
provides: [isDemo-gate-on-globe-state-post]
affects: [src/app/api/globe/state/route.ts, src/app/api/globe/state/route.test.ts]
tech_stack:
added: []
patterns: [vi.hoisted-mutable-mock, isDemo-gate-before-auth]
key_files:
created: []
modified:
- src/app/api/globe/state/route.ts
- src/app/api/globe/state/route.test.ts
decisions:
- "vi.hoisted() required for editionMock because vi.mock factories are hoisted above const declarations"
- "isDemo gate placed after rate limiter and before userId resolution (matching canonical api-keys pattern)"
- "useGlobeStateSync requires no code change: fetch only catches thrown exceptions, not HTTP 4xx responses"
metrics:
duration: "8 minutes"
completed: "2026-05-30"
tasks_completed: 2
tasks_total: 2
---
# Phase 18 Plan 05: isDemo Gate for POST /api/globe/state Summary
**One-liner:** Added isDemo edition gate (403 before auth) to POST /api/globe/state, closing GAP-01 identified in the PR #215 pre-merge audit.
## Tasks Completed
| Task | Name | Commit | Files |
|------|------|--------|-------|
| 1 | Add isDemo gate + test | ee16ec0 | route.ts, route.test.ts |
| 2 | Verify useGlobeStateSync handles 403 gracefully | (no commit - read-only verification) | useGlobeStateSync.ts |
## What Was Built
**GAP-01 closed:** `POST /api/globe/state` previously had no demo-edition gate, allowing demo users to write globe state to Redis. The fix:
1. `route.ts`: Added `import { isDemo } from "@/core/edition"` and an early-return guard immediately after the rate limiter check and before the `userId` resolution block. Returns `{ error: "Not available in demo edition" }` with HTTP 403.
2. `route.test.ts`: Added `editionMock` via `vi.hoisted()` (required because `vi.mock` factories are hoisted above normal `const` declarations), a `vi.mock("@/core/edition", () => editionMock)` call, reset `editionMock.isDemo = false` in the existing `beforeEach`, and a new describe block "POST /api/globe/state -- demo edition gate" with one test asserting 403 + correct body + `writeGlobeState` not called.
**Task 2 - verification only:** `useGlobeStateSync.pushState` wraps `fetch` in a `try/catch` that only catches thrown exceptions (network errors). HTTP 403 responses are awaited and silently discarded because the hook never inspects `response.ok` or `response.status`. No code change was required.
## Deviations from Plan
### Auto-fixed Issues
**1. [Rule 1 - Bug] vi.hoisted() required for editionMock**
- **Found during:** Task 1 (test run)
- **Issue:** The plan suggested `const editionMock = { isDemo: false }` declared at module scope, but Vitest hoists `vi.mock` calls above all `const` declarations. The mock factory referenced `editionMock` before it was initialized, causing `ReferenceError: Cannot access 'editionMock' before initialization`.
- **Fix:** Wrapped the declaration in `vi.hoisted()`: `const editionMock = vi.hoisted(() => ({ isDemo: false }))`. This runs the factory function before hoisting resolves, making the variable available to the `vi.mock` factory.
- **Files modified:** `src/app/api/globe/state/route.test.ts`
- **Commit:** ee16ec0
## Test Results
All 6 tests passed (5 pre-existing + 1 new):
```
POST /api/globe/state (RSRC-01)
- returns 200 and calls writeGlobeState with valid auth + body
- returns 401 and does NOT call writeGlobeState when auth returns null
- returns 400 and does NOT call writeGlobeState when sessionId is missing
- returns 400 when snapshot is missing from body
- still yields 200 when writeGlobeState rejects (fire-and-forget contract)
POST /api/globe/state -- demo edition gate
- returns 403 and does NOT call writeGlobeState when isDemo is true
```
TypeScript strict mode: zero errors (`pnpm tsc --noEmit`).
## Threat Model Coverage
| Threat ID | Mitigation | Status |
|-----------|-----------|--------|
| T-18-05-01 | isDemo gate before auth prevents demo users writing state even with a valid session | MITIGATED |
| T-18-05-02 | 403 returned before writeGlobeState is called; no Redis write occurs on demo edition | MITIGATED |
## Known Stubs
None.
## Threat Flags
None - this plan closes a security gap rather than introducing new surface.
## Self-Check: PASSED
- [x] `src/app/api/globe/state/route.ts` modified - `isDemo` import and guard present
- [x] `src/app/api/globe/state/route.test.ts` modified - new describe block present
- [x] Commit ee16ec0 exists in git log
- [x] 6/6 tests green
- [x] tsc --noEmit: zero errors
+1 -2
View File
@@ -131,7 +131,6 @@ See `.agents/context/INDEX.md` for the full navigation table. Key files:
| `/debugging-coolify` | Troubleshoot deployed apps on Coolify via MCP/SSH | `.agents/workflows/debugging-coolify.md` |
| `/five` | Five Whys root cause analysis | `.agents/workflows/five.md` |
| `/stitch-to-nextjs` | Generate UI with Stitch MCP, port into Next.js | `.agents/workflows/stitch-to-nextjs.md` |
| `/branch-cleanup` | **Post-merge teardown** — commit leftovers, delete plan file, remove worktree via worktree-manager | `.agents/skills/branch-cleanup/SKILL.md` |
---
@@ -154,7 +153,7 @@ See `.agents/context/INDEX.md` for the full navigation table. Key files:
- **Commit format**: Conventional Commits (`feat:`, `fix:`, `refactor:`, `perf:`). Use `/commit` workflow.
- **Required checks**: `pnpm test` and `pnpm build` must pass before merge.
- **Review**: Use `/pr-review` for comprehensive multi-role review.
- **Worktrees**: Use `git-wt switch --create <branch>` and `git-wt remove` (never `rm -rf` a worktree — orphans the PostgreSQL Docker volume). After a PR merges, use `/branch-cleanup` to commit leftovers, delete the plan file, and remove the worktree in one flow.
- **Worktrees**: Use `git-wt switch --create <branch>` and `git-wt remove` (never `rm -rf` a worktree — orphans the PostgreSQL Docker volume).
---
-32
View File
@@ -148,38 +148,6 @@ WorldWideView is distributed across several specialized repositories:
5. **`worldwideview-web`**: Marketing and landing site.
6. **[`szski/wwv-mcp`](https://github.com/szski/wwv-mcp)**: Reference MCP server for driving the globe from an LLM agent (Claude Code, Claude Desktop, Cursor, Cline, …) via the [Agent Bus](docs/agent-bus.md). Contributor-hosted today; may move under the project org.
## MCP (Model Context Protocol)
WorldWideView exposes a built-in MCP endpoint at `/api/mcp`. Any MCP-compatible client (Claude Desktop, Claude Code, Cursor, Cline, etc.) can connect to it directly without a separate sidecar server.
### Authentication
Generate an API key from the `/setup` page (Settings > API Keys). Pass it as a Bearer token in every request.
### Client config
Add the following block to your MCP client configuration. Replace `https://your-host` with the actual origin where WorldWideView is running (e.g. `http://localhost:3000` for local development).
```json
{
"mcpServers": {
"worldwideview": {
"type": "http",
"url": "https://your-host/api/mcp",
"headers": {
"Authorization": "Bearer <your-api-key>"
}
}
}
}
```
### Plugin tools and tools/list
Plugin tools (tools contributed by loaded plugins) appear in `tools/list` only after the browser tab that is running WorldWideView has loaded the relevant plugin and published its catalog. The server snapshots the catalog at the moment each request arrives.
Because the server is stateless and per-request, it cannot push `list_changed` notifications proactively. If you load a new plugin in the browser after your MCP client has already called `tools/list`, call `tools/list` again to pick up the new tools.
## Development & Workflow
- **Branching & Commits:** We strictly enforce [Conventional Commits](https://www.conventionalcommits.org/) (`feat:`, `fix:`, `refactor:`). Every commit should utilize our semantic versioning `[/commit]` workflow.
@@ -1,20 +1,7 @@
# ADR-0003: Shared Identity & Ecosystem Authentication Host
## Status
Accepted *(Phase 2A-D implementation complete 2026-05-25)*
### Implementation notes (Phase 2 closeout)
- `worldwideview/src/lib/supabase/cookieOptions.ts`, `server.ts`, `client.ts` added — mirrors marketplace recipe
- `worldwideview/src/app/login/actions.ts` migrated from `@supabase/supabase-js` to `@supabase/ssr` `createServerClient` so login actually persists a Supabase session cookie
- `getSafeRedirect` updated to allow trusted subdomains via `NEXT_PUBLIC_WWV_COOKIE_DOMAIN`
- `/auth/callback` route added to handle Supabase PKCE email verification flow
- `/signup` page and server action added at the auth host
- `worldwideview-marketplace/src/app/api/install/start/route.ts` — new server-side auth gate for plugin installs
- `InstallButton.handleInstall` retargeted to `/api/install/start` (was building WWV URL client-side with no auth check)
- `requireSupabaseUser` redirect param renamed `?next=` to `?callbackUrl=` to match WWV login page
- `install-redirect/route.ts` updated to use Supabase auth for cloud edition, NextAuth for local/demo
- `httpOnly: true` added to `buildCookieOptions()` in both repos
- NextAuth (Credentials provider) retained for local and demo editions
Proposed *(Pending Review)*
## Date
2026-05-22
@@ -1,88 +0,0 @@
# ADR-0004: Supabase Session Cookies Are httpOnly Across the Ecosystem
## Status
Accepted *(2026-05-25)*
## Date
2026-05-25
## Related
- **Builds on:** ADR-0003 (Shared Identity & Ecosystem Auth Host) — defines the parent-domain cookie scoping (`.worldwideview.dev` prod, `.wwv.local` dev) that this ADR hardens.
---
## Context
ADR-0003 established that a single Supabase session cookie scoped to the parent domain is inherited by every product subdomain (`marketplace.*`, `app.*`, the apex). The cookie carries a base64-encoded access token, refresh token, and the full user record. By default `@supabase/ssr` does NOT set `httpOnly` on this cookie because the browser client (`createBrowserClient`) needs to read it via `document.cookie` to compute `getSession()` synchronously without a network round-trip.
That default is fine for apps that perform their auth reads on the client. It is **not fine for this ecosystem.** Three observations forced the decision:
1. The `worldwideview-marketplace` repo already shipped with `httpOnly: true` in `buildCookieOptions()`. Its install gate (`requireSupabaseUser` in `src/lib/auth/requireSession.ts`) and every API route read the session through the server client (`createServerClient`), so a non-readable cookie costs nothing.
2. The `worldwideview-web` repo originally omitted `httpOnly` and had a single client-side reader: `src/app/hub/layout.tsx` ran `supabase.auth.getSession()` inside a `useEffect`. During audit (2026-05-25), this was discovered as the *only* JS path that needed cookie readability — and it was an anti-pattern (auth gate via `useEffect` introduces a flash-of-unauthenticated-content and is replaceable with a server component).
3. With the cookie readable to JS, an XSS payload on **any** subdomain (`marketplace.*`, `app.*`, the apex itself) can exfiltrate the full session token. Given marketplace renders user-submitted plugin metadata, this is a real attack surface.
The choice was: keep client-side `getSession()` calls and accept the XSS exfiltration risk, or refactor reads to server-side and harden the cookie.
---
## Decision
**All Supabase session cookies in the WorldWideView ecosystem are set with `httpOnly: true`.** This is enforced in each app's `buildCookieOptions()` helper:
- `worldwideview-web/src/lib/supabase/cookieOptions.ts`
- `worldwideview-marketplace/src/lib/supabase/cookieOptions.ts`
- `worldwideview/src/lib/supabase/cookieOptions.ts` (cloud edition only — file is gated by `isCloud`)
Companion hard constraint: **client-side `supabase.auth.getSession()` / `getUser()` / `getClaims()` is forbidden everywhere in the ecosystem.** All auth reads happen server-side via:
- Server components / pages — call `createClient()` from `src/lib/supabase/server.ts`, then `auth.getClaims()` (validates against the auth server, doesn't trust the cookie).
- API routes — same pattern, typically through a wrapper like `requireSupabaseUser(returnTo)`.
- Middleware / proxy — Next.js 16's `proxy.ts` refreshes the session on every matched request and propagates rotated cookies onto the response.
If a future feature requires reading the session on the client, it MUST go through a server-rendered prop or a fetch to an internal API route that performs the read server-side. Adding a new client-side auth read is a regression.
### Implementation details
```ts
// Shared shape across all three apps:
export function buildCookieOptions() {
return {
domain: resolveCookieDomain(process.env.NEXT_PUBLIC_WWV_COOKIE_DOMAIN),
path: '/',
sameSite: 'lax' as const,
secure: true,
httpOnly: true, // ← this ADR
}
}
```
The `proxy.ts` middleware in worldwideview-web (and the equivalent server-client invocations everywhere else) passes `cookieOptions: buildCookieOptions()` into `createServerClient` so token refreshes inherit the flag automatically.
---
## Consequences
**Positive:**
- XSS on any subdomain can no longer read the session token. The auth server still validates each request, so even a forged cookie has no value.
- Auth state is server-rendered, removing the flash-of-unauthenticated-content on protected routes (previously visible at `/hub`).
- Codebase is consistent — three apps, one cookie configuration, one set of read patterns.
**Negative / accepted tradeoffs:**
- Client-side `supabase.auth.*` reads are dead. Any contributor who has worked with `@supabase/supabase-js` will instinctively reach for `getSession()` in a `useEffect`. Code review must catch this.
- Token refresh on stale tabs requires a server round-trip through `proxy.ts` rather than a client-side refresh. Adds ~10-50ms to the first request after expiry; acceptable.
- Browser-side debugging is slightly harder — you can't paste `await supabase.auth.getUser()` into the devtools console. Use server logs or a dedicated `/api/debug/whoami` route instead.
**Verification at adoption (2026-05-25):**
- `worldwideview-web/src/app/hub/layout.tsx` refactored from client useEffect to server component (`getClaims()` + `redirect()`); test rewritten accordingly.
- Browser confirmation: after sign-out + sign-in, `document.cookie` no longer contains `sb-<ref>-auth-token.0/.1`; only the dev-mode `__next_hmr_refresh_hash__` cookie remains visible to JS.
- Marketplace and WWV cloud edition were already compliant; no behavior change for them.
---
## Enforcement
- Code review must flag any import of `createBrowserClient` from `@supabase/ssr` if the call site is in a `"use client"` component and invokes `auth.*` directly.
- Future ESLint rule: ban `createBrowserClient` + `auth.getUser|getSession|getClaims` combinations. (Not implemented yet — follow-up.)
- The marketplace browser client (`worldwideview-marketplace/src/lib/supabase/client.ts`) exists but is unused; consider deleting it in a future cleanup pass.
@@ -1,212 +0,0 @@
# ADR-001 Consolidation Plan: Dual-Issue Path to Decentralized Plugin Auth
## Context
ADR-001 (`docs/architecture/decisions/adr-0001-decentralized-plugin-auth-and-ssrf-mitigation.md`, still marked **Proposed**) defines a decentralized auth model for plugin data streams: Ed25519-signed, 5-minute, audience-bound JWT tickets issued by the Marketplace and verified offline by each Data Engine using JWKS.
The current state is a partial, incoherent implementation across three repos. A recent attempt to land the Data-Engine half landed broken, took production down, and was rolled back. This plan consolidates everything required so the next attempt is **once-and-done** with a feature-flagged dual-issue strategy — the existing HS256 marketplace-session token keeps working untouched while a new Ed25519 data-plane ticket system is built end-to-end and flipped on per-plugin.
**Scope:** all three deployables — Local App (this repo), Marketplace (`marketplace.worldwideview.dev`), Data Engine (`wwv-data-engine-public`).
**Honest framing on "once and done":** this plan delivers **zero-downtime** but is **4–5 sequenced deploys across 3 repos**, not literally one deploy. Compressing the steps is exactly what caused the previous outage. Sequenced + dormant-flagged is the safest path that meets the no-downtime intent.
---
## Current State (what we found)
### ✅ Already in place
- **AES-256-GCM at rest** for credentials — `src/lib/auth/encryption.ts` (PBKDF2 100k, v1 versioning, nonce+salt in DB). Schema `MarketplaceCredential` at `prisma/schema.prisma:92-104`.
- **PKCE initiation** — `src/app/api/marketplace/connect/route.ts` (openid-client v6, S256, `__Host-` cookies).
- **PKCE callback** — `src/app/api/marketplace/callback/route.ts` (state check, encrypted persist). Tests at `src/app/api/marketplace/pkce.spec.ts`.
- **JWT claims type contract** — `packages/wwv-plugin-sdk/src/auth-contracts.ts` (`PluginJwtClaims`, `WebSocketAuthMessage`).
- **SSRF partial defenses** in the JSON proxy — `src/lib/security/ssrf.ts` (post-DNS private-IP block, pinned-IP agent, size/time limits) used by `src/app/api/camera/proxy/route.ts` and `stream/route.ts`.
### ⚠️ Half-finished / risky (must be resolved before any deploy)
1. `src/lib/marketplace/marketplaceToken.ts` — HS256, 4h, hard-coded `aud`. **Acceptable as the marketplace-session bridge token only**; must NOT be used as a data-plane ticket. C1 enforces this via branded types (`MarketplaceSessionToken` vs `PluginTicket`) so the compiler refuses misuse.
2. `src/app/api/marketplace/callback/route.ts:27-29` — currently calls a non-existent `/api/tickets/exchange` endpoint and casts `openid-client` config as `any`. A4 ships the correct endpoint as `/api/oauth/token`; C1 updates the callback URL and removes the cast. Suspected source of the previous deploy break.
3. `src/lib/auth/encryption.ts:9` — silent fallback to a zeroed key if `ENCRYPTION_MASTER_KEY` env var unset. Must throw at startup.
4. `src/app/api/camera/proxy/iframe/route.ts` — bypasses `safeFetch` entirely; raw `fetch` with default redirect-follow and no size cap.
5. No host allowlist on any proxy route (ADR-001D constraint 1) — accepts any HTTPS host.
6. `src/core/data/WsClient.ts:67-69` — sends `subscribe` immediately on `onopen`; no first-message auth, contract `WebSocketAuthMessage` is defined but unused.
### ❌ Missing entirely
- **Ed25519 keypair** + **JWKS endpoint** on the Marketplace.
- **Ticket-issuance endpoint** on the Marketplace (`POST /api/tickets/exchange` → returns 5-min EdDSA JWT for a named audience).
- **JWT verification** on the Data Engine (`@fastify/jwt` + `get-jwks`, audience binding, unknown-`kid` re-fetch, fail-closed cold start).
- **First-message auth handshake** on the Data Engine WebSocket (3s timeout, close code 4003, no re-auth, log redaction).
- **Marketplace-side PKCE token endpoint** that the Local App callback expects.
- **Origin / Host / TLS-state validation** on the Data Engine WS upgrade.
- ADR doc still `Proposed`; never moved to `Accepted`.
---
## Strategy: Dual-Issue, Feature-Flagged, Coordinated Cutover
Two token systems live side by side. **Nothing existing is removed in this phase.**
| Token | Algo | Lifetime | Issued by | Verified by | Purpose |
|---|---|---|---|---|---|
| Marketplace-session (existing) | HS256 | 4h | Marketplace | Marketplace API only | UI login session. **Unchanged.** |
| Plugin-ticket (new) | EdDSA / Ed25519 | 5 min | Marketplace | Each Data Engine via JWKS | Data-plane WebSocket auth. **New, flag-gated.** |
Canonical flag names used throughout this plan:
- **Client (Local App), per-plugin opt-in:** `NEXT_PUBLIC_WWV_TICKET_AUTH_PLUGINS` (comma-separated plugin IDs).
- **Engine (Data Engine), per-plugin enforcement:** `REQUIRE_TICKET_AUTH_PLUGINS` (comma-separated).
- **Engine, global default-deny:** `REQUIRE_TICKET_AUTH` (boolean, default `false`).
- **Engine, Origin allowlist enforcement:** `ENFORCE_ORIGIN_ALLOWLIST` (boolean, default `false`).
- **Local App SSRF allowlist:** `PROXY_HOST_ALLOWLIST` (`*` sentinel, empty/unset = deny, or concrete list).
When all three engine flags are off, Data Engine accepts unauthenticated connections (today's behavior). When `REQUIRE_TICKET_AUTH_PLUGINS` includes a plugin, the client must send `{type:"auth", v:1, token}` first message and engine validates.
This means each repo can be deployed independently in the safe order below; the new code is dormant until the flag is flipped per-plugin.
---
## Workstreams
### A. Marketplace repo (separate codebase — produce a hand-off spec from this plan)
Deliverables the Marketplace team must ship:
1. **Ed25519 keypair generation + storage**, with overlap support (current + next `kid`).
2. `GET /.well-known/jwks.json` — public keys array, `Cache-Control: public, max-age=300`.
3. **Data-plane ticket exchange** — `POST /api/tickets/exchange` — body `{ api_key, audience, plugin_id }`; returns `{ token, expires_in: 300 }`. Claims per ADR-001B: `iss, sub, aud, exp, iat, jti, tier, scope`. Redis-backed `jti` issuance counter for rate-limiting; Postgres fallback per ADR failure-mode table.
4. Library: `jose` for sign + JWKS publication.
### A4. Marketplace API-Key sub-spec (gating workstream — must complete BEFORE step 2)
Pulled out of Workstream A because it is not a one-endpoint change. The Marketplace does not currently have an "API Key" concept — today's PKCE callback stores the OAuth `access_token` directly. A4 invents the API-Key end-to-end:
- New DB table for API Keys (device-bound, manually rotated per ADR-001C).
- `POST /api/oauth/token` — the PKCE token endpoint consumed by `src/app/api/marketplace/callback/route.ts:27`. Renamed from `/api/tickets/exchange` to avoid path collision with A3.
- API-Key issuance and rotation API.
- Rotation UI in the Marketplace (per ADR-001C).
- A4 is treated as its own workstream with its own sub-plan; do not bundle it into the A1–A3 deploy.
### B. Data Engine repo (`wwv-data-engine-public` — hand-off spec)
**SDK delivery mechanism (decide before B starts):** the Data Engine must consume the same `PluginJwtClaims` and `WebSocketAuthMessage` types that live in `packages/wwv-plugin-sdk` here. Pick one:
- **(Recommended)** Publish `wwv-plugin-sdk` to a private npm registry (or GitHub Packages) with a SemVer tag; Data Engine pins `^x.y.z`.
- Vendor a copy in Data Engine and add a CI check that diffs against the upstream SDK on push.
- Git submodule pointing at a tagged commit.
Drift between repos = silent audience-binding mismatches. Pick one before any Workstream B code lands.
Deliverables:
1. `@fastify/jwt` + `get-jwks` integration; LRU JWKS cache 5 min; unknown-`kid` triggers single background re-fetch then reject.
2. **Fail-closed cold start** — if JWKS unreachable on boot, refuse all connections. **Gated by `REQUIRE_TICKET_AUTH || REQUIRE_TICKET_AUTH_PLUGINS !== "" || ENFORCE_ORIGIN_ALLOWLIST`** — when no ticket auth is active (the dormant-deploy state), JWKS unreachable is non-fatal at boot; engine boots, retries fetch in background. Today's engine has no Marketplace boot dependency, and the dormant deploy must preserve that.
3. WS upgrade handler validates `Origin` (allowlist via env), `Host`, TLS state, expected tenant/plugin metadata. **Flag-gated by `ENFORCE_ORIGIN_ALLOWLIST` (default `false`)** so first deploy does not reject existing production clients. Flipped to `true` only after the allowlist has been populated and observed in logs (step 7).
4. **First-message auth state machine** — connection state `AWAITING_AUTH`; 3s timer; reject any non-auth frame; close 4003 on failure; promote to `AUTHENTICATED` on success; no re-auth.
5. Audience binding — reject if `aud !== process.env.ENGINE_ID`.
6. Log redaction — JWT payloads never serialized to logs.
7. Feature flag `REQUIRE_TICKET_AUTH` defaults `false`; production cuts over per-plugin. Separate from `ENFORCE_ORIGIN_ALLOWLIST` so the two protections can be rolled out independently.
### C. Local App (this repo — concrete file changes)
**C1. Cleanup & safety (must land first, no behavior change)**
- `src/lib/auth/encryption.ts:9` — throw on missing **`ENCRYPTION_MASTER_KEY`** (correct var name; not `MASTER_KEY`) instead of zero fallback. **Pre-deploy gate:** verify `ENCRYPTION_MASTER_KEY` is set in every Coolify environment (local-self-host template, cloud, demo) and in `.env.example` before merging this change — otherwise next deploy fails on boot.
- `src/app/api/marketplace/callback/route.ts:29` — replace `config as any` with proper `client.discovery()` once Marketplace publishes its discovery doc; until then, type the config explicitly.
- **Type-level separation instead of runtime guard:** Introduce branded types in `packages/wwv-plugin-sdk/src/auth-contracts.ts`:
- `type MarketplaceSessionToken = string & { __brand: "marketplace-session" }`
- `type PluginTicket = string & { __brand: "plugin-ticket" }`
- `marketplaceToken.ts` signs/returns `MarketplaceSessionToken`; new `ticketClient.ts` returns `PluginTicket`; `WsClient`'s auth message accepts only `PluginTicket`. Compiler enforces the separation — no fragile runtime detection.
- Add doc comment to `marketplaceToken.ts`: "Marketplace UI session only — NOT a data-plane ticket."
- Existing consumers ([src/lib/marketplace/auth.ts](src/lib/marketplace/auth.ts), [install-redirect/route.ts](src/app/api/marketplace/install-redirect/route.ts), [grant-token/route.ts](src/app/api/marketplace/grant-token/route.ts)) continue to receive `MarketplaceSessionToken` — no behavior change.
**C2. SSRF hardening (ADR-001D)**
- Add `PROXY_HOST_ALLOWLIST` env var (comma-separated). Enforce in `src/lib/security/ssrf.ts` before scheme/IP checks. **Default behavior:**
- `PROXY_HOST_ALLOWLIST="*"` → permissive (current behavior), logs a `WARN` per request naming the host so operators can populate the list from observed traffic. **Sentinel-based instead of relying on unset-vs-empty** because Coolify and similar PaaS UIs may send `""` for "unset."
- `PROXY_HOST_ALLOWLIST=""` or unset → deny all (safe default).
- `PROXY_HOST_ALLOWLIST="a.example,b.example"` → only those hosts allowed.
Ship C2 with the var set to `*` in all environments; populate over a week of log observation; flip to a concrete allowlist in step 7c.
- `src/app/api/camera/proxy/iframe/route.ts` — route through `safeFetch`; add 5 MB size cap; reject redirects to non-allowlisted hosts (don't follow by default).
- Document the allowlist procedure in `.agents/context/environment-config.md`.
**C3. Ticket client (new, dormant behind flag)**
- New module `src/lib/auth/ticketClient.ts`: given a `pluginId` and the encrypted API Key from `MarketplaceCredential`, POSTs to Marketplace `/api/tickets/exchange`, caches the 5-min JWT in memory keyed by `(pluginId, audience)`, refreshes at 4.5 min.
- New hook `src/core/data/useTicket.ts` consumed by `WsClient`.
**C4. WebSocket first-message auth (client side)**
- `src/core/data/WsClient.ts:67-69` — when `ticketAuth` enabled for plugin, on `onopen` send `WebSocketAuthMessage` (from `auth-contracts.ts`) *before* any subscribe; queue subscribes until server `ack`.
- Add 3 s client-side auth timeout that closes and reconnects.
**C5. Feature flag**
- `src/core/edition.ts` — new flag `ticketAuthEnabledForPlugin(pluginId): boolean`, sourced from `NEXT_PUBLIC_WWV_TICKET_AUTH_PLUGINS` (comma list) so we can enable one plugin at a time.
**C6. ADR status**
- Move `docs/architecture/decisions/adr-0001-...md:4` from `Proposed` to `Accepted` once Workstreams A/B/C land in their repos but BEFORE any plugin is flag-flipped.
---
## Parallelization: what runs concurrently vs sequentially
**Code writing — fully parallel.** Three agents can work in three repos simultaneously with zero file conflicts:
- Agent 1: Marketplace (`worldwideview-marketplace`, `feat/adr-001-marketplace`) — A1, A2, A3 fixes.
- Agent 2: Data Engine (`wwv-data-engine`, `feat/adr-001-data-engine`) — B1-B5 fixes.
- Agent 3: Local App (`worldwideview.feat-adr-001-local-app`, `feat/adr-001-local-app`) — C1, C2 first; then C3, C4, C5 once contract is confirmed.
The wire contract (POST body shape for `/api/auth/exchange`, JWT claim shape, WS first-message format) is already defined in `packages/wwv-plugin-sdk/src/auth-contracts.ts` and the existing Marketplace/Data Engine code — agents don't need to coordinate on it.
**Workstream A4 (Marketplace PKCE / API Key invention) can also run in parallel** with the other three, since it touches new files (`/api/oauth/token`, new Prisma model) and doesn't conflict with A1-A3. Treat it as a 4th parallel track.
**Deploys — must be sequenced** (this is the part that took down the previous attempt by compressing it). The order in the next section enforces that.
**Bottom line:** parallelize the *writing*, sequence the *shipping*.
## Deploy Order (zero-downtime sequenced choreography)
Steps 1–5 are individually deployable and dormant for the new ticket flow. **Step 1 introduces a hard `ENCRYPTION_MASTER_KEY` requirement** that the step-0 audit must confirm is satisfied — that is the riskiest item in the pre-ticket-cutover phase. Production behavior for ticket auth does not change until step 6.
0. **Pre-flight audit (before any code lands).** The previous failed deploy may have left debris. Audit and reconcile, producing an artifact at `docs/deploys/2026-MM-DD-adr-001-preflight.md` (use today's date) that records:
- Coolify env vars on the Local App, Marketplace, and Data Engine services — list every `JWT_*`, `TICKET_*`, `JWKS_*`, `AUTH_*` var and decide keep/remove.
- `wwv-data-engine-public` `main` branch — confirm the rolled-back state is clean (no half-merged `@fastify/jwt` wiring, no env-gated bypass middleware).
- Any feature flags or dormant routes already in production from the prior attempt.
- Confirm `ENCRYPTION_MASTER_KEY` is set on every environment (gate for C1).
- Confirm decision on SDK delivery mechanism (gate for B).
- Create `docs/feature-flags.md` listing all 5 flags introduced by this plan (`NEXT_PUBLIC_WWV_TICKET_AUTH_PLUGINS`, `REQUIRE_TICKET_AUTH`, `REQUIRE_TICKET_AUTH_PLUGINS`, `ENFORCE_ORIGIN_ALLOWLIST`, `PROXY_HOST_ALLOWLIST`) with: which repo, default value, what each setting means, and a per-environment table (dev/staging/prod) the operator keeps current. Drift between staging and prod is a known paper-cut; the doc is the single source of truth.
- Confirm A4 sub-spec (Marketplace API-Key invention) has its own plan and is on track — A4 must ship before step 2.
1. **Local App** ships C1 + C2 (cleanup + SSRF). Pure hardening, no protocol change. Verify and deploy.
2. **Marketplace** ships A1–A3 (Ed25519 keys, JWKS endpoint, data-plane ticket-exchange). **A4 must already be shipped** as a prerequisite (see Workstream A4 above) — confirmed in step 0. New endpoints only, no existing endpoint changes. Verify JWKS reachable.
3. **Local App** ships C3 + C4 + C5 (ticket client, first-message auth, flag) with flag set to `""` (empty = nothing enabled). Dormant. Deploy.
4. **Data Engine** ships B1–B7 with `REQUIRE_TICKET_AUTH=false` **and** `ENFORCE_ORIGIN_ALLOWLIST=false`. Both new verification paths exist but are dormant. Deploy.
5. **Smoke test in staging** — flip flag for one synthetic plugin only; verify end-to-end ticket → WS auth → stream → 5-min refresh → revocation latency.
6. **Cut over one real plugin at a time** by adding it to `NEXT_PUBLIC_WWV_TICKET_AUTH_PLUGINS` and the engine's `REQUIRE_TICKET_AUTH_PLUGINS`. Monitor each for 24h before adding the next.
7. **Default-deny rollout — split into three independent steps with 24h soak between each** so attribution is clear if anything breaks:
- **7a.** Set `REQUIRE_TICKET_AUTH=true` on the Data Engine. Soak 24h.
- **7b.** Set `ENFORCE_ORIGIN_ALLOWLIST=true` on the Data Engine (after populating the Origin allowlist from observed connections). Soak 24h.
- **7c.** Replace `PROXY_HOST_ALLOWLIST="*"` with the concrete observed list on the Local App. Soak 24h.
Each step is independently revertable by flipping its single flag back.
8. **Accept** the ADR; schedule HS256 marketplace-session token review as a follow-up (out of scope here).
**Re-verification rule (anti-dormant-code rot):** between steps 3 and 5, run the staging smoke test (step 5) **weekly** if the cutover is paused. Dormant feature-flagged code rots — catching it weekly in staging beats finding it broken on cutover day.
Rollback at any step: remove the plugin from the env list. Old path is still wired up because we never removed it.
---
## Critical Files (this repo)
- `src/lib/auth/encryption.ts` — fix silent fallback (C1)
- `src/lib/security/ssrf.ts` — add host allowlist (C2)
- `src/app/api/camera/proxy/iframe/route.ts` — route via safeFetch (C2)
- `src/app/api/marketplace/callback/route.ts` — remove `as any` cast (C1)
- `src/lib/marketplace/marketplaceToken.ts` — guard against data-plane misuse (C1)
- `src/lib/auth/ticketClient.ts` — **new** (C3)
- `src/core/data/WsClient.ts` — first-message auth (C4)
- `src/core/edition.ts` — per-plugin flag accessor (C5)
- `packages/wwv-plugin-sdk/src/auth-contracts.ts` — already has the types, reuse
- `docs/architecture/decisions/adr-0001-...md` — status flip (C6)
---
## Verification
End-to-end (run in staging, gated by `NEXT_PUBLIC_WWV_TICKET_AUTH_PLUGINS=test-synthetic`):
1. `pnpm test` — Vitest suite green including new tests for `ticketClient` and `WsClient` first-message flow.
2. `pnpm build` — no TS errors after removing the `as any` cast.
3. Manual: with engine `REQUIRE_TICKET_AUTH_PLUGINS=test-synthetic`, connect WS without auth message → engine closes with 4003 within 3 s.
4. Manual: valid ticket → stream flows; wait 4.5 min → client transparently refreshes; revoke API Key on Marketplace → next refresh fails, current stream finishes its 5-min window then disconnects (matches ADR revocation table).
5. Stop Marketplace → Data Engine cold-start refuses connections (fail-closed); restart Marketplace → connections resume.
6. SSRF: attempt `POST /api/camera/proxy` to a host not in `PROXY_HOST_ALLOWLIST` → 403. Attempt to a host resolving to 127.0.0.1 → 403.
7. Encryption fallback: unset `ENCRYPTION_MASTER_KEY`, start app → process refuses to start.
Deploy gate: all of the above pass in staging before step 6 of the deploy order touches any production plugin.
@@ -1,264 +0,0 @@
# ADR-001 Workstream A: Marketplace Hardening (A1–A3)
## Context
The Marketplace already ships two stubbed JWT endpoints from a prior attempt:
- `GET /api/auth/jwks` — Ed25519 JWKS publication
- `POST /api/auth/exchange` — issues 5-min audience-bound EdDSA tickets
Both are functional but wrong in production-critical ways: the JWKS cache is set to 24h (mismatched against the 5-min ticket lifetime), the API-key check is a hardcoded string, and the JWT claims (`sub`, `aud`, `tier`, `scope`) are stubs.
ADR-001 (`C:\dev\worldwideview\docs\architecture\decisions\adr-0001-decentralized-plugin-auth-and-ssrf-mitigation.md`) makes this the foundation of decentralized plugin auth: each Data Engine verifies tickets offline via JWKS, with no Marketplace DB roundtrip per frame. A wrong cache TTL or missing `kid` silently breaks every consumer.
This plan delivers **A1 + A2 + A3 only** — the bounded hardening of the two existing endpoints, plus the minimum schema needed to back A3 with real DB lookups. **A4 (PKCE token endpoint + API-Key issuance UI) is deferred** to its own branch/PR per the master plan's explicit instruction: *"Treat it as its own workstream with its own sub-plan; do not bundle it into the A1–A3 deploy."*
A4 also depends on design decisions outside this scope (how a user authenticates to the Marketplace, whether there's a consent screen, how `openid-client` on the Local App expects the token response shaped) that should not be force-fit into a hardening pass.
---
## Decisions (with rationale)
### D1. Separate `User` table (vs. denormalize onto API key)
- Per ADR-001B, `sub` is "the user's unique ID" and `tier` is subscription-scoped. These are identity facts, not credential facts. A user with two devices (two API keys) shares one tier.
- Denormalizing tier onto each `MarketplaceApiKey` row creates multi-row update consistency problems on tier change.
### D2. SHA-256 for `keyHash` (vs. bcrypt per HANDOFF)
- The HANDOFF says "bcrypt of the raw key," but bcrypt is non-deterministic — `WHERE keyHash = bcrypt(input)` is impossible. The exchange endpoint would have to load every row and `bcrypt.compare` each, which is O(n) per request.
- API keys here are **32 bytes of CSPRNG output** (256 bits of entropy). Slow hashing is for low-entropy passwords; for high-entropy secrets, SHA-256 with constant-time compare is the standard (GitHub, AWS, Stripe all use deterministic hashes for API keys).
- No new dependency. Uses `node:crypto` (already a Node built-in).
### D3. Defer A4 entirely
- Master plan: A4 is its own workstream.
- A4 is gated by product decisions (Marketplace auth/login, consent UI) that don't exist yet.
- A1–A3 deliver real value independently — they make the existing JWT issuance pipeline production-correct.
- The Local App's PKCE callback is already broken (calls `/api/tickets/exchange` which doesn't exist); deferring A4 doesn't regress anything.
### D4. Scope derivation from tier (not stored)
- ADR-001B says `scope` is "specific entitlements." If stored separately from tier, the two can drift.
- Derive at JWT issuance: `free → "plugins:read"`, `pro → "plugins:read plugins:write"`, `enterprise → "plugins:read plugins:write plugins:admin"`. Single function, easy to extend.
---
## Implementation
### A1. Fix JWKS cache TTL
**File:** `src/app/api/auth/jwks/route.ts:26`
Change `max-age=86400, stale-while-revalidate=86400` → `max-age=300, stale-while-revalidate=60`.
Rationale: ticket lifetime is 5 min (300s) per ADR-001B. Cache TTL matching the ticket lifetime means a Data Engine can never have stale JWKS during a valid ticket's lifetime. `swr=60` lets engines tolerate brief Marketplace downtime — matches the ADR's "JWKS cache expires (Marketplace down) → grace period" failure mode row.
### A2. Ensure `kid` present in JWK
**File:** `src/app/api/auth/jwks/route.ts`
Currently relies on the env-var JWK having a `kid` field, which the spread `{ ...privateJwk }` carries through. But there's no defensive check — if `MARKETPLACE_JWK_PRIVATE` was generated without `kid`, the endpoint silently publishes a key with no identifier, and `get-jwks` on the Data Engine can't cache it.
Fix: after parsing, throw if `privateJwk.kid` is falsy. Return 500 with a clear error log (`MARKETPLACE_JWK_PRIVATE is missing 'kid' field — regenerate per README`).
Also: ensure `alg: "EdDSA"` is present on the published JWK so JWKS consumers can pre-validate algorithm. The spread carries it through if set; add the same defensive check.
> **Pre-mortem mitigation (TIGER, medium):** this check is fail-closed — if the *current production* `MARKETPLACE_JWK_PRIVATE` lacks `kid`, this deploy turns "JWKS works without kid" into "JWKS 500s every Data Engine fetch." **Before merging**, audit the env var in every environment (local-self-host, cloud, demo):
> ```
> node -e "const j=JSON.parse(process.env.MARKETPLACE_JWK_PRIVATE);console.log('kid:',j.kid,'alg:',j.alg)"
> ```
> If any environment prints `kid: undefined`, regenerate the keypair per the HANDOFF README *before* this code reaches that environment.
### A3. Harden the token exchange endpoint
#### A3a. Schema additions
**File:** `prisma/schema.prisma`
Add two models:
```prisma
model User {
id String @id @default(cuid())
email String @unique
tier String @default("free") // "free" | "pro" | "enterprise"
createdAt DateTime @default(now())
apiKeys MarketplaceApiKey[]
}
model MarketplaceApiKey {
id String @id @default(cuid())
userId String
keyHash String @unique // SHA-256 hex of raw key, used for O(1) lookup
deviceId String?
name String? // user-given label, e.g. "Home Server"
createdAt DateTime @default(now())
lastUsedAt DateTime?
revokedAt DateTime?
user User @relation(fields: [userId], references: [id])
@@index([userId])
}
```
Run `pnpm prisma migrate dev --name add_user_and_marketplace_api_key`.
Note: `MarketplaceApiKey` model is added here even though A4 (issuance) is deferred — A3's DB lookup needs the model to exist. A4 will add the *issuance path*; A3 only consumes the table.
#### A3b. Add a `scopeFor(tier)` helper
**File (new):** `src/lib/auth/tierScope.ts`
```ts
export function scopeFor(tier: string): string {
switch (tier) {
case "enterprise": return "plugins:read plugins:write plugins:admin";
case "pro": return "plugins:read plugins:write";
case "free":
default: return "plugins:read";
}
}
```
Single source of truth; A4 will also use this.
#### A3c. Add a key-hashing helper
**File (new):** `src/lib/auth/apiKeyHash.ts`
```ts
import { createHash, timingSafeEqual } from "node:crypto";
/** Deterministic SHA-256 hex of an API key. Use for DB lookup. */
export function hashApiKey(rawKey: string): string {
return createHash("sha256").update(rawKey, "utf8").digest("hex");
}
/** Constant-time compare of two hex strings. */
export function timingSafeEqualHex(a: string, b: string): boolean {
if (a.length !== b.length) return false;
return timingSafeEqual(Buffer.from(a, "hex"), Buffer.from(b, "hex"));
}
```
#### A3d. Rewrite the exchange endpoint
**File:** `src/app/api/auth/exchange/route.ts`
- Parse body as `{ apiKey, audience?, plugin_id? }`.
- Reject if `apiKey` missing (400).
- Hash the apiKey via `hashApiKey()` and look up `prisma.marketplaceApiKey.findUnique({ where: { keyHash }, include: { user: true } })`.
- If not found, or `revokedAt != null` → 401.
- Update `lastUsedAt = new Date()` (fire-and-forget: `prisma.marketplaceApiKey.update(...).catch(err => console.warn("lastUsedAt update failed:", err.message))` — never await, never let it fail the request).
- JWT claims:
- `iss`: `"https://marketplace.worldwideview.dev"` (unchanged)
- `sub`: `apiKeyRecord.userId` (was: `"user-id-stub"`)
- `aud`: `audience ?? "wwv-data-engine"` (default for backwards-compat per HANDOFF)
- `tier`: `apiKeyRecord.user.tier` (was: `"pro"`)
- `scope`: `scopeFor(apiKeyRecord.user.tier)` (was: `"plugins:read"`)
- `exp`, `iat`, `nbf`, `jti`: unchanged
- Header `kid`: unchanged
- Return `{ token }` as before.
#### A3e. Update tests
**File:** `src/app/api/auth/exchange/route.spec.ts`
- Mock `@/lib/prisma` with `vi.mock`. Provide a fake `prisma.marketplaceApiKey.findUnique` that returns a fixture user/key for one specific hashed key, and `null` otherwise.
- Test: invalid key → 401.
- Test: valid key → 200, JWT contains `sub` matching user id, `aud` matching request body, `tier`/`scope` matching user fixture.
- Test: revoked key (revokedAt set) → 401.
- Test: `aud` defaults to `"wwv-data-engine"` when `audience` omitted from body.
- Keep the existing claim-shape assertions.
**File:** `src/app/api/auth/jwks/route.spec.ts`
- Add: when env JWK has no `kid` → 500.
- Add: response has `Cache-Control` with `max-age=300`.
### Seed data
**File:** `prisma/seed.ts`
Add a test user + API key (only run in dev/test) so the exchange endpoint can be exercised by `curl`:
```ts
// Add to existing seed
const testUser = await prisma.user.upsert({
where: { email: "[email protected]" },
update: {},
create: { email: "[email protected]", tier: "pro" }
});
await prisma.marketplaceApiKey.upsert({
where: { keyHash: hashApiKey("dev-key-do-not-use-in-prod") },
update: {},
create: {
userId: testUser.id,
keyHash: hashApiKey("dev-key-do-not-use-in-prod"),
name: "dev seed key"
}
});
```
Raw key `dev-key-do-not-use-in-prod` is committed intentionally — it only works against a local dev DB seeded with it.
---
## Critical files
| File | Change | Task |
|---|---|---|
| `src/app/api/auth/jwks/route.ts` | Cache header + kid defensive check | A1, A2 |
| `src/app/api/auth/jwks/route.spec.ts` | New tests for kid + cache | A1, A2 |
| `prisma/schema.prisma` | Add User + MarketplaceApiKey models | A3a |
| `src/lib/auth/tierScope.ts` | **NEW** — scope derivation | A3b |
| `src/lib/auth/apiKeyHash.ts` | **NEW** — SHA-256 + constant-time compare | A3c |
| `src/app/api/auth/exchange/route.ts` | Real DB lookup, real claims | A3d |
| `src/app/api/auth/exchange/route.spec.ts` | Prisma-mocked tests | A3e |
| `prisma/seed.ts` | Dev seed user + key | (dev only) |
Existing utilities reused:
- `src/lib/prisma.ts` — singleton Prisma client.
- `jose` (already installed) — JWT signing.
- `node:crypto` — SHA-256, timing-safe compare.
---
## Verification
1. `pnpm prisma migrate dev` — migration applies cleanly to a fresh DB.
2. `pnpm test` — vitest suite green: new + updated exchange tests, jwks tests.
3. `pnpm lint && pnpm build` — no TS errors.
4. Manual end-to-end (dev):
- `pnpm prisma db seed` — seeds dev user + key.
- `curl http://localhost:3000/api/auth/jwks` — returns `{ keys: [{ kty:"OKP", crv:"Ed25519", x, kid, alg:"EdDSA" }] }`, no `d`, `Cache-Control: public, max-age=300, stale-while-revalidate=60`.
- `curl -X POST http://localhost:3000/api/auth/exchange -d '{"apiKey":"dev-key-do-not-use-in-prod","audience":"wwv-aviation-engine","plugin_id":"aviation"}' -H 'Content-Type: application/json'` — returns `{ token }`.
- Decode the JWT at jwt.io: `iss === "https://marketplace.worldwideview.dev"`, `sub` is the seeded user id, `aud === "wwv-aviation-engine"`, `tier === "pro"`, `scope === "plugins:read plugins:write"`, `exp - iat === 300`, header has `kid`.
- Same curl with `apiKey: "valid-key-for-testing"` (the old stub) → 401.
- Same curl with `apiKey` omitted → 400.
---
## Deployment checklist (pre-mortem mitigations)
1. **Pre-deploy env audit** — before this code reaches any environment, run:
```
node -e "const j=JSON.parse(process.env.MARKETPLACE_JWK_PRIVATE);console.log('kid:',j.kid,'alg:',j.alg)"
```
Both must be non-null. If `kid` is missing, regenerate the keypair (the HANDOFF README has the Node script). Do this in every environment (local-self-host, cloud, demo) before merging.
2. **Production migration** — the deploy pipeline must run `pnpm prisma migrate deploy` (not `migrate dev`) on boot before the app starts. Verify this is wired into the Coolify start command or Dockerfile CMD. The new `User` and `MarketplaceApiKey` tables must exist before any request reaches the exchange endpoint.
3. **Empty-table gap** — after A3 deploys and until A4 ships, the exchange endpoint requires a real `MarketplaceApiKey` row to issue a JWT. There are no such rows in production until A4's issuance path lands. This is acceptable because: (a) the Local App's callback is already broken today (`/api/tickets/exchange` 404s), so no real traffic is hitting this endpoint, and (b) the `dev-key-do-not-use-in-prod` seed entry gives a working dev/staging test path through the real code. Document this gap in the PR description so the next person isn't confused by a blank table.
---
## Out of scope (explicit deferrals)
- **A4 — PKCE token endpoint, OAuth authorize page, API-Key issuance UI.** Its own workstream. The `MarketplaceApiKey` model is added in A3 so A4 can plug straight in without a schema migration.
- **Key rotation overlap** (multi-key JWKS array). The JWKS endpoint already returns an array of one key; multi-key rotation is a separate concern.
---
## Risk Mitigations (Pre-Mortem)
**Pre-mortem run:** 2026-05-21 | Mode: quick | Tigers: 3 | Elephants: 1
### Tigers addressed
1. **Production migration path unclear** (medium) — Mitigation: Deployment checklist item 2 above. Plan explicitly requires `prisma migrate deploy` step.
2. **`kid` defensive check fails-closed on deploy** (medium) — Mitigation: Deployment checklist item 1 above. Mandatory pre-deploy env audit.
3. **Fire-and-forget `lastUsedAt` unhandled rejection** (low) — Mitigation: Added `.catch(warn)` to A3d spec.
### Elephant addressed
- **No user creation path until A4** — Mitigation: dev seed provides a working test path; prod gap documented in deployment checklist item 3. Acceptable because the endpoint has zero real traffic today.
### Accepted risks (none — all addressed)
- **Redis-backed `jti` counter / rate limiting.** Per ADR-001 failure-mode table, this is part of issuance hardening — not bundled here.
- **`/api/tickets/exchange` legacy path.** The Local App currently points there but it doesn't exist. Renaming the callback URL is a Local App change (workstream C1), not Marketplace.
@@ -1,173 +0,0 @@
# ADR-001 Production Auth Rollout — Two-Track Strategy
## Context
PR #9 (`wwv-data-engine`) merged + deployed. Auth is bypassed (`WWV_SKIP_WS_AUTH=true`).
`WWV_SKIP_WS_AUTH` lives on the **engine** — flipping it affects every app connecting
to that engine (demo, local, future cloud). The demo has anonymous visitors with no
user session. It needs a **service credential** (machine-to-machine API key) so
`/api/auth/ticket` can issue JWTs on behalf of visitors with no session.
ADR-003 (Supabase identity) is not yet implemented and blocks the full PKCE user
flow. It must not block the demo deployment.
**Strategy:**
- **Track A (demo, immediate):** Seed a service credential into production DBs → flip
auth on → zero user-facing change. No PKCE needed.
- **Track B (local/self-hosted, medium-term):** Implement the PKCE OAuth server on
the marketplace using admin-password auth (no Supabase). When ADR-003 lands,
replace the password check with Supabase session check — same OAuth interface,
no client changes.
---
## Track A — Demo Service Credential (Production)
### Why this works
**`src/core/edition.ts` line 70:** `export const isAuthEnabled: boolean = !isDemo;`
`isAuthEnabled = false` in demo edition → `/api/auth/ticket` skips the session
gate entirely and calls `getTicket()` directly. The admin's "special login page"
uses `WWV_DEMO_ADMIN_SECRET` — a completely separate mechanism unrelated to the
ticket flow. Every anonymous visitor can call `/api/auth/ticket` and get a JWT,
provided `MarketplaceCredential { tenantId: "local" }` exists in the demo DB.
Also confirmed: `ticketAuthEnabledForPlugin()` parses `NEXT_PUBLIC_WWV_TICKET_AUTH_PLUGINS`
as CSV. The current `"\"\""` bug in Coolify makes it parse to no valid IDs → all
plugins skip ticket auth → they connect unauthenticated → `WWV_SKIP_WS_AUTH=true`
lets them in. Fixing the env var + adding the credential + flipping auth is the
complete transition.
### Prerequisites (Coolify → `worldwideview-demo` → Environment)
```
ENCRYPTION_MASTER_KEY=<generate: node -e "console.log(require('crypto').randomBytes(32).toString('base64'))">
MARKETPLACE_URL=https://marketplace.worldwideview.dev
```
Set these before running the seed script. The master key is used to encrypt the
credential at seed time and decrypt it at runtime.
### Step A1 — Write seed script
**File:** `local-scripts/seed-demo-credential.mjs`
One idempotent run does:
1. Connects to marketplace SQLite on the server (SSH + direct file access or
Prisma with `DATABASE_URL=file:<path>`)
- Upserts `User { email: "[email protected]", tier: "pro" }`
- Generates raw API key: `crypto.randomBytes(32).toString('hex')`
- SHA256-hashes it → upserts `MarketplaceApiKey { userId, keyHash, revokedAt: null }`
2. Connects to demo PostgreSQL
- AES-256-GCM encrypts the raw API key using `ENCRYPTION_MASTER_KEY`
(matching the format in `src/lib/auth/encryption.ts`:
`pbkdf2Sync(key, salt, 100000, 32, 'sha256')` → cipher → `base64(ct).base64(tag)`)
- Upserts `marketplace_credentials { tenantId: "local", version:"v1", salt, nonce, ciphertext }`
3. Prints `[seed] ✅ Demo service credential installed`
### Step A2 — Fix env var + rebuild demo
Coolify → `worldwideview-demo` → Environment:
- `NEXT_PUBLIC_WWV_TICKET_AUTH_PLUGINS` → `aviation,maritime` (remove literal quotes)
- Trigger **rebuild** (build-time env var — restart alone doesn't pick it up)
### Step A3 — Verify ticket route
```bash
curl https://<demo-fqdn>/api/auth/ticket?pluginId=aviation
# expect: { "token": "eyJ..." }
```
- 500 → `ENCRYPTION_MASTER_KEY` mismatch or credential not seeded
- 200 with no `token` field → `ticketClient` returned wrong shape
- 401 → `isAuthEnabled` is true (shouldn't happen in demo edition — check `NEXT_PUBLIC_WWV_EDITION=demo`)
### Step A4 — Flip auth on engine
Coolify → `wwv-engine-stack` → Environment → `WWV_SKIP_WS_AUTH=false` → Deploy.
### Step A5 — Smoke test
```bash
node local-scripts/local-smoke-test.mjs # pointed at production URLs
# Step 1: JWKS ✅ Step 2: Exchange ✅ Step 3: WS welcome ✅
```
---
## Track B — PKCE OAuth Server on Marketplace (Local/Self-Hosted)
Implements the missing `GET /oauth/authorize` (consent page) and
`POST /api/oauth/token` (code exchange) on the marketplace. Uses admin-password
auth for now — Supabase replaces the password check when ADR-003 lands. Same
OAuth interface; no client-side changes needed during that migration.
### Files to create/modify in `c:/dev/worldwideview-marketplace/`
**`prisma/schema.prisma`** — add:
```prisma
model OAuthCode {
id String @id @default(cuid())
code String @unique
userId String
user User @relation(fields: [userId], references: [id])
codeChallenge String
redirectUri String
clientId String
scope String
state String
expiresAt DateTime
usedAt DateTime?
@@index([code])
}
```
Add `oauthCodes OAuthCode[]` to `User`. Run `npx prisma db push` (SQLite, local).
**`src/app/oauth/authorize/page.tsx`** — consent UI:
- Reads URL params: `client_id`, `code_challenge`, `state`, `redirect_uri`, `scope`
- Renders form: "Local App is requesting access to your data. Enter your admin password."
- Stores params in hidden fields, POSTs to `/api/oauth/authorize`
**`src/app/api/oauth/authorize/route.ts`** (POST):
- Validates `password === process.env.ADMIN_PASSWORD`
- Finds or creates `User { email: "local@marketplace", tier: "pro" }`
- Creates `OAuthCode { code: randomBytes(32).hex, expiresAt: now+10min, ... }`
- Redirects to `redirectUri?code=<code>&state=<state>`
- **ADR-003 migration point:** swap password check for `supabase.auth.getUser()` session check
**`src/app/api/oauth/token/route.ts`** (POST, `application/x-www-form-urlencoded`):
- Parses `{ grant_type, code, code_verifier, client_id, redirect_uri }`
- Finds `OAuthCode`; verifies not expired, not used, clientId + redirectUri match
- PKCE: `createHash('sha256').update(code_verifier).digest('base64url') === codeChallenge`
- Marks `usedAt = now` (replay protection)
- Generates raw API key: `randomBytes(32).hex`
- SHA256-hash → creates `MarketplaceApiKey { userId, keyHash }`
- Returns `{ access_token: rawApiKey, token_type: "bearer" }`
**Note:** `connect/route.ts` in local app redirects to `/oauth/authorize` (no `/api/` prefix)
— the consent page is at that path. The POST handler is at `/api/oauth/authorize`.
### Local app env (`.env.local`):
```
NEXT_PUBLIC_WWV_MARKETPLACE_URL=http://localhost:3001
ENCRYPTION_MASTER_KEY=<any-32+-char-string>
```
### Verification (Track B)
1. Start all 3 services locally
2. Browser: `http://localhost:3000/api/marketplace/connect`
→ redirects to `http://localhost:3001/oauth/authorize?...`
→ enter admin password → redirects back to local app
3. `GET http://localhost:3000/api/auth/ticket?pluginId=aviation` → `{ token: "eyJ..." }`
4. Check WebSocket in devtools → `{ type: "welcome" }`
---
## ADR-003 Migration Path (future, no client changes needed)
When Supabase Auth is wired into the marketplace:
- `/api/oauth/authorize` replaces `password === ADMIN_PASSWORD` with `supabase.auth.getUser()`
- If no session → redirect to `worldwideview.dev/login?redirect_to=marketplace.../oauth/authorize?...`
- Everything else stays identical — same code format, same token endpoint, same local app callback
---
## Order of Execution
1. **Track A first** (demo service credential + auth flip) — unblocks production
2. **Track B in parallel or after** (PKCE for local instances) — enables proper self-hosted UX
3. **ADR-003 later** — migrates marketplace auth internals without touching clients
@@ -1,874 +0,0 @@
# WorldWideView Auth Rollout — ADR-001 + ADR-003
> **For agentic workers:** REQUIRED SUB-SKILL: use `superpowers:executing-plans` to implement
> this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
**Goal:** Get ADR-001 (decentralized plugin auth) and ADR-003 (shared ecosystem identity)
into production with the least possible downtime. `/cloud` plan registration and billing
are deferred.
**Architecture:** One shared Supabase project is the identity store. `worldwideview.dev`
(the `worldwideview-web` repo) becomes the SSR auth host; a parent-domain cookie
(`.worldwideview.dev`) is inherited by every subdomain. The Marketplace becomes a PKCE
OAuth server on top of that session. The Local App is the PKCE client. The Data Engine
validates 5-minute JWTs on its WebSocket.
**Tech Stack:** Next.js 16 App Router, `@supabase/ssr`, Supabase Auth, Ed25519 JWTs
(`jose`), Prisma, Docker + Coolify.
---
## Northstar Sequence (the map)
Build everything locally on `wwv.local` HTTPS first, then deploy production one repo at a
time in this dependency order. Each deploy is independently revertible.
1. **Phase 0** — Provision Supabase + local HTTPS dev (human setup — see checklist below).
2. **Phase 1** — `worldwideview-web` becomes the SSR auth host. ← **detailed below**
3. **Phase 2** — Marketplace consumes the session + adds the PKCE OAuth server.
4. **Phase 3** — Merge the ADR-001 Local App worktree.
5. **Phase 4** — Data Engine WebSocket first-message JWT auth (report-only → enforcing).
6. **Phase 5** — Final cutover, link the nav, update docs, mark ADRs Accepted.
Phases 2–5 are detailed once Phase 1 lands and Assumption #2 (below) is confirmed.
---
## Phase 0 — Human Setup Checklist (do this while Phase 1 code is written)
These cannot be automated — they need accounts, browsers, admin rights, or a DNS panel.
- [ ] **0.1 Create the Supabase project.** supabase.com → New project. Pick a region close
to the Coolify host. Save the DB password. Copy three values: Project URL, `anon` public
key, `service_role` secret key.
- [ ] **0.2 Enable auth providers** (Supabase dashboard → Authentication → Providers):
- Email: enable, "Confirm email" ON.
- Google: create an OAuth Client in Google Cloud Console (APIs & Services → Credentials →
OAuth client ID → Web app). Authorized redirect URI:
`https://<project-ref>.supabase.co/auth/v1/callback`. Paste Client ID + Secret into
Supabase.
- GitHub: GitHub → Settings → Developer settings → OAuth Apps → New. Authorization
callback URL: `https://<project-ref>.supabase.co/auth/v1/callback`. Paste Client ID +
Secret into Supabase.
- [ ] **0.3 Set URL configuration** (Authentication → URL Configuration):
- Site URL: `https://worldwideview.dev`
- Redirect URLs (add all): `https://worldwideview.dev/**`,
`https://*.worldwideview.dev/**`, `https://wwv.local/**`, `https://*.wwv.local/**`
- [ ] **0.4 Install mkcert** (one-time, per dev machine). PowerShell as admin:
`choco install mkcert` then `mkcert -install`.
- [ ] **0.5 Generate the local TLS cert.** In `c:\dev\wwv`:
`mkcert "*.wwv.local" wwv.local` — produces a `.pem` cert + key pair.
- [ ] **0.6 Edit the hosts file** (admin). Add to `C:\Windows\System32\drivers\etc\hosts`:
`127.0.0.1 wwv.local marketplace.wwv.local app.wwv.local`
- [ ] **0.7 Confirm production DNS.** `worldwideview.dev`, `marketplace.worldwideview.dev`,
`app.worldwideview.dev` (and a `*.worldwideview.dev` wildcard) must point at the Coolify
host. Check the DNS provider; add records if missing.
- [ ] **0.8 Decide Assumption #2** — Marketplace keeps SQLite, or migrates to Supabase
Postgres? (See "Assumptions" at the end. This unblocks the Phase 2 detailed plan.)
- [ ] **0.9 Hand me the Supabase values** (or paste them yourself) so they can go into the
local `.env` files and the Coolify env vars — never committed to git. Needed:
`NEXT_PUBLIC_SUPABASE_URL`, `NEXT_PUBLIC_SUPABASE_ANON_KEY`, `SUPABASE_SERVICE_ROLE_KEY`.
- [ ] **0.10 (Optional) Resend SMTP** — only if launching with magic-link login. Create a
Resend account, verify a sending domain, paste SMTP creds into Supabase → Project
Settings → Auth → SMTP. Skip if launching with email+password + OAuth only.
---
# Phase 1 Implementation Plan — `worldwideview-web` Auth Host
**Repo:** `c:\dev\wwv\worldwideview-web`
**Outcome:** `worldwideview.dev` serves the landing site (static) plus working SSR
`/login`, `/signup`, `/accounts`, and `/api/auth/callback`, with sessions written to a
parent-domain cookie. The container runs as Node, not Nginx.
### File Structure
| File | Responsibility |
|---|---|
| `next.config.ts` | Drop static export, switch to `standalone`, add headers |
| `vitest.config.ts` (new) | Unit test runner config |
| `src/lib/supabase/cookieOptions.ts` (new) | Pure helper: resolve cookie domain from env |
| `src/lib/supabase/cookieOptions.test.ts` (new) | Unit tests for the helper |
| `src/lib/supabase/client.ts` | Browser client via `@supabase/ssr` |
| `src/lib/supabase/server.ts` (new) | Server client via `@supabase/ssr` |
| `middleware.ts` (new) | Refresh the session on every request |
| `src/app/api/health/route.ts` (new) | Health endpoint for Coolify deploy gating |
| `src/app/login/page.tsx` | SSR login: email+password + OAuth |
| `src/app/login/actions.ts` (new) | Server action: password sign-in |
| `src/app/login/oauth-buttons.tsx` (new) | Client component: Google/GitHub buttons |
| `src/app/signup/page.tsx` (new) | SSR signup |
| `src/app/signup/actions.ts` (new) | Server action: account creation |
| `src/app/accounts/page.tsx` (new) | Account view + logout |
| `src/app/accounts/actions.ts` (new) | Server action: sign out |
| `src/app/api/auth/callback/route.ts` (new) | OAuth code → session exchange |
| `Dockerfile` | Node standalone runtime instead of Nginx |
| `docker-compose.yml` | Port mapping for the Node container |
| `.env.local` (new, not committed) | Local Supabase + cookie-domain vars |
---
### Task 1: Add dependencies and the test runner
**Files:** Modify `package.json`; Create `vitest.config.ts`; Create `.env.local`.
- [ ] **Step 1: Install `@supabase/ssr`.**
Run in `c:\dev\wwv\worldwideview-web`:
```bash
pnpm add @supabase/ssr
```
Expected: `@supabase/ssr` appears under `dependencies` in `package.json`.
- [ ] **Step 2: Add a `test` script.** In `package.json` `scripts`, add:
```json
"test": "vitest run"
```
- [ ] **Step 3: Create `vitest.config.ts`.**
```ts
import { defineConfig } from 'vitest/config'
import react from '@vitejs/plugin-react'
export default defineConfig({
plugins: [react()],
test: {
environment: 'jsdom',
globals: true,
},
})
```
- [ ] **Step 4: Create `.env.local`** (NOT committed — confirm it is in `.gitignore`):
```
NEXT_PUBLIC_SUPABASE_URL=<from Phase 0.1>
NEXT_PUBLIC_SUPABASE_ANON_KEY=<from Phase 0.1>
NEXT_PUBLIC_WWV_COOKIE_DOMAIN=.wwv.local
```
- [ ] **Step 5: Verify the `@supabase/ssr` API.** Use context7 (`resolve-library-id`
→ `query-docs` for `@supabase/ssr` "Next.js App Router server client middleware") to
confirm `createServerClient` / `createBrowserClient` signatures match the installed
version before writing Tasks 3–4. The `getAll`/`setAll` cookie pattern below is the
stable pattern; adjust only if the installed version differs.
- [ ] **Step 6: Commit.**
```bash
git add package.json pnpm-lock.yaml vitest.config.ts
git commit -m "chore: add @supabase/ssr and vitest runner"
```
---
### Task 2: Cookie-domain resolver helper (TDD)
The cookie `domain` must be `.worldwideview.dev` in production, `.wwv.local` in dev, and
**unset** when running on plain `localhost` (browsers reject a `domain` on `localhost`).
This is the one piece of pure logic worth a unit test.
**Files:** Create `src/lib/supabase/cookieOptions.ts`,
`src/lib/supabase/cookieOptions.test.ts`.
- [ ] **Step 1: Write the failing test.** `src/lib/supabase/cookieOptions.test.ts`:
```ts
import { describe, it, expect } from 'vitest'
import { resolveCookieDomain } from './cookieOptions'
describe('resolveCookieDomain', () => {
it('returns undefined when nothing is configured', () => {
expect(resolveCookieDomain(undefined)).toBeUndefined()
})
it('returns undefined for an empty string', () => {
expect(resolveCookieDomain('')).toBeUndefined()
})
it('returns the configured domain', () => {
expect(resolveCookieDomain('.worldwideview.dev')).toBe('.worldwideview.dev')
})
it('trims surrounding whitespace', () => {
expect(resolveCookieDomain(' .wwv.local ')).toBe('.wwv.local')
})
})
```
- [ ] **Step 2: Run the test, confirm it fails.**
Run: `pnpm test`
Expected: FAIL — `resolveCookieDomain` is not exported / file not found.
- [ ] **Step 3: Implement `src/lib/supabase/cookieOptions.ts`.**
```ts
export function resolveCookieDomain(raw: string | undefined): string | undefined {
const trimmed = raw?.trim()
return trimmed ? trimmed : undefined
}
export function buildCookieOptions() {
return {
domain: resolveCookieDomain(process.env.NEXT_PUBLIC_WWV_COOKIE_DOMAIN),
path: '/',
sameSite: 'lax' as const,
secure: true,
}
}
```
- [ ] **Step 4: Run the test, confirm it passes.**
Run: `pnpm test`
Expected: PASS — 4 passing.
- [ ] **Step 5: Commit.**
```bash
git add src/lib/supabase/cookieOptions.ts src/lib/supabase/cookieOptions.test.ts
git commit -m "feat: add cookie-domain resolver for cross-subdomain sessions"
```
---
### Task 3: Supabase browser + server clients
**Files:** Modify `src/lib/supabase/client.ts`; Create `src/lib/supabase/server.ts`.
- [ ] **Step 1: Rewrite `src/lib/supabase/client.ts`** (browser client — now a factory):
```ts
import { createBrowserClient } from '@supabase/ssr'
import { buildCookieOptions } from './cookieOptions'
export function createClient() {
return createBrowserClient(
process.env.NEXT_PUBLIC_SUPABASE_URL!,
process.env.NEXT_PUBLIC_SUPABASE_ANON_KEY!,
{ cookieOptions: buildCookieOptions() },
)
}
```
- [ ] **Step 2: Create `src/lib/supabase/server.ts`** (server client):
```ts
import { createServerClient } from '@supabase/ssr'
import { cookies } from 'next/headers'
import { buildCookieOptions } from './cookieOptions'
export async function createClient() {
const cookieStore = await cookies()
return createServerClient(
process.env.NEXT_PUBLIC_SUPABASE_URL!,
process.env.NEXT_PUBLIC_SUPABASE_ANON_KEY!,
{
cookieOptions: buildCookieOptions(),
cookies: {
getAll() {
return cookieStore.getAll()
},
setAll(cookiesToSet) {
try {
cookiesToSet.forEach(({ name, value, options }) =>
cookieStore.set(name, value, options),
)
} catch {
// Called from a Server Component — middleware refreshes instead.
}
},
},
},
)
}
```
- [ ] **Step 3: Migrate existing importers of the old `supabase` singleton.**
Run: `grep -rn "lib/supabase/client" src/` — every hit that did `import { supabase }`
must change to `import { createClient }` + `const supabase = createClient()`. The `hub/`
page is a known importer. Update each.
- [ ] **Step 4: Verify the build typechecks.**
Run: `pnpm build`
Expected: build proceeds past type-checking (it may still fail later on `output: export`
+ middleware — that is fixed in Task 5; a clean `tsc` pass here is the goal).
- [ ] **Step 5: Commit.**
```bash
git add src/lib/supabase/client.ts src/lib/supabase/server.ts src/app/hub
git commit -m "feat: add @supabase/ssr browser and server clients"
```
---
### Task 4: Session-refresh middleware
**Files:** Create `middleware.ts` (repo root, sibling of `src/`).
- [ ] **Step 1: Create `middleware.ts`.**
```ts
import { createServerClient } from '@supabase/ssr'
import { NextResponse, type NextRequest } from 'next/server'
import { buildCookieOptions } from './src/lib/supabase/cookieOptions'
export async function middleware(request: NextRequest) {
let response = NextResponse.next({ request })
const supabase = createServerClient(
process.env.NEXT_PUBLIC_SUPABASE_URL!,
process.env.NEXT_PUBLIC_SUPABASE_ANON_KEY!,
{
cookieOptions: buildCookieOptions(),
cookies: {
getAll() {
return request.cookies.getAll()
},
setAll(cookiesToSet) {
cookiesToSet.forEach(({ name, value }) =>
request.cookies.set(name, value),
)
response = NextResponse.next({ request })
cookiesToSet.forEach(({ name, value, options }) =>
response.cookies.set(name, value, options),
)
},
},
},
)
// Refreshes the session cookie if expired. Do NOT add logic between
// createServerClient and getUser().
await supabase.auth.getUser()
return response
}
export const config = {
matcher: ['/((?!_next/static|_next/image|favicon.ico|.*\\.(?:svg|png|jpg|jpeg|gif|webp)$).*)'],
}
```
- [ ] **Step 2: Verify it compiles.**
Run: `pnpm build`
Expected: no type errors from `middleware.ts`.
- [ ] **Step 3: Commit.**
```bash
git add middleware.ts
git commit -m "feat: add session-refresh middleware"
```
---
### Task 5: Convert `next.config.ts` from static export to SSR
**Files:** Modify `next.config.ts`.
- [ ] **Step 1: Replace `next.config.ts` entirely.**
```ts
import type { NextConfig } from "next";
const nextConfig: NextConfig = {
output: "standalone",
images: {
unoptimized: true,
},
async headers() {
return [
{
source: "/:path*",
headers: [
{ key: "X-Content-Type-Options", value: "nosniff" },
{ key: "X-Frame-Options", value: "DENY" },
{ key: "Referrer-Policy", value: "strict-origin-when-cross-origin" },
],
},
];
},
};
export default nextConfig;
```
Note: any custom headers currently in `nginx.conf` should be folded into the `headers()`
array. Keep CSP out for now unless `nginx.conf` already had one — an over-strict CSP
silently breaks Supabase XHR calls.
- [ ] **Step 2: Build and confirm SSR mode.**
Run: `pnpm build`
Expected: build output lists routes as a mix of `○ (Static)` and `ƒ (Dynamic)`; auth
routes appear as Dynamic. No `out/` directory is produced; `.next/standalone` is.
- [ ] **Step 3: Commit.**
```bash
git add next.config.ts
git commit -m "feat: switch worldwideview-web from static export to standalone SSR"
```
---
### Task 6: Health endpoint
**Files:** Create `src/app/api/health/route.ts`.
- [ ] **Step 1: Create the route.**
```ts
import { NextResponse } from 'next/server'
export const dynamic = 'force-dynamic'
export function GET() {
return NextResponse.json({ status: 'ok' })
}
```
- [ ] **Step 2: Verify.**
Run: `pnpm build && pnpm start` then in another shell: `curl -i http://localhost:3000/api/health`
Expected: `200 OK`, body `{"status":"ok"}`.
- [ ] **Step 3: Commit.**
```bash
git add src/app/api/health/route.ts
git commit -m "feat: add health endpoint for deploy gating"
```
---
### Task 7: Login page (email+password + OAuth)
**Files:** Modify `src/app/login/page.tsx`; Create `src/app/login/actions.ts`,
`src/app/login/oauth-buttons.tsx`.
- [ ] **Step 1: Create `src/app/login/actions.ts`.**
```ts
'use server'
import { redirect } from 'next/navigation'
import { createClient } from '../../lib/supabase/server'
export async function signInWithPassword(formData: FormData) {
const email = String(formData.get('email') ?? '')
const password = String(formData.get('password') ?? '')
const next = String(formData.get('next') ?? '/accounts')
const supabase = await createClient()
const { error } = await supabase.auth.signInWithPassword({ email, password })
if (error) {
redirect(`/login?error=${encodeURIComponent(error.message)}&next=${encodeURIComponent(next)}`)
}
redirect(next)
}
```
- [ ] **Step 2: Create `src/app/login/oauth-buttons.tsx`.**
```tsx
'use client'
import { createClient } from '../../lib/supabase/client'
import styles from '../hub/hub.module.css'
export function OAuthButtons({ next }: { next: string }) {
const signIn = async (provider: 'google' | 'github') => {
const supabase = createClient()
await supabase.auth.signInWithOAuth({
provider,
options: {
redirectTo: `${window.location.origin}/api/auth/callback?next=${encodeURIComponent(next)}`,
},
})
}
return (
<div style={{ display: 'flex', flexDirection: 'column', gap: 'var(--space-sm)' }}>
<button className={styles.submitButton} onClick={() => signIn('google')} type="button">
Continue with Google
</button>
<button className={styles.submitButton} onClick={() => signIn('github')} type="button">
Continue with GitHub
</button>
</div>
)
}
```
- [ ] **Step 3: Rewrite `src/app/login/page.tsx`** as a server component:
```tsx
import { signInWithPassword } from './actions'
import { OAuthButtons } from './oauth-buttons'
import styles from '../hub/hub.module.css'
export default async function LoginPage({
searchParams,
}: {
searchParams: Promise<{ error?: string; next?: string }>
}) {
const { error, next = '/accounts' } = await searchParams
return (
<div className={styles.hubContainer}>
<div className={styles.glassCard} style={{ maxWidth: '400px', marginTop: '10vh' }}>
<h1 className={styles.title}>Welcome Back</h1>
<p style={{ textAlign: 'center', marginBottom: 'var(--space-lg)', color: 'var(--text-secondary)' }}>
Sign in to your WorldWideView account
</p>
<form action={signInWithPassword}>
<input type="hidden" name="next" value={next} />
<input className={styles.inputField} type="email" name="email"
placeholder="Email address" required />
<input className={styles.inputField} type="password" name="password"
placeholder="Password" required />
<button className={styles.submitButton} type="submit">Sign In</button>
</form>
<div style={{ textAlign: 'center', margin: 'var(--space-md) 0', color: 'var(--text-secondary)' }}>
or
</div>
<OAuthButtons next={next} />
{error && (
<p style={{ marginTop: 'var(--space-md)', textAlign: 'center', fontSize: '0.9rem',
color: 'var(--color-accent)' }}>
{error}
</p>
)}
<p style={{ marginTop: 'var(--space-md)', textAlign: 'center', fontSize: '0.9rem' }}>
No account? <a href={`/signup?next=${encodeURIComponent(next)}`}>Create one</a>
</p>
</div>
</div>
)
}
```
- [ ] **Step 4: Verify.**
Run: `pnpm build`
Expected: `/login` builds as a Dynamic route, no type errors.
- [ ] **Step 5: Commit.**
```bash
git add src/app/login
git commit -m "feat: SSR login with email/password and OAuth"
```
---
### Task 8: Signup page
**Files:** Create `src/app/signup/page.tsx`, `src/app/signup/actions.ts`.
- [ ] **Step 1: Create `src/app/signup/actions.ts`.**
```ts
'use server'
import { redirect } from 'next/navigation'
import { createClient } from '../../lib/supabase/server'
export async function signUp(formData: FormData) {
const email = String(formData.get('email') ?? '')
const password = String(formData.get('password') ?? '')
const next = String(formData.get('next') ?? '/accounts')
const supabase = await createClient()
const { error } = await supabase.auth.signUp({
email,
password,
options: { emailRedirectTo: `${process.env.NEXT_PUBLIC_SITE_URL ?? ''}/api/auth/callback?next=${encodeURIComponent(next)}` },
})
if (error) {
redirect(`/signup?error=${encodeURIComponent(error.message)}&next=${encodeURIComponent(next)}`)
}
redirect(`/login?message=${encodeURIComponent('Check your email to confirm your account.')}&next=${encodeURIComponent(next)}`)
}
```
Note: add `NEXT_PUBLIC_SITE_URL` to `.env.local` (`https://wwv.local`) and to Coolify
(`https://worldwideview.dev`).
- [ ] **Step 2: Create `src/app/signup/page.tsx`.**
```tsx
import { signUp } from './actions'
import styles from '../hub/hub.module.css'
export default async function SignupPage({
searchParams,
}: {
searchParams: Promise<{ error?: string; next?: string }>
}) {
const { error, next = '/accounts' } = await searchParams
return (
<div className={styles.hubContainer}>
<div className={styles.glassCard} style={{ maxWidth: '400px', marginTop: '10vh' }}>
<h1 className={styles.title}>Create Account</h1>
<form action={signUp}>
<input type="hidden" name="next" value={next} />
<input className={styles.inputField} type="email" name="email"
placeholder="Email address" required />
<input className={styles.inputField} type="password" name="password"
placeholder="Password (8+ characters)" minLength={8} required />
<button className={styles.submitButton} type="submit">Sign Up</button>
</form>
{error && (
<p style={{ marginTop: 'var(--space-md)', textAlign: 'center', fontSize: '0.9rem',
color: 'var(--color-accent)' }}>
{error}
</p>
)}
<p style={{ marginTop: 'var(--space-md)', textAlign: 'center', fontSize: '0.9rem' }}>
Already have an account? <a href={`/login?next=${encodeURIComponent(next)}`}>Sign in</a>
</p>
</div>
</div>
)
}
```
- [ ] **Step 3: Verify.** Run: `pnpm build` — `/signup` builds as Dynamic.
- [ ] **Step 4: Commit.**
```bash
git add src/app/signup
git commit -m "feat: SSR signup page"
```
---
### Task 9: OAuth callback route
**Files:** Create `src/app/api/auth/callback/route.ts`.
- [ ] **Step 1: Create the route.**
```ts
import { NextResponse } from 'next/server'
import { createClient } from '../../../../lib/supabase/server'
export async function GET(request: Request) {
const { searchParams, origin } = new URL(request.url)
const code = searchParams.get('code')
const next = searchParams.get('next') ?? '/accounts'
if (code) {
const supabase = await createClient()
const { error } = await supabase.auth.exchangeCodeForSession(code)
if (!error) {
return NextResponse.redirect(`${origin}${next}`)
}
}
return NextResponse.redirect(`${origin}/login?error=${encodeURIComponent('Authentication failed')}`)
}
```
- [ ] **Step 2: Verify.** Run: `pnpm build` — route compiles.
- [ ] **Step 3: Commit.**
```bash
git add src/app/api/auth/callback/route.ts
git commit -m "feat: OAuth callback route for code-to-session exchange"
```
---
### Task 10: Accounts page + logout
**Files:** Create `src/app/accounts/page.tsx`, `src/app/accounts/actions.ts`.
- [ ] **Step 1: Create `src/app/accounts/actions.ts`.**
```ts
'use server'
import { redirect } from 'next/navigation'
import { createClient } from '../../lib/supabase/server'
export async function signOut() {
const supabase = await createClient()
await supabase.auth.signOut()
redirect('/login')
}
```
- [ ] **Step 2: Create `src/app/accounts/page.tsx`** (guards unauthenticated users):
```tsx
import { redirect } from 'next/navigation'
import { createClient } from '../../lib/supabase/server'
import { signOut } from './actions'
import styles from '../hub/hub.module.css'
export default async function AccountsPage() {
const supabase = await createClient()
const { data: { user } } = await supabase.auth.getUser()
if (!user) {
redirect('/login?next=/accounts')
}
return (
<div className={styles.hubContainer}>
<div className={styles.glassCard} style={{ maxWidth: '480px', marginTop: '10vh' }}>
<h1 className={styles.title}>Your Account</h1>
<p style={{ textAlign: 'center', color: 'var(--text-secondary)' }}>{user.email}</p>
<form action={signOut}>
<button className={styles.submitButton} type="submit">Sign Out</button>
</form>
</div>
</div>
)
}
```
- [ ] **Step 3: Verify.** Run: `pnpm build` — `/accounts` builds as Dynamic.
- [ ] **Step 4: Commit.**
```bash
git add src/app/accounts
git commit -m "feat: account page with parent-domain logout"
```
---
### Task 11: Dockerfile → Node standalone runtime
**Files:** Modify `Dockerfile`; verify `.dockerignore`; modify `docker-compose.yml`.
- [ ] **Step 1: Replace `Dockerfile` entirely.**
```dockerfile
# Stage 1: Build
FROM node:22-alpine AS builder
WORKDIR /app
RUN corepack enable pnpm
COPY package.json pnpm-lock.yaml* ./
RUN pnpm i --frozen-lockfile
COPY . .
# NEXT_PUBLIC_* vars are inlined at build time — must be present as build args.
ARG NEXT_PUBLIC_SUPABASE_URL
ARG NEXT_PUBLIC_SUPABASE_ANON_KEY
ARG NEXT_PUBLIC_WWV_COOKIE_DOMAIN
ARG NEXT_PUBLIC_SITE_URL
ENV NEXT_PUBLIC_SUPABASE_URL=$NEXT_PUBLIC_SUPABASE_URL
ENV NEXT_PUBLIC_SUPABASE_ANON_KEY=$NEXT_PUBLIC_SUPABASE_ANON_KEY
ENV NEXT_PUBLIC_WWV_COOKIE_DOMAIN=$NEXT_PUBLIC_WWV_COOKIE_DOMAIN
ENV NEXT_PUBLIC_SITE_URL=$NEXT_PUBLIC_SITE_URL
RUN pnpm run build
# Stage 2: Runtime
FROM node:22-alpine AS runner
WORKDIR /app
ENV NODE_ENV=production
RUN addgroup --system --gid 1001 nodejs \
&& adduser --system --uid 1001 nextjs
COPY --from=builder /app/public ./public
COPY --from=builder --chown=nextjs:nodejs /app/.next/standalone ./
COPY --from=builder --chown=nextjs:nodejs /app/.next/static ./.next/static
USER nextjs
EXPOSE 3000
ENV PORT=3000 HOSTNAME=0.0.0.0
CMD ["node", "server.js"]
```
- [ ] **Step 2: Confirm `.dockerignore` excludes `node_modules`, `.next`, `out`.**
If missing, create `.dockerignore` with: `node_modules`, `.next`, `out`, `.git`.
- [ ] **Step 3: Update `docker-compose.yml`** so the web service maps a host port to
container `3000` (was Nginx `80`), e.g. `ports: - "3001:3000"`, and passes the four
`NEXT_PUBLIC_*` build args.
- [ ] **Step 4: Build the image locally.**
Run: `docker build -t wwv-web-test .`
Expected: build completes; the final image runs `node server.js`.
- [ ] **Step 5: Commit.**
```bash
git add Dockerfile .dockerignore docker-compose.yml
git commit -m "feat: run worldwideview-web as a Node standalone container"
```
Note: `nginx.conf` is now unused — delete it in Phase 5 cleanup once production is verified.
---
### Task 12: Bump version
**Files:** Modify `package.json`.
- [ ] **Step 1:** Bump `version` `0.5.0` → `0.6.0` (a feature release — new auth surface).
- [ ] **Step 2: Commit.**
```bash
git add package.json
git commit -m "chore: bump version to 0.6.0"
```
---
## Phase 1 Verification
**Local (on `wwv.local` HTTPS — requires Phase 0.4–0.6 done):**
1. Run the dev server bound to `wwv.local` over HTTPS (via the mkcert cert or a local TLS
proxy). Open `https://wwv.local/signup`.
2. Create an account → confirm the email → land back authenticated.
3. Sign out, then `https://wwv.local/login` with email+password → reach `/accounts`.
4. In DevTools → Application → Cookies: the Supabase `sb-*` cookies have
`Domain = .wwv.local`.
5. Test Google and GitHub buttons → provider → `/api/auth/callback` → `/accounts`.
6. `https://wwv.local/accounts` while logged out → redirect to `/login?next=/accounts`.
7. Sign out → cookies cleared on `.wwv.local`.
8. `curl -k https://wwv.local/api/health` → `{"status":"ok"}`.
9. Marketing pages (`/`, `/about`, `/docs`) still load and are statically generated.
**Production deploy (the one downtime-sensitive cutover):**
1. Push the branch, open a PR, run `/pr-review`.
2. In Coolify: update the `worldwideview-web` app — new Dockerfile build, set the four
`NEXT_PUBLIC_*` build args + runtime env vars, set the health check to `/api/health`.
3. **Safest:** deploy first to a temporary domain (`web-next.worldwideview.dev`), smoke-test
the full flow, then repoint `worldwideview.dev` (blue-green). Otherwise rely on Coolify's
health-check-gated deploy so traffic only switches after `/api/health` passes.
4. After cutover: smoke-test `worldwideview.dev` landing + `/login` + `/signup`.
---
## Assumptions / Open Questions
1. **No Supabase project exists yet** — Phase 0 creates one.
2. **Marketplace keeps SQLite (Option A)** — recommended: the Marketplace consumes the
Supabase *session* for identity but keeps plugin/API-key data in SQLite. If you want a
full SQLite → Supabase-Postgres migration instead, Phase 2 is much larger. **This is the
one decision that changes scope — please confirm (Phase 0.8).**
3. **`/cloud` + billing deferred** — `/accounts` is minimal (profile + logout), no Stripe.
4. **API-key management UI lives in the Marketplace** (ADR-001C wording), not in `/accounts`.
5. **`SigningKey` rotation table** treated as a fast-follow, not a launch blocker.
6. **The ADR-001 worktree branch needs consolidation** before merge (messy branch name,
uncommitted changes) — handled in Phase 3.
7. **`hub.module.css` is reused** for auth page styling to avoid new CSS scope. If you want
dedicated auth styling, that is an easy add but not assumed here.
8. **No Playwright in `worldwideview-web`** — Phase 1 verification is browser-manual. Adding
a Playwright E2E suite is a reasonable but separate follow-up.
---
## Phase 1 — Implemented & Verified (2026-05-22)
All 12 implementation tasks done. `pnpm test` 16/16 green; `pnpm build` green (marketing
pages Static, the 5 auth routes Dynamic, middleware active). `.env.local` verified against
the live Supabase project `kvlnzjtcstnaqkpqrquf` — URL + publishable key match. The new
API-key system is used throughout (`NEXT_PUBLIC_SUPABASE_PUBLISHABLE_KEY`); `getClaims()`
for session checks.
### Discovered at commit time: pre-existing uncommitted work (NOT Phase 1)
The `worldwideview-web` working tree already held a separate, coherent change —
**"consolidate docs off the `docs.` subdomain back onto `/docs`"**:
- `src/proxy.ts` — **deleted** (was the `docs.` subdomain rewrite middleware; used Next 16's
`proxy.ts` convention with `export function proxy`)
- `src/app/docs/layout.tsx` — dropped the `NEXT_PUBLIC_IS_DOCS_SUBDOMAIN` branching
- `src/components/Footer.tsx` — docs links `docs.worldwideview.dev` → `/docs`
- `vercel.json` — removed the docs-subdomain redirects (now `{}`)
- `.vscode/settings.json` — new untracked file
This is unrelated to the auth host. **Decision (user, 2026-05-22): commit the
docs-consolidation work first as its own commit, then Phase 1.**
### Finalized commit sequence (branch `feat/auth-host` off `main`)
1. **Commit 1 — docs consolidation.** Stage only: `src/proxy.ts` (deletion),
`src/app/docs/layout.tsx`, `src/components/Footer.tsx`, `vercel.json`. Message:
`refactor(docs): serve docs from /docs instead of the docs. subdomain`.
2. **Rename** Phase 1's root `middleware.ts` → `src/proxy.ts` (`export function proxy`).
Clean now — Commit 1 removed the old `src/proxy.ts`, so Commit 2 just adds the new auth
one. Matches the repo's Next-16 `proxy.ts` convention.
3. **Commit 2 — Phase 1.** Version bump `0.5.0 → 0.6.0`; stage all Phase 1 files. Message:
`feat: add SSR auth host with shared Supabase identity (ADR-003)`.
4. Push `feat/auth-host`, open the PR (remote: `github.com/silvertakana/worldwideview-web`).
`.vscode/settings.json` (stray untracked file) is left out of both commits.
@@ -1,306 +0,0 @@
# Plan: Marketplace Authentication (Supabase Auth — Shared Identity)
## Context
The Marketplace (`marketplace.worldwideview.dev`, repo: `c:/dev/worldwideview-marketplace`) and the Cloud App (`app.worldwideview.dev`) share one identity ecosystem. A user who creates an account on the Marketplace can use that same account to access the Cloud App, and vice versa — no second signup required. This is called **Federated Identity / Single Sign-On (SSO)** and is the same model used by GitHub (one account → GitHub + GitHub Marketplace + GitHub Actions), Atlassian (one account → Jira + Confluence + Marketplace), and Stripe (one account → Dashboard + Stripe Apps).
The Marketplace has no database yet, so Supabase is the right choice: one Supabase project serves as the **shared identity store** for all WorldWideView products. Both the Marketplace and the Cloud App point at the same Supabase project URL — the same `auth.users` table — so a user is one entity across the entire ecosystem.
The Marketplace also acts as the **OAuth 2.0 Authorization Server** in the ADR-001 PKCE flow — when a Local App user clicks "Connect to Marketplace," they are redirected here to authorize, and the Marketplace issues a short-lived code that the Local App exchanges for an API Key.
> **⚠️ Rule conflict to resolve:** `.agents/rules/cloud-auth-architecture.md` currently says `app.worldwideview.dev` owns all login UI and the Marketplace redirects there. This plan reverses that — **Marketplace is the identity source**, and the Cloud App will defer to it. That rule file must be updated after this plan is approved.
---
## Decision: Shared Supabase Auth + Supabase Postgres
**Why:**
- **One user table for the whole ecosystem** — `auth.users` is shared. A user created on Marketplace is immediately valid on Cloud App. No sync job, no duplication.
- No database to spin up — Supabase project gives you Postgres + Auth in one dashboard click
- Google OAuth, GitHub OAuth, magic link, email+password all configured in the Supabase dashboard (no code for provider setup)
- Prisma connects to Supabase Postgres via standard `DATABASE_URL` — app data (plugins, installs, etc.) lives alongside auth tables
- Well-trodden Stripe + Supabase integration path via the [Vercel subscription payments starter](https://github.com/vercel/nextjs-subscription-payments)
- Free tier covers early-stage Marketplace comfortably
- `@supabase/ssr` package has first-class Next.js App Router support (server components, middleware, server actions)
**Accepted tradeoff:** Supabase vendor dependency — if pricing changes at scale, migration to self-hosted Postgres + Better Auth is straightforward because Prisma abstracts the DB layer.
**Known gotcha — cookie domain across subdomains:** Supabase session cookies are scoped per-app by default. A session from `marketplace.worldwideview.dev` does not automatically carry over to `app.worldwideview.dev`. Fix: after login on either app, set a shared JWT cookie on the parent domain `worldwideview.dev` (via custom middleware) that both subdomains can read and validate against the shared Supabase public key. This is the only custom plumbing required — both apps still use the same Supabase JWT secret for validation.
**Upgrade path:** If the Marketplace later needs to support third-party app integrations (external developers building plugins), flip to the **Marketplace-as-OAuth-Provider** model — the user table stays the same, you add the PKCE authorization layer on top. No user migration needed.
---
## What to Build
### Step 1 — Create Supabase Project
1. Go to supabase.com → New project → name it `wwv-marketplace`
2. In **Authentication → Providers**, enable:
- **Email** (email+password + magic link)
- **Google** (add Client ID + Secret from Google Cloud Console)
- **GitHub** (add Client ID + Secret from GitHub OAuth App)
3. Save the project URL and anon key (go into `.env.local`)
### Step 2 — Install Packages
```bash
pnpm add @supabase/supabase-js @supabase/ssr
pnpm add -D prisma @prisma/client
```
### Step 3 — Configure Environment Variables
```env
# Supabase
NEXT_PUBLIC_SUPABASE_URL=https://xxxx.supabase.co
NEXT_PUBLIC_SUPABASE_ANON_KEY=eyJ...
SUPABASE_SERVICE_ROLE_KEY=eyJ... # server-only, never expose to client
# Prisma (Supabase Postgres — use the "Transaction" pooler URL)
DATABASE_URL=postgresql://postgres.[ref]:[password]@aws-0-[region].pooler.supabase.com:6543/postgres?pgbouncer=true
DIRECT_URL=postgresql://postgres.[ref]:[password]@aws-0-[region].pooler.supabase.com:5432/postgres
# Ed25519 key for plugin ticket signing (ADR-001B)
MARKETPLACE_ED25519_PRIVATE_KEY= # base64 PEM
MARKETPLACE_ED25519_KID= # key ID (for JWKS rotation)
# OAuth server (PKCE for Local App)
OAUTH_CLIENT_REDIRECT_URI_ALLOWLIST=http://localhost:3000/api/marketplace/callback,https://app.worldwideview.dev/api/marketplace/callback
```
### Step 4 — Supabase Client Helpers
**`src/lib/supabase/server.ts`** — server-side client (server components, server actions, API routes):
```ts
import { createServerClient } from "@supabase/ssr"
import { cookies } from "next/headers"
export function createClient() {
const cookieStore = cookies()
return createServerClient(
process.env.NEXT_PUBLIC_SUPABASE_URL!,
process.env.NEXT_PUBLIC_SUPABASE_ANON_KEY!,
{ cookies: { getAll: () => cookieStore.getAll(), setAll: (c) => c.forEach(({ name, value, options }) => cookieStore.set(name, value, options)) } }
)
}
```
**`src/lib/supabase/client.ts`** — browser client:
```ts
import { createBrowserClient } from "@supabase/ssr"
export const supabase = createBrowserClient(
process.env.NEXT_PUBLIC_SUPABASE_URL!,
process.env.NEXT_PUBLIC_SUPABASE_ANON_KEY!
)
```
### Step 5 — Auth Middleware (session refresh)
**`src/middleware.ts`** — refreshes Supabase session on every request:
```ts
import { createServerClient } from "@supabase/ssr"
import { NextResponse, type NextRequest } from "next/server"
export async function middleware(request: NextRequest) {
let response = NextResponse.next({ request })
const supabase = createServerClient(
process.env.NEXT_PUBLIC_SUPABASE_URL!,
process.env.NEXT_PUBLIC_SUPABASE_ANON_KEY!,
{ cookies: { getAll: () => request.cookies.getAll(), setAll: (c) => c.forEach(({ name, value, options }) => response.cookies.set(name, value, options)) } }
)
await supabase.auth.getUser() // refreshes session token
return response
}
export const config = { matcher: ["/((?!_next/static|_next/image|favicon.ico).*)"] }
```
### Step 6 — Auth Callback Route
**`src/app/auth/callback/route.ts`** — Supabase redirects here after social login to exchange the code for a session:
```ts
import { createClient } from "@/lib/supabase/server"
import { NextResponse } from "next/server"
export async function GET(request: Request) {
const { searchParams, origin } = new URL(request.url)
const code = searchParams.get("code")
const next = searchParams.get("next") ?? "/"
if (code) {
const supabase = createClient()
await supabase.auth.exchangeCodeForSession(code)
}
return NextResponse.redirect(`${origin}${next}`)
}
```
### Step 7 — Login UI
**`src/app/login/page.tsx`** — single page with four auth methods:
- Email + password (`supabase.auth.signInWithPassword()`)
- Google (`supabase.auth.signInWithOAuth({ provider: "google" })`)
- GitHub (`supabase.auth.signInWithOAuth({ provider: "github" })`)
- Magic link (`supabase.auth.signInWithOtp({ email })`)
All OAuth redirects must include `redirectTo: "${origin}/auth/callback"`.
### Step 8 — Prisma Schema for App Data
Supabase manages the `auth.users` table internally. For app-specific data, add to `prisma/schema.prisma`:
```prisma
// References Supabase auth.users UUID — do NOT store password/session data here
model MarketplaceUser {
id String @id // matches auth.users.id (UUID)
email String @unique
plan String @default("free") // "free" | "pro" | "team"
stripeCustomerId String?
createdAt DateTime @default(now())
plugins InstalledPlugin[]
}
model InstalledPlugin {
id String @id @default(cuid())
userId String
pluginId String
installedAt DateTime @default(now())
user MarketplaceUser @relation(fields: [userId], references: [id])
@@unique([userId, pluginId])
}
```
Run `prisma migrate dev` pointing at the Supabase Postgres URL.
### Step 9 — OAuth 2.0 PKCE Authorization Server (for Local App)
This is the custom layer that lets the Local App "connect" to the Marketplace. Supabase handles user identity; these routes handle the code exchange.
**`src/app/api/oauth/authorize/route.ts`**
- Validates: `client_id`, `redirect_uri` (must match allowlist), `code_challenge`, `code_challenge_method=S256`, `state`
- Calls `supabase.auth.getUser()` — if not logged in, redirects to `/login?redirect_to=...`
- Shows a consent screen ("Allow WorldWideView Local App to access your account?")
- On approval: stores `{ code, code_challenge, user_id, expires: now+60s }` in Supabase DB, redirects to `redirect_uri?code=...&state=...`
**`src/app/api/oauth/token/route.ts`**
- Validates `code` + `code_verifier` (SHA-256(`code_verifier`) must equal stored `code_challenge`)
- Issues a long-lived API Key (random 32-byte token), stores hashed in DB
- Returns `{ access_token, token_type: "Bearer" }` — this is what the Local App encrypts and stores
**`src/app/api/auth/exchange/route.ts`** *(ADR-001B)*
- Accepts API Key + `audience` (plugin ID)
- Verifies API Key against DB, checks subscription tier
- Issues a 5-minute EdDSA JWT signed with the Marketplace Ed25519 key, scoped to the requested plugin
- Data Engines verify this JWT via the public JWKS endpoint
**`src/app/.well-known/jwks.json/route.ts`**
- Returns the Marketplace's Ed25519 public key in JWKS format
- Data Engines call this on startup and cache it (ADR-001A)
---
## Critical Files (Marketplace repo: `c:/dev/worldwideview-marketplace`)
| File | Purpose |
|---|---|
| `src/lib/supabase/server.ts` | Server-side Supabase client |
| `src/lib/supabase/client.ts` | Browser-side Supabase client |
| `src/middleware.ts` | Session refresh on every request |
| `src/app/auth/callback/route.ts` | OAuth callback from Supabase (social login) |
| `src/app/login/page.tsx` | Login UI (all 4 methods) |
| `prisma/schema.prisma` | App data schema (MarketplaceUser, InstalledPlugin) |
| `src/app/api/oauth/authorize/route.ts` | PKCE auth endpoint (Local App connects here) |
| `src/app/api/oauth/token/route.ts` | PKCE token exchange |
| `src/app/api/auth/exchange/route.ts` | API Key → 5-min EdDSA JWT (ADR-001B) |
| `src/app/.well-known/jwks.json/route.ts` | Public JWKS for Data Engine JWT verification |
| `.env.local` | Supabase + Prisma + Ed25519 keys |
---
## Stripe Integration Path (Future)
When ready to add payments:
1. Install `stripe` + `@stripe/stripe-js`
2. On user signup (Supabase `auth.onAuthStateChange`), create a Stripe customer and save `stripeCustomerId` to `MarketplaceUser`
3. Use Stripe Billing for subscriptions — webhook updates `MarketplaceUser.plan`
4. The `exchange` endpoint reads `plan` to enforce tier in the issued JWT (`tier` claim per ADR-001B)
Reference: [Vercel Next.js + Supabase + Stripe starter](https://github.com/vercel/nextjs-subscription-payments)
---
---
## Risk Mitigations (Pre-Mortem — 2026-05-22)
### Tigers Addressed
**[HIGH] RLS not configured on app tables**
- All tables that contain user-scoped data (`MarketplaceUser`, `InstalledPlugin`) MUST have Row Level Security enabled in Supabase before going live.
- Add to `prisma/migrations/` as a raw SQL migration (Prisma doesn't support RLS natively):
```sql
ALTER TABLE "MarketplaceUser" ENABLE ROW LEVEL SECURITY;
CREATE POLICY "Users can only access own row" ON "MarketplaceUser"
USING (id = auth.uid());
ALTER TABLE "InstalledPlugin" ENABLE ROW LEVEL SECURITY;
CREATE POLICY "Users can only access own plugins" ON "InstalledPlugin"
USING ("userId" = auth.uid());
```
- Server-side routes that need full access use `SUPABASE_SERVICE_ROLE_KEY` (bypasses RLS by design — keep it server-only, never expose to client).
**[HIGH] Authorization code not invalidated after use**
- In `/api/oauth/token/route.ts`, wrap the validation and API Key issuance in a transaction:
1. `SELECT` the code row (validate `code_challenge`)
2. `DELETE` the code row
3. Issue the API Key
4. Commit
- If the DELETE fails or the row is already gone, return `400 invalid_grant`. This prevents replay attacks within the 60s window.
### Medium Risks Addressed
**[MEDIUM] Supabase email rate limit breaks magic link**
- Configure a custom SMTP provider in Supabase Dashboard → Project Settings → Auth → SMTP Settings before enabling magic link in production.
- Recommended: Resend (free tier = 3,000 emails/month, 100/day). Add `RESEND_API_KEY` to env and configure in Supabase dashboard — no code change needed.
**[MEDIUM] MarketplaceUser orphan on auth.users delete**
- Add a Postgres trigger via raw SQL migration:
```sql
CREATE OR REPLACE FUNCTION handle_deleted_user()
RETURNS TRIGGER LANGUAGE plpgsql SECURITY DEFINER AS $$
BEGIN
DELETE FROM "MarketplaceUser" WHERE id = OLD.id;
RETURN OLD;
END;
$$;
CREATE TRIGGER on_auth_user_deleted
AFTER DELETE ON auth.users
FOR EACH ROW EXECUTE PROCEDURE handle_deleted_user();
```
**[MEDIUM] Ed25519 key rotation — no zero-downtime strategy**
- Do not store the Ed25519 key only in an env var. Add a `SigningKey` table to Prisma schema:
```prisma
model SigningKey {
kid String @id
publicKey String // PEM, served in JWKS
privateKey String // encrypted at rest
status String // "active" | "retiring" | "revoked"
expiresAt DateTime?
createdAt DateTime @default(now())
}
```
- The JWKS endpoint serves ALL `status != "revoked"` keys. During rotation: add a new key (`active`), flip the old key to `retiring`, wait 10 minutes for Data Engine caches to expire, then flip old key to `revoked`.
### Accepted Risks
- None — all identified risks have been mitigated in plan.
---
## Verification
1. **Social login:** Click "Continue with Google" → Google OAuth flow → redirects to `/auth/callback` → lands on dashboard logged in
2. **Magic link:** Enter email → receive email → click link → logged in
3. **PKCE flow (end-to-end):** From Local App, click "Connect to Marketplace" → redirected to `/api/oauth/authorize` → must log in first if not logged in → consent screen → approve → back to Local App with `code` → Local App calls `/api/oauth/token` → receives API Key → stored encrypted in Local App DB
4. **Plugin ticket:** Local App POSTs API Key + `audience` to `/api/auth/exchange` → receives 5-min JWT → browser sends as WebSocket first message to Data Engine → engine verifies via JWKS → stream opens
5. **Unauthenticated protection:** Hit `/api/oauth/authorize` without a session → must redirect to login, not error
@@ -1,578 +0,0 @@
# Phase 4 Coolify Cutover Implementation Plan
> **For agentic workers:** REQUIRED SUB-SKILL: Use `superpowers:subagent-driven-development` (recommended) or `superpowers:executing-plans` to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
**Goal:** Flip the production WorldWideView data engine from "auth disabled" (`WWV_SKIP_WS_AUTH=true`) to "auth enforced" (`JWKS_URL` set + ticket-auth plugin allowlist enumerated) in a single atomic redeploy window, with smoke tests proving every plugin still works.
**Architecture:** Pure ops change — no source code modifications. Two Coolify services receive env-var updates (`wwv-data-engine` and `wwv` frontend), both are redeployed near-simultaneously, smoke-tested through the live UI. Rollback is a single env-var revert + redeploy.
**Tech Stack:** Coolify CLI (context `wwv` pre-configured), `curl` for registry/JWKS probing, optional SSH to the host for low-level diagnostics. No code, no tests, no commits — only env config + deploy actions.
---
## Context — why this plan exists
Phase 4 of the Northstar Sequence shipped the data engine's WebSocket JWT enforcement code (engine-side: PR #10 merged on `wwv-data-engine`; frontend ticket-issuance and PKCE flow merged earlier in worldwideview/marketplace). The enforcement is currently **dormant** in production because the Coolify env still has `WWV_SKIP_WS_AUTH=true`, which the engine reads at boot to bypass first-message auth.
Local end-to-end testing in the prior session proved the full chain works:
- PKCE → marketplace API key (encrypted in `MarketplaceCredential`)
- `/api/auth/ticket` → 5-min EdDSA JWT signed by marketplace
- WebSocket first frame: `{type:"auth", v:1, token}` → engine validates against marketplace JWKS → `welcome` → subscribe → data flows
The pre-mortem identified four tigers blocking the cutover. The companion plan (`joyful-napping-gizmo.md`) addresses three via CI hardening and the `BaseIncidentPlugin` payload fix. **This plan handles the remaining tiger (T3 — production JWKS reachability never verified) and the elephants (E1 — plugin enumeration never done, E2 — current Coolify state unknown), then performs the actual flip.**
**Prerequisites — the companion plan must be fully landed:**
- PR1 `feat/ci-coverage-phase4` merged on `wwv-data-engine` main (CI guards live)
- PR2 `test/base-incident-plugin-payload` merged on `worldwideview` main (test guards live)
- PR3 `fix/base-incident-plugin-payload` merged on `worldwideview` main (earthquakes data fix live)
- Worldwideview Coolify service has redeployed since PR3 landed (so the new frontend bundle contains the `mapWebsocketPayload` fix)
Without PR3 deployed, enabling Earthquakes in the cutover allowlist will show an empty layer — the test we ran locally would re-occur in production.
---
## File Structure
This plan modifies **zero source files**. All changes are Coolify env-var updates plus the documents created during audit:
| Path | Role | Lifecycle |
|---|---|---|
| `C:\Users\silve\.claude\plans\coolify-baseline-<timestamp>.md` | Captured baseline of current Coolify env state (rollback reference) | created in Task 3 |
| `C:\Users\silve\.claude\plans\coolify-cutover-allowlist-<timestamp>.md` | Computed plugin allowlist (CSV + rationale) | created in Task 4 |
| `worldwideview/docs/architecture/decisions/adr-0001b-status.md` *(optional)* | ADR-001B status update — Proposed → Accepted | Task 13 (post-success only) |
The Coolify services touched:
- **`wwv-data-engine`** — set `JWKS_URL`, remove `WWV_SKIP_WS_AUTH`
- **`wwv`** (the worldwideview frontend) — set `NEXT_PUBLIC_WWV_TICKET_AUTH_PLUGINS`
---
## Task 1: Confirm prerequisites are live in production
**Files:** none. Read-only verification.
- [ ] **Step 1: Verify all three companion-plan PRs are merged**
```powershell
gh pr list --repo silvertakana/wwv-data-engine --state merged --limit 5
gh pr list --repo silvertakana/worldwideview --state merged --limit 5
```
Expected: see merged PRs titled `test(auth): Phase 4 WebSocket auth + port-contract CI guards` (wwv-data-engine), `test(incidents): guard BaseIncidentPlugin WebSocket payload contract` and `fix(incidents): add mapWebsocketPayload default to BaseIncidentPlugin` (worldwideview). If any is missing, STOP — return to the companion plan.
- [ ] **Step 2: Verify the worldwideview Coolify deploy includes PR3**
```powershell
coolify deploy status wwv
```
Expected: `last deploy` timestamp is AFTER PR3 merge. If older, trigger a redeploy first:
```powershell
coolify deploy trigger wwv
```
Wait for completion (`coolify deploy status wwv --watch`). The cutover later will trigger this again, but verifying the fix is deployed BEFORE the cutover ensures we can isolate any failure to env vs. code.
- [ ] **Step 3: Confirm the data engine code already has Phase 4 (PR #10)**
```powershell
curl https://dataenginev2.worldwideview.dev/health
```
Expected: JSON with `version: "1.7.3"` (or newer if CI tests bumped further). If `1.7.1` or earlier, PR #10 is not deployed — STOP.
---
## Task 2: Verify the marketplace JWKS endpoint is reachable from the public internet
**Files:** none.
- [ ] **Step 1: Curl from the local workstation**
```powershell
curl -v https://marketplace.worldwideview.dev/api/auth/jwks
```
Expected: HTTP 200, body is JSON like `{"keys":[{"kty":"OKP","crv":"Ed25519","x":"...","kid":"marketplace-key-v1","alg":"EdDSA"}]}`. If 404/500/timeout → STOP, fix the marketplace before continuing.
- [ ] **Step 2: Validate the JWK shape**
The response must include at least one key with `kty: "OKP"`, `crv: "Ed25519"`, `alg: "EdDSA"`, and a non-empty `kid`. If any of those are missing, the engine's `verifyEngineToken` will reject every token with "no applicable key found" — same symptom as our local test isolation bug. STOP if shape is wrong.
---
## Task 3: Capture Coolify baseline (rollback reference)
**Files:**
- Create: `C:\Users\silve\.claude\plans\coolify-baseline-<YYYYMMDD-HHMMSS>.md`
- [ ] **Step 1: Dump current data engine env vars**
```powershell
coolify env list wwv-data-engine
```
- [ ] **Step 2: Dump current frontend env vars**
```powershell
coolify env list wwv
```
- [ ] **Step 3: Save baseline to file**
Create `C:\Users\silve\.claude\plans\coolify-baseline-<YYYYMMDD-HHMMSS>.md` with the exact current values of these keys (copy from the `coolify env list` output):
```markdown
# Coolify baseline — captured <ISO timestamp>
## wwv-data-engine
- JWKS_URL: <current value or "(unset)">
- WWV_SKIP_WS_AUTH: <current value or "(unset)">
- ENGINE_ID: <current value or "(unset)">
- REDIS_URL: <redacted — confirm present>
- SENTRY_DSN: <redacted — confirm present>
## wwv (frontend)
- NEXT_PUBLIC_WWV_TICKET_AUTH_PLUGINS: <current value or "(unset)">
- NEXT_PUBLIC_WWV_PLUGIN_DATA_ENGINE_URL: <current value>
- NEXT_PUBLIC_WWV_MARKETPLACE_URL: <current value>
- NEXT_PUBLIC_WWV_EDITION: <current value>
## Rollback procedure
If anything breaks during/after cutover:
1. coolify env set wwv-data-engine WWV_SKIP_WS_AUTH=true
2. coolify env unset wwv-data-engine JWKS_URL (or set to "(unset)" value above)
3. coolify env unset wwv NEXT_PUBLIC_WWV_TICKET_AUTH_PLUGINS (or restore "(unset)" value above)
4. coolify deploy trigger wwv-data-engine
5. coolify deploy trigger wwv
6. curl https://dataenginev2.worldwideview.dev/health — confirm version unchanged, engine running
```
This file is the **single source of truth** for rollback. Do not skip it.
---
## Task 4: Compute the plugin allowlist from the live data engine manifest
**Files:**
- Create: `C:\Users\silve\.claude\plans\coolify-cutover-allowlist-<YYYYMMDD-HHMMSS>.md`
- [ ] **Step 1: Fetch the live manifest from production**
```powershell
curl https://dataenginev2.worldwideview.dev/manifest
```
Expected: JSON like `{"engine":"wwv-data-engine","version":"1.7.3","plugins":["earthquakes","wildfires","sanctions","conflict-events","gps-jamming","cyber-attacks","civil-unrest","iss","satellites","aviation-history","maritime-history","iranwar-events"],"websocket":"/stream","timestamp":...}`. The exact list will vary — capture whatever is returned.
- [ ] **Step 2: Build the CSV**
Take the `plugins` array from the response and join with commas (no spaces). Example:
```
earthquakes,wildfires,sanctions,conflict-events,gps-jamming,cyber-attacks,civil-unrest,iss,satellites,aviation-history,maritime-history,iranwar-events
```
- [ ] **Step 3: Cross-check against the marketplace verified registry (sanity)**
```powershell
curl https://marketplace.worldwideview.dev/api/registry
```
Compare the engine manifest's plugin IDs against the registry's verified IDs. Note any discrepancies:
- Manifest plugins NOT in registry → in-house seeders not advertised on marketplace (fine, still need auth)
- Registry plugins NOT in manifest → marketplace-only or static plugins (fine, don't need auth)
The **engine manifest is authoritative** for the allowlist — anything that streams from this engine needs to be there.
- [ ] **Step 4: Save the computed allowlist**
Create `C:\Users\silve\.claude\plans\coolify-cutover-allowlist-<YYYYMMDD-HHMMSS>.md`:
```markdown
# Phase 4 Coolify cutover — plugin allowlist
Captured from: https://dataenginev2.worldwideview.dev/manifest at <ISO timestamp>
## Allowlist (CSV form for NEXT_PUBLIC_WWV_TICKET_AUTH_PLUGINS)
<the CSV from Step 2>
## Per-plugin rationale
| Plugin | In manifest | In registry | Action |
|---|---|---|---|
| earthquakes | yes | yes | include |
| wildfires | yes | yes | include |
| <one row per plugin> | | | |
## Notes
- Engine manifest is authoritative.
- Registry sanity-checked at Step 3 — discrepancies noted above.
```
---
## Task 5: Smoke-test JWKS reachability FROM the production data engine container
**Files:** none.
The engine's startup check (`startup-checks.ts`) verifies JWKS is reachable at boot — but a one-shot check at boot doesn't catch transient network policy issues that appear later. Pre-flight this from inside the actual container to catch firewall/DNS surprises.
- [ ] **Step 1: Exec into the container via Coolify**
```powershell
coolify exec wwv-data-engine -- curl -v https://marketplace.worldwideview.dev/api/auth/jwks
```
If `coolify exec` is not available, fall back to SSH:
```bash
ssh root@<coolify-host>
docker exec -it $(docker ps --filter "name=wwv-data-engine" -q) curl -v https://marketplace.worldwideview.dev/api/auth/jwks
```
The Coolify host IP is in your server-dev memory; if unsure, run `coolify host info`.
- [ ] **Step 2: Verify the response**
Expected: HTTP 200, same JWKS body as Task 2 Step 1. If the local workstation can reach JWKS but the container cannot, you have a Coolify network policy issue (egress firewall, DNS) — STOP and fix the network before continuing.
---
## Task 6: Set the new env vars on Coolify (no redeploy yet)
**Files:** none. Coolify state-only.
Setting env vars in Coolify does NOT trigger a deploy by default — the new values take effect at the next start. We set BOTH services' new env vars first, then trigger redeploys near-simultaneously so the atomic-flip window is small.
- [ ] **Step 1: Set `JWKS_URL` on the data engine**
```powershell
coolify env set wwv-data-engine JWKS_URL=https://marketplace.worldwideview.dev/api/auth/jwks
```
Expected: confirmation message. Verify with `coolify env get wwv-data-engine JWKS_URL`.
- [ ] **Step 2: Remove (or set false) `WWV_SKIP_WS_AUTH` on the data engine**
```powershell
coolify env unset wwv-data-engine WWV_SKIP_WS_AUTH
```
(If `unset` is not supported, use `coolify env set wwv-data-engine WWV_SKIP_WS_AUTH=false` — the engine reads `=== 'true'` so anything else is treated as enforcement-on.)
Verify with `coolify env get wwv-data-engine WWV_SKIP_WS_AUTH` — should return empty/false.
- [ ] **Step 3: Set `NEXT_PUBLIC_WWV_TICKET_AUTH_PLUGINS` on the frontend**
Use the CSV from Task 4 Step 2. Example:
```powershell
coolify env set wwv NEXT_PUBLIC_WWV_TICKET_AUTH_PLUGINS=earthquakes,wildfires,sanctions,conflict-events,gps-jamming,cyber-attacks,civil-unrest,iss,satellites,aviation-history,maritime-history,iranwar-events
```
Verify with `coolify env get wwv NEXT_PUBLIC_WWV_TICKET_AUTH_PLUGINS` — should return the CSV verbatim.
- [ ] **Step 4: Cross-check the saved values against the baseline file**
Open `C:\Users\silve\.claude\plans\coolify-baseline-<...>.md` and confirm the three values you just changed are different from baseline. This is the "I really did change them" sanity check.
---
## Task 7: Trigger redeploys for both services
**Files:** none.
Both redeploys should be triggered as close together as possible (within ~1 minute) so the user-facing flip is single-window. Coolify deploys are non-blocking; we trigger both then watch both.
- [ ] **Step 1: Trigger data engine deploy**
```powershell
coolify deploy trigger wwv-data-engine
```
Capture the deploy ID printed.
- [ ] **Step 2: Trigger frontend deploy**
```powershell
coolify deploy trigger wwv
```
Capture the deploy ID printed.
- [ ] **Step 3: Watch both deploys complete**
In one terminal:
```powershell
coolify deploy status wwv-data-engine --watch
```
In another (parallel):
```powershell
coolify deploy status wwv --watch
```
Expected: both reach `running` state within ~3-5 minutes. The data engine log will show `[Server] JWKS reachable at https://marketplace.worldwideview.dev/api/auth/jwks` — that is the moment enforcement goes live.
- [ ] **Step 4: STOP and rollback if either deploy fails**
If `wwv-data-engine` startup fails with `FATAL: JWKS unreachable at startup`, the network/JWKS smoke (Task 5) gave a false positive. Rollback immediately (Task 12) and investigate before retrying.
If `wwv` build fails, the env var format may be wrong (Coolify can sometimes mangle commas) — check the build log and re-set the env var with proper escaping.
---
## Task 8: Post-deploy smoke — engine endpoints
**Files:** none.
- [ ] **Step 1: Health check**
```powershell
curl https://dataenginev2.worldwideview.dev/health
```
Expected: `{"status":"ok","engine":"wwv-data-engine","timestamp":<recent>,"seeders":{...}}`. The `seeders` object should list every plugin in the manifest with a recent `lastSuccess` timestamp.
- [ ] **Step 2: Manifest check**
```powershell
curl https://dataenginev2.worldwideview.dev/manifest
```
Expected: same plugin list as Task 4 Step 1 (no regression). If a plugin dropped out, a seeder failed to load — investigate before continuing.
---
## Task 9: Post-deploy smoke — WebSocket auth handshake (curl-based)
**Files:** none.
Confirms enforcement is actually ON before testing the UI.
- [ ] **Step 1: Attempt unauthenticated WebSocket connection**
```powershell
# Use websocat or wscat — install via: pnpm dlx wscat
pnpm dlx wscat -c wss://dataenginev2.worldwideview.dev/stream
```
Then send (paste into the open prompt):
```
{"action":"subscribe","pluginId":"earthquakes"}
```
Expected: connection closes immediately with code `4003`. If the connection stays open and you receive data, **enforcement is NOT active** — the env var didn't apply, restart the data engine via `coolify deploy trigger wwv-data-engine` and retry.
---
## Task 10: Post-deploy smoke — UI walkthrough (the real test)
**Files:** none.
- [ ] **Step 1: Open the production app in a browser**
Navigate to `https://worldwideview.dev` (or whatever the production frontend URL is). Sign in if not already.
- [ ] **Step 2: Open browser DevTools — Network tab, filter "WS"**
You'll see WebSocket connections appear as you enable plugins.
- [ ] **Step 3: Enable EACH plugin in the allowlist, one at a time**
For each plugin in the CSV from Task 4:
1. Enable it via the UI control (sidebar, search, etc.)
2. Observe the new `wss://dataenginev2.worldwideview.dev/stream` WebSocket in the Network tab
3. Click it → Messages tab → confirm:
- First **outbound** message: `{"type":"auth","v":1,"token":"<long EdDSA JWT>"}`
- First **inbound** message: `{"type":"welcome","engine":"wwv-data-engine","plugins":[...]}`
- Subsequent inbound: `{"type":"data","pluginId":"<this plugin>","payload":...}`
4. Confirm visible markers/data appear on the globe
5. **If any plugin closes with `4003`** in the Network tab → that plugin is broken under enforcement. Capture the plugin ID, then either:
- Continue smoke-testing remaining plugins (note all failures)
- OR rollback immediately if it's a high-traffic plugin (your call)
- [ ] **Step 4: Specific check — Earthquakes**
This is the plugin we expected to break before PR3 landed. Confirm:
- WebSocket connects with auth handshake
- `data` messages arrive
- Earthquake markers appear on the globe (NOT an empty layer)
If markers don't appear but the auth handshake succeeded and `data` messages are arriving, PR3 didn't deploy — re-trigger the worldwideview deploy.
- [ ] **Step 5: Negative smoke — sign out, retry**
Sign out of the app. Try to enable a cloud plugin. Expected: WebSocket either doesn't connect or closes with `4003`/`4001` quickly. The frontend's `/api/auth/ticket` route should return 401 because no NextAuth session exists. If unauthenticated users CAN still get data, the ticket route's session check is broken — file an issue.
---
## Task 11: Decision point — happy path OR rollback
**Files:** none.
- [ ] **Step 1: Evaluate smoke results**
| Outcome | Action |
|---|---|
| All plugins green, all WebSockets auth-handshake-then-data | Skip to Task 13 (mark Phase 4 done) |
| 1-2 minor plugins break (low-traffic, niche) | Document failures, keep enforcement on, file fix issues — Task 13 with caveats |
| Earthquakes or any high-traffic plugin breaks | EXECUTE TASK 12 (rollback), then debug |
| Engine won't stay up (crashloop) | EXECUTE TASK 12 immediately, investigate logs |
The threshold for rollback is qualitative — your call. The intent is enforcement-on by default; rollback should be the exception.
---
## Task 12: Rollback procedure (conditional — only if Task 11 says so)
**Files:** none. Coolify state restoration.
- [ ] **Step 1: Re-enable bypass on the data engine**
```powershell
coolify env set wwv-data-engine WWV_SKIP_WS_AUTH=true
```
- [ ] **Step 2: Restore (or unset) `JWKS_URL` to baseline value**
Open `C:\Users\silve\.claude\plans\coolify-baseline-<...>.md` and use the recorded value:
```powershell
# If baseline value was "(unset)":
coolify env unset wwv-data-engine JWKS_URL
# If baseline had a value:
coolify env set wwv-data-engine JWKS_URL=<baseline value>
```
- [ ] **Step 3: Restore (or unset) `NEXT_PUBLIC_WWV_TICKET_AUTH_PLUGINS` to baseline value**
```powershell
# If baseline was "(unset)":
coolify env unset wwv NEXT_PUBLIC_WWV_TICKET_AUTH_PLUGINS
# If baseline had a value:
coolify env set wwv NEXT_PUBLIC_WWV_TICKET_AUTH_PLUGINS=<baseline value>
```
- [ ] **Step 4: Trigger both redeploys**
```powershell
coolify deploy trigger wwv-data-engine
coolify deploy trigger wwv
coolify deploy status wwv-data-engine --watch
coolify deploy status wwv --watch
```
- [ ] **Step 5: Confirm rollback via the same smoke as Task 9**
```powershell
pnpm dlx wscat -c wss://dataenginev2.worldwideview.dev/stream
# Send: {"action":"subscribe","pluginId":"earthquakes"}
```
Expected: connection stays open, data arrives WITHOUT sending an auth message. That's the pre-cutover bypass behavior restored.
- [ ] **Step 6: Document the rollback**
Append to the baseline file:
```markdown
## Rollback executed <ISO timestamp>
- Trigger: <what broke — e.g. "Earthquakes 4003 close in production browser test">
- Plugins broken: <list>
- Next steps: <e.g. "Investigate WsClient auth path for plugin X — file issue">
```
STOP. Do not retry the cutover until the root cause is fixed and re-tested locally.
---
## Task 13: Mark Phase 4 complete (happy path only)
**Files:**
- Modify: `worldwideview/docs/architecture/decisions/adr-0001-decentralized-plugin-auth.md` (or wherever ADR-001B lives)
- Modify: any Northstar Sequence tracker doc (likely in `worldwideview/docs/` or `.agents/context/`)
- [ ] **Step 1: Find the ADR**
```powershell
cd C:\dev\wwv\worldwideview
cymbal search --text "ADR-001B"
# OR
Get-ChildItem -Recurse docs/architecture -Filter "adr-0001*.md"
```
- [ ] **Step 2: Update the ADR status from "Proposed" to "Accepted"**
Edit the relevant ADR — typically there's a status block near the top:
```diff
-Status: Proposed
+Status: Accepted (cutover completed <YYYY-MM-DD>)
```
- [ ] **Step 3: Find and update the Northstar Sequence tracker**
```powershell
cymbal search --text "Northstar Sequence"
# OR
Grep --pattern "Phase 4" docs/
```
Update Phase 4's status from in-progress to complete. Add a one-line note referencing the cutover date.
- [ ] **Step 4: Commit on a new branch**
```powershell
git checkout main
git pull --ff-only origin main
git checkout -b docs/phase4-cutover-complete
git add docs/
git commit -m "docs: mark ADR-001B accepted and Northstar Phase 4 complete
Production data engine cutover completed <YYYY-MM-DD>. WebSocket JWT
enforcement active for: <list from allowlist>. Local + production E2E
verified. Baseline + allowlist files retained in personal notes."
git push -u origin docs/phase4-cutover-complete
gh pr create --title "docs: mark Phase 4 Coolify cutover complete" --body "Status update only — production cutover landed <YYYY-MM-DD>. See plan at .claude/plans/2026-05-24-phase4-coolify-cutover.md."
```
---
## Verification — end-to-end
After Task 13 (or Task 12 if rolled back):
1. **If happy path:** the WebSocket `/stream` endpoint rejects unauthenticated connections with `4003` (proven via wscat in Task 9), accepts authenticated UI sessions (proven in Task 10), shows earthquake markers and every other allowlisted plugin's data on the globe.
2. **If rolled back:** the baseline file documents what broke. A follow-up plan is needed to fix the broken plugin(s) before re-attempting cutover. The system is back to "auth disabled" — same security posture as before this plan started, no production regression.
The Coolify dashboard should show both services healthy and the deploy timestamps should match the cutover window.
---
## Out of scope (separate plans / future work)
- **Port-5000 codification** (engine startup warning + ADR) — purely local DX, not cutover-blocking.
- **`better-sqlite3` → `node:sqlite` refactor of `seeder-sdk`** — Docker prebuilt binaries make this non-urgent for production; local-dev DX only.
- **Silent auth logging fixes** — engine should log `[WS] Closing unauthenticated connection from <ip>: <reason>` and use a more specific close reason than `"Invalid auth message"`. Improves operator UX; track as a small follow-up issue.
- **Plugin `streamUrl` hardcoding sweep** — plugins still hardcode `wss://dataenginev2.worldwideview.dev/stream`; saved today by local-first routing in `resolveEngineUrl.ts`. Worth a sweep, not cutover-blocking.
- **Northstar Sequence Phase 5** — final cutover (link nav, docs, marketing) — starts after Phase 4 is stable in production for ~1 week.
- **Multi-engine audience separation** — currently every JWT has audience `wwv-data-engine` (broadcast). Per-engine audiences are a future ADR-001 follow-up.
---
## Assumptions
1. The Coolify CLI is installed and the `wwv` context is pre-configured (per personal memory). If not, fall back to the Coolify web dashboard for the same actions — the env-set + deploy-trigger workflow is identical, just clicked instead of typed.
2. The production data engine URL is `https://dataenginev2.worldwideview.dev`. If it's something different, substitute throughout.
3. The production marketplace URL is `https://marketplace.worldwideview.dev`. Confirmed in Task 2.
4. The production frontend URL is `https://worldwideview.dev`. If it's `app.worldwideview.dev` or similar, substitute in Task 10.
5. The data engine reads `JWKS_URL` at boot — confirmed via `startup-checks.ts` in PR #10. The cutover redeploy triggers a fresh boot.
6. The frontend reads `NEXT_PUBLIC_*` vars at BUILD time — Coolify's `deploy trigger wwv` does a rebuild, so the new env var is baked into the bundle.
7. The PKCE flow has already been used at least once on the production marketplace by an actual user, so `MarketplaceCredential` rows exist for that user. If you're testing as a fresh user, you'll need to complete `/api/marketplace/connect` once before Task 10 Step 3 — otherwise `/api/auth/ticket` will fail with "no marketplace credential."
8. Coolify deploys complete in under ~5 minutes typical. If much longer, check the Coolify host disk space — per deployment rules, builds silently die when disk fills up.
-1
View File
@@ -9,7 +9,6 @@ const config = [
"out/**",
"src/generated/**",
"public/cesium/**",
"public/plugins-local/**",
"packages/*/dist/**",
".worktrees/**",
"local-scripts/**",
-25
View File
@@ -19,7 +19,6 @@ const nextConfig: NextConfig = {
async headers() {
return [
{
// Global security headers for all routes
source: "/(.*)",
headers: [
{
@@ -52,30 +51,6 @@ const nextConfig: NextConfig = {
},
],
},
{
// Camera proxy routes serve content that must be embeddable in an <iframe>
// within the same origin. Override the global frame-ancestors 'none' and
// X-Frame-Options: DENY so the browser allows the proxy response to be
// framed by our own app (localhost:3000 / production domain).
// The upstream camera HTML already had its own X-Frame-Options stripped by
// the iframe proxy route; this override ensures our response headers don't
// re-block it.
source: "/api/camera/proxy/:path*",
headers: [
{
key: "Content-Security-Policy",
// Proxy routes serve arbitrary external HTML (YouTube embeds, camera feeds, etc.).
// Restricting style-src/script-src here would break the proxied content since we
// cannot predict which CDN domains it loads from. The meaningful security boundary
// is frame-ancestors 'self' — only our own origin can embed these proxy responses.
value: "default-src * 'unsafe-inline' 'unsafe-eval' data: blob:; frame-ancestors 'self'",
},
// Remove the DENY override — SAMEORIGIN allows our app to frame this response
{ key: "X-Frame-Options", value: "SAMEORIGIN" },
{ key: "X-Content-Type-Options", value: "nosniff" },
{ key: "Referrer-Policy", value: "strict-origin-when-cross-origin" },
],
},
];
},
+13 -15
View File
@@ -1,6 +1,6 @@
{
"name": "worldwideview",
"version": "2.31.1",
"version": "2.18.3",
"private": true,
"type": "module",
"packageManager": "[email protected]",
@@ -28,23 +28,21 @@
"dependencies": {
"@auth/prisma-adapter": "^2.11.2",
"@auth/supabase-adapter": "^1.11.2",
"@modelcontextprotocol/sdk": "^1.29.0",
"@prisma/adapter-pg": "^7.8.0",
"@prisma/client": "^7.8.0",
"@prisma/client-runtime-utils": "7.8.0",
"@sentry/nextjs": "^10.54.0",
"@supabase/ssr": "^0.10.3",
"@supabase/supabase-js": "^2.106.2",
"@sentry/nextjs": "^10.53.1",
"@supabase/supabase-js": "^2.106.0",
"@types/pg": "^8.20.0",
"@vercel/analytics": "^1.6.1",
"@worldwideview/wwv-plugin-sdk": "workspace:*",
"bcryptjs": "^3.0.3",
"cesium": "^1.141.0",
"cheerio": "^1.2.0",
"hls.js": "^1.6.16",
"insecam-api": "^0.0.3",
"ioredis": "^5.4.1",
"jose": "^6.2.3",
"lucide-react": "^1.17.0",
"lucide-react": "^1.16.0",
"next": "16.2.6",
"next-auth": "^5.0.0-beta.30",
"openid-client": "^6.8.4",
@@ -55,10 +53,10 @@
"react-player": "^3.4.0",
"resium": "1.21.1",
"satellite.js": "^7.0.1",
"stripe": "^22.2.0",
"stripe": "^22.1.1",
"tslib": "^2.8.1",
"undici": "^8.3.0",
"ws": "^8.21.0",
"ws": "^8.20.1",
"zod": "^4.4.3",
"zustand": "^5.0.13"
},
@@ -76,10 +74,10 @@
"@types/react": "^19",
"@types/react-dom": "^19",
"@types/ws": "^8.18.1",
"@typescript-eslint/eslint-plugin": "^8.60.0",
"@typescript-eslint/parser": "^8.60.0",
"@typescript-eslint/eslint-plugin": "^8.59.4",
"@typescript-eslint/parser": "^8.59.4",
"@vitejs/plugin-react": "^6.0.2",
"@vitest/coverage-v8": "4.1.7",
"@vitest/coverage-v8": "4.1.6",
"concurrently": "^9.2.1",
"copy-webpack-plugin": "^14.0.0",
"eslint": "^9",
@@ -94,10 +92,10 @@
"jsdom": "^28.1.0",
"rollup-plugin-external-globals": "^0.13.0",
"tree-kill": "^1.2.2",
"tsx": "^4.22.3",
"tsx": "^4.22.2",
"typescript": "^5",
"vite": "^8.0.14",
"vitest": "^4.1.7"
"vite": "^8.0.13",
"vitest": "^4.1.6"
},
"pnpm": {
"overrides": {
+2 -3
View File
@@ -1,6 +1,6 @@
{
"name": "@worldwideview/wwv-cli",
"version": "1.3.1",
"version": "1.2.0",
"type": "module",
"description": "WorldWideView CLI",
"main": "dist/index.js",
@@ -10,8 +10,7 @@
"scripts": {
"build": "tsc",
"dev": "tsc --watch",
"clean": "rm -rf dist",
"test": "node --test \"test/**/*.test.mjs\""
"clean": "rm -rf dist"
},
"dependencies": {
"commander": "^14.0.3",
+71 -179
View File
@@ -4,74 +4,29 @@ import fs from 'fs';
import path from 'path';
import prompts from 'prompts';
const VALID_CATEGORIES = ['aviation', 'maritime', 'space', 'weather', 'custom'];
const VALID_ARCHITECTURES = ['polling', 'websocket'];
const VALID_SEEDER_TIERS = ['community', 'private'];
const VALID_RENDER_STYLES = ['billboard', 'model', 'point'];
interface CreateOptions {
core?: boolean;
displayName?: string;
description?: string;
category?: string;
architecture?: string;
seederTier?: string;
renderStyle?: string;
yes?: boolean;
}
function assertChoice(label: string, value: string | undefined, valid: string[]): void {
if (value !== undefined && !valid.includes(value)) {
console.error(`Error: invalid --${label} "${value}". Valid values: ${valid.join(', ')}`);
process.exit(1);
}
}
export const createCommand = new Command('create')
.description('Scaffold a new WorldWideView plugin (flags skip the matching prompt; omit them for interactive mode)')
.argument('[pluginId]', 'Unique kebab-case plugin ID (e.g. my-tracker)')
.description('Interactively scaffold a new WorldWideView plugin')
.option('-c, --core', 'Create inside packages instead of local-plugins (for core contributors)')
.option('-n, --display-name <name>', 'Display name (e.g. "My Live Tracker")')
.option('-d, --description <text>', 'Short description')
.option('--category <category>', `One of: ${VALID_CATEGORIES.join(', ')}`)
.option('--architecture <architecture>', `One of: ${VALID_ARCHITECTURES.join(', ')}`)
.option('--seeder-tier <tier>', `One of: ${VALID_SEEDER_TIERS.join(', ')} (websocket only)`)
.option('--render-style <style>', `One of: ${VALID_RENDER_STYLES.join(', ')}`)
.option('-y, --yes', 'Non-interactive: fail instead of prompting for missing values')
.action(async (pluginIdArg: string | undefined, options: CreateOptions) => {
// Validate any flags supplied up front so we fail fast and consistently.
assertChoice('category', options.category, VALID_CATEGORIES);
assertChoice('architecture', options.architecture, VALID_ARCHITECTURES);
assertChoice('seeder-tier', options.seederTier, VALID_SEEDER_TIERS);
assertChoice('render-style', options.renderStyle, VALID_RENDER_STYLES);
// Build the prompt list, skipping any value already supplied via flag/arg.
// In --yes mode we never prompt; missing required values are an error.
const promptList: prompts.PromptObject[] = [];
if (!pluginIdArg) {
promptList.push({
.action(async (options) => {
const response = await prompts([
{
type: 'text',
name: 'pluginId',
message: 'What is the unique ID for your plugin? (e.g. my-tracker)',
validate: value => value.length > 0 ? true : 'Plugin ID is required',
});
}
if (options.displayName === undefined) {
promptList.push({
validate: value => value.length > 0 ? true : 'Plugin ID is required'
},
{
type: 'text',
name: 'displayName',
message: 'What is the display name? (e.g. My Live Tracker)',
initial: pluginIdArg
? pluginIdArg.replace(/-/g, ' ').replace(/\b\w/g, (l: string) => l.toUpperCase())
: undefined,
});
}
if (options.description === undefined) {
promptList.push({ type: 'text', name: 'description', message: 'Enter a short description:' });
}
if (options.category === undefined) {
promptList.push({
initial: (prev: string) => prev.replace(/-/g, ' ').replace(/\b\w/g, (l: string) => l.toUpperCase())
},
{
type: 'text',
name: 'description',
message: 'Enter a short description:'
},
{
type: 'select',
name: 'category',
message: 'Which category does this plugin belong to?',
@@ -80,78 +35,46 @@ export const createCommand = new Command('create')
{ title: 'Maritime (Ship icon)', value: 'maritime' },
{ title: 'Space (Satellite icon)', value: 'space' },
{ title: 'Weather (Cloud icon)', value: 'weather' },
{ title: 'Custom (Box icon)', value: 'custom' },
],
});
}
if (options.architecture === undefined) {
promptList.push({
{ title: 'Custom (Box icon)', value: 'custom' }
]
},
{
type: 'select',
name: 'architecture',
message: 'How will your plugin receive data?',
choices: [
{ title: 'REST Polling / Static Data (Frontend only)', value: 'polling' },
{ title: 'Real-Time WebSockets (Generates backend seeder)', value: 'websocket' },
],
});
}
if (options.seederTier === undefined) {
promptList.push({
// Only ask about seeder tier when architecture resolves to websocket.
type: (_prev, values) => {
const arch = options.architecture ?? values.architecture;
return arch === 'websocket' ? 'select' : null;
},
{ title: 'Real-Time WebSockets (Generates backend seeder)', value: 'websocket' }
]
},
{
type: (prev: string) => prev === 'websocket' ? 'select' : null,
name: 'seederTier',
message: 'Which tier should your backend seeder belong to?',
choices: [
{ title: 'Community (Open data, shared)', value: 'community' },
{ title: 'Private (Requires API keys, restricted)', value: 'private' },
],
});
}
if (options.renderStyle === undefined) {
promptList.push({
{ title: 'Private (Requires API keys, restricted)', value: 'private' }
]
},
{
type: 'select',
name: 'renderStyle',
message: 'How should entities be rendered on the 3D globe?',
choices: [
{ title: '2D Billboard (Great for static icons)', value: 'billboard' },
{ title: '3D Model with LOD (Transitions 2D to 3D)', value: 'model' },
{ title: 'Simple Point (High performance dots)', value: 'point' },
],
});
}
if (options.yes && promptList.length > 0) {
// Only count prompts that would actually fire. seederTier is irrelevant
// unless the resolved architecture is websocket, so a polling plugin with
// --yes must not be forced to supply it.
const missing = promptList
.map(p => p.name as string)
.filter(name => (name === 'seederTier' ? options.architecture === 'websocket' : true));
if (missing.length > 0) {
console.error(`Error: --yes given but these required values are missing: ${missing.join(', ')}`);
process.exit(1);
{ title: 'Simple Point (High performance dots)', value: 'point' }
]
}
}
]);
const answers = promptList.length > 0 ? await prompts(promptList) : {};
// Merge flags/args (authoritative) over interactive answers.
const pluginId: string | undefined = pluginIdArg ?? answers.pluginId;
const displayName: string = options.displayName ?? answers.displayName ?? '';
const description: string = options.description ?? answers.description ?? '';
const category: string = options.category ?? answers.category;
const architecture: string = options.architecture ?? answers.architecture;
const seederTier: string | undefined = options.seederTier ?? answers.seederTier;
const renderStyle: string = options.renderStyle ?? answers.renderStyle;
if (!pluginId) {
if (!response.pluginId) {
console.log('Plugin creation cancelled.');
return;
}
const { pluginId, displayName, description, category, architecture, seederTier, renderStyle } = response;
const targetBaseDir = options.core ? 'packages' : 'local-plugins';
const pluginDir = path.join(process.cwd(), targetBaseDir, `wwv-plugin-${pluginId}`);
@@ -162,19 +85,21 @@ export const createCommand = new Command('create')
fs.mkdirSync(pluginDir, { recursive: true });
// Determine category icon
const iconMap: Record<string, string> = {
aviation: 'Plane',
maritime: 'Ship',
space: 'Satellite',
weather: 'Cloud',
custom: 'Box',
custom: 'Box'
};
const defaultIcon = iconMap[category] || 'Box';
const streamUrlField = architecture === 'websocket'
? `\n "streamUrl": "wss://dataenginev2.worldwideview.dev/stream",`
// Package.json boilerplate
const streamUrlField = architecture === 'websocket'
? `\n "streamUrl": "wss://dataenginev2.worldwideview.dev/stream",`
: '';
const packageJsonContent = `{
"name": "@worldwideview/wwv-plugin-${pluginId}",
"version": "1.0.0",
@@ -206,6 +131,7 @@ export const createCommand = new Command('create')
}
`;
// index.ts boilerplate
const renderBoilerplate = {
billboard: `return {
type: 'billboard',
@@ -226,14 +152,14 @@ export const createCommand = new Command('create')
size: 8,
outlineColor: '#ffffff',
outlineWidth: 2
};`,
};`
}[renderStyle as string] || `return {};`;
const fetchBoilerplate = architecture === 'polling'
const fetchBoilerplate = architecture === 'polling'
? `\n async fetch(): Promise<GeoEntity[]> {
// Implement REST polling logic here
return [];
}\n`
}\n`
: '';
const indexTsContent = `import { WorldPlugin, GeoEntity, CesiumEntityOptions } from '@worldwideview/wwv-plugin-sdk';
@@ -269,75 +195,41 @@ ${fetchBoilerplate}
console.log(`\n✅ Successfully created plugin frontend at ${pluginDir}`);
// Generate seeder backend if requested. Matches the real community/private
// seeder layout: local-seeders/<tier>/packages/<pluginId>/ with a tsup build
// and a src/index.ts that exports the canonical { name, cron, fn } shape.
// Generate seeder backend if requested
if (architecture === 'websocket' && seederTier) {
const seederDir = path.join(process.cwd(), 'local-seeders', seederTier, 'packages', pluginId);
if (fs.existsSync(seederDir)) {
console.error(`Warning: Seeder directory ${seederDir} already exists. Skipping seeder scaffold.`);
} else {
fs.mkdirSync(path.join(seederDir, 'src'), { recursive: true });
const seederPackageJson = `{
"name": "@wwv-seeders/${pluginId}",
"version": "1.0.0",
"main": "dist/index.mjs",
"scripts": {
"build": "tsup"
},
"dependencies": {
"@worldwideview/seeder-sdk": "^1.0.0"
}
const seederDir = path.join(process.cwd(), 'local-seeders', seederTier, pluginId);
if (!fs.existsSync(seederDir)) {
fs.mkdirSync(seederDir, { recursive: true });
const seederContent = `// WWV Data Engine Seeder for ${pluginId}
export default async function run(context) {
console.log('Starting ${displayName} seeder...');
// Example: Emit a mock entity to the engine
setInterval(() => {
context.emit({
id: '${pluginId}-mock-1',
lat: 0,
lng: 0,
alt: 10000,
heading: 0,
velocity: 250,
properties: {
status: 'active'
}
});
}, 1000);
}
`;
const tsupConfig = `import { defineConfig } from 'tsup';
export default defineConfig({
entry: ['src/index.ts'],
format: ['esm'],
clean: true,
shims: true,
noExternal: [/@wwv-seeders\\/.*/],
external: [/^(?!@wwv-seeders)[a-z@].*/],
});
`;
// The seeder's exported `name` MUST equal the frontend plugin id
// (kebab-case). See ADR-0002 / data-engine-architecture.md section 6.
const seederIndexTs = `import { setLiveSnapshot, fetchWithTimeout, withRetry } from '@worldwideview/seeder-sdk';
async function seed${pluginId.replace(/-/g, '')}() {
// TODO: fetch from the real data source and map into GeoEntity-shaped items.
const items: Array<{ id: string; lat: number; lon: number; name: string }> = [];
await setLiveSnapshot('${pluginId}', {
source: '${pluginId}',
fetchedAt: new Date().toISOString(),
items,
totalCount: items.length,
}, 3600);
}
export default {
name: '${pluginId}', // MUST match the frontend plugin id exactly (kebab-case)
cron: '*/15 * * * *',
fn: seed${pluginId.replace(/-/g, '')},
};
`;
fs.writeFileSync(path.join(seederDir, 'package.json'), seederPackageJson);
fs.writeFileSync(path.join(seederDir, 'tsup.config.ts'), tsupConfig);
fs.writeFileSync(path.join(seederDir, 'src', 'index.ts'), seederIndexTs);
fs.writeFileSync(path.join(seederDir, 'seeder.mjs'), seederContent);
console.log(`✅ Successfully created plugin seeder at ${seederDir}`);
}
}
console.log(`\nNext steps:
1. Run \`pnpm install\` from the project root.
2. Edit \`src/index.ts\` in your plugin directory.
${architecture === 'websocket' && seederTier ? `3. Edit \`src/index.ts\` in local-seeders/${seederTier}/packages/${pluginId}/, then \`pnpm build\` it.` : ''}
2. Edit \`index.ts\` in your plugin directory.
${architecture === 'websocket' ? '3. Edit `seeder.mjs` in your local-seeders directory.' : ''}
`);
});
+1 -1
View File
@@ -9,7 +9,7 @@ const program = new Command();
program
.name('wwv')
.description('WorldWideView Plugin CLI')
.version('1.3.1');
.version('1.1.2');
program.addCommand(createCommand);
program.addCommand(publishCommand);
-219
View File
@@ -1,219 +0,0 @@
// Integration tests for `wwv create`.
//
// These spawn the *built* CLI (dist/index.js) into a throwaway temp directory
// and assert the generated file tree. No mocking — this is the real scaffold
// path agents and humans use. Run with: `node --test test/` (the CLI must be
// built first; the cli.yml workflow runs `pnpm build` before this).
import { test } from 'node:test';
import assert from 'node:assert/strict';
import { spawnSync } from 'node:child_process';
import { mkdtempSync, readFileSync, existsSync, rmSync } from 'node:fs';
import { tmpdir } from 'node:os';
import { join, resolve } from 'node:path';
const CLI = resolve(import.meta.dirname, '..', 'dist', 'index.js');
function freshDir() {
return mkdtempSync(join(tmpdir(), 'wwv-cli-test-'));
}
function runCreate(cwd, args) {
const res = spawnSync(process.execPath, [CLI, 'create', ...args], {
cwd,
encoding: 'utf8',
// Empty stdin so a stray prompt (a bug) fails fast instead of hanging CI.
input: '',
});
return { status: res.status, stdout: res.stdout ?? '', stderr: res.stderr ?? '' };
}
function readPkg(dir, ...segments) {
return JSON.parse(readFileSync(join(dir, ...segments), 'utf8'));
}
function readText(dir, ...segments) {
return readFileSync(join(dir, ...segments), 'utf8');
}
// --- Happy-path plugin types -------------------------------------------------
test('polling plugin: frontend only, no seeder, no streamUrl, has fetch()', () => {
const dir = freshDir();
try {
const r = runCreate(dir, [
'poll-test', '-n', 'Poll Test', '-d', 'A polling plugin',
'--category', 'custom', '--architecture', 'polling', '--render-style', 'point', '--yes',
]);
assert.equal(r.status, 0, r.stderr);
const pkg = readPkg(dir, 'local-plugins', 'wwv-plugin-poll-test', 'package.json');
assert.equal(pkg.name, '@worldwideview/wwv-plugin-poll-test');
assert.equal(pkg.worldwideview.pluginId, 'poll-test');
assert.equal(pkg.worldwideview.category, 'custom');
assert.equal(pkg.worldwideview.streamUrl, undefined, 'polling plugin must not declare streamUrl');
const idx = readText(dir, 'local-plugins', 'wwv-plugin-poll-test', 'src', 'index.ts');
assert.match(idx, /id = 'poll-test'/);
assert.match(idx, /getPollingInterval\(\): number \{\s*return 10000/);
assert.match(idx, /async fetch\(/);
assert.match(idx, /type: 'point'/);
assert.ok(!existsSync(join(dir, 'local-seeders')), 'polling plugin must not scaffold a seeder');
} finally {
rmSync(dir, { recursive: true, force: true });
}
});
test('websocket + community + billboard: seeder under community/packages, streamUrl, Plane icon', () => {
const dir = freshDir();
try {
const r = runCreate(dir, [
'ws-comm', '-n', 'WS Comm', '-d', 'desc',
'--category', 'aviation', '--architecture', 'websocket', '--seeder-tier', 'community',
'--render-style', 'billboard', '--yes',
]);
assert.equal(r.status, 0, r.stderr);
const pkg = readPkg(dir, 'local-plugins', 'wwv-plugin-ws-comm', 'package.json');
assert.equal(pkg.worldwideview.icon, 'Plane');
assert.ok(pkg.worldwideview.streamUrl, 'websocket plugin must declare a streamUrl');
const seederBase = join(dir, 'local-seeders', 'community', 'packages', 'ws-comm');
assert.ok(existsSync(join(seederBase, 'package.json')), 'seeder package.json missing');
assert.ok(existsSync(join(seederBase, 'tsup.config.ts')), 'seeder tsup.config.ts missing');
const seederPkg = readPkg(dir, 'local-seeders', 'community', 'packages', 'ws-comm', 'package.json');
assert.equal(seederPkg.name, '@wwv-seeders/ws-comm');
assert.equal(seederPkg.main, 'dist/index.mjs');
const seederIdx = readText(dir, 'local-seeders', 'community', 'packages', 'ws-comm', 'src', 'index.ts');
assert.match(seederIdx, /name: 'ws-comm'/, "seeder name must equal the plugin id (ADR-0002)");
assert.match(seederIdx, /cron:/);
const idx = readText(dir, 'local-plugins', 'wwv-plugin-ws-comm', 'src', 'index.ts');
assert.match(idx, /getPollingInterval\(\): number \{\s*return 0;/);
assert.match(idx, /type: 'billboard'/);
assert.doesNotMatch(idx, /async fetch\(/, 'websocket plugin should not generate a fetch() stub');
} finally {
rmSync(dir, { recursive: true, force: true });
}
});
test('websocket + private + model: seeder under private/packages, Satellite icon, model render', () => {
const dir = freshDir();
try {
const r = runCreate(dir, [
'ws-priv', '-n', 'WS Priv', '-d', 'desc',
'--category', 'space', '--architecture', 'websocket', '--seeder-tier', 'private',
'--render-style', 'model', '--yes',
]);
assert.equal(r.status, 0, r.stderr);
assert.ok(
existsSync(join(dir, 'local-seeders', 'private', 'packages', 'ws-priv', 'src', 'index.ts')),
'private seeder must land under local-seeders/private/packages/',
);
const pkg = readPkg(dir, 'local-plugins', 'wwv-plugin-ws-priv', 'package.json');
assert.equal(pkg.worldwideview.icon, 'Satellite');
const idx = readText(dir, 'local-plugins', 'wwv-plugin-ws-priv', 'src', 'index.ts');
assert.match(idx, /modelUrl/, 'model render style should reference modelUrl');
} finally {
rmSync(dir, { recursive: true, force: true });
}
});
test('--core flag scaffolds into packages/ instead of local-plugins/', () => {
const dir = freshDir();
try {
const r = runCreate(dir, [
'core-test', '--core', '-n', 'Core', '-d', 'desc',
'--category', 'custom', '--architecture', 'polling', '--render-style', 'point', '--yes',
]);
assert.equal(r.status, 0, r.stderr);
assert.ok(existsSync(join(dir, 'packages', 'wwv-plugin-core-test', 'package.json')));
assert.ok(!existsSync(join(dir, 'local-plugins', 'wwv-plugin-core-test')));
} finally {
rmSync(dir, { recursive: true, force: true });
}
});
test('multi-word id: kebab id preserved, class and seeder fn names strip dashes', () => {
const dir = freshDir();
try {
const r = runCreate(dir, [
'multi-word-plugin', '-n', 'Multi Word', '-d', 'desc',
'--category', 'custom', '--architecture', 'websocket', '--seeder-tier', 'community',
'--render-style', 'point', '--yes',
]);
assert.equal(r.status, 0, r.stderr);
const idx = readText(dir, 'local-plugins', 'wwv-plugin-multi-word-plugin', 'src', 'index.ts');
assert.match(idx, /class multiwordpluginPlugin/);
assert.match(idx, /id = 'multi-word-plugin'/);
const seederIdx = readText(dir, 'local-seeders', 'community', 'packages', 'multi-word-plugin', 'src', 'index.ts');
assert.match(seederIdx, /name: 'multi-word-plugin'/);
assert.match(seederIdx, /function seedmultiwordplugin\(/);
} finally {
rmSync(dir, { recursive: true, force: true });
}
});
test('category icon mapping: maritime -> Ship, weather -> Cloud', () => {
const dir = freshDir();
try {
const base = ['-d', 'desc', '--architecture', 'polling', '--render-style', 'point', '--yes'];
assert.equal(runCreate(dir, ['sea', '-n', 'Sea', '--category', 'maritime', ...base]).status, 0);
assert.equal(runCreate(dir, ['sky', '-n', 'Sky', '--category', 'weather', ...base]).status, 0);
assert.equal(readPkg(dir, 'local-plugins', 'wwv-plugin-sea', 'package.json').worldwideview.icon, 'Ship');
assert.equal(readPkg(dir, 'local-plugins', 'wwv-plugin-sky', 'package.json').worldwideview.icon, 'Cloud');
} finally {
rmSync(dir, { recursive: true, force: true });
}
});
// --- Failure / guard cases ---------------------------------------------------
test('--yes with a missing required value exits non-zero and creates nothing', () => {
const dir = freshDir();
try {
const r = runCreate(dir, ['fail-test', '--architecture', 'websocket', '--yes']);
assert.notEqual(r.status, 0, 'should fail when required values are missing');
assert.match(r.stderr, /missing/i);
assert.ok(!existsSync(join(dir, 'local-plugins', 'wwv-plugin-fail-test')), 'must not create a partial scaffold');
} finally {
rmSync(dir, { recursive: true, force: true });
}
});
test('invalid --category exits non-zero with a clear message', () => {
const dir = freshDir();
try {
const r = runCreate(dir, [
'bad-cat', '-n', 'x', '-d', 'y', '--category', 'nonsense',
'--architecture', 'polling', '--render-style', 'point', '--yes',
]);
assert.notEqual(r.status, 0);
assert.match(r.stderr, /invalid --category/i);
} finally {
rmSync(dir, { recursive: true, force: true });
}
});
test('creating into an existing plugin directory exits non-zero', () => {
const dir = freshDir();
try {
const args = [
'dup-test', '-n', 'x', '-d', 'y', '--category', 'custom',
'--architecture', 'polling', '--render-style', 'point', '--yes',
];
assert.equal(runCreate(dir, args).status, 0);
const second = runCreate(dir, args);
assert.notEqual(second.status, 0);
assert.match(second.stderr, /already exists/i);
} finally {
rmSync(dir, { recursive: true, force: true });
}
});
-32
View File
@@ -52,38 +52,6 @@ export class MyPlugin implements WorldPlugin {
}
```
## Property Tag Helpers
Wrap entity property values with these helpers so the WorldWideView Intel panel renders them as rich UI elements instead of plain text.
```ts
import { dtProp, urlProp, imageProp, videoProp } from "@worldwideview/wwv-plugin-sdk";
```
| Helper | Wraps | Panel renders as |
|---|---|---|
| `dtProp(iso: string \| null)` | ISO 8601 date string | Expandable datetime row — local time collapsed, UTC + relative time expanded |
| `urlProp(href: string \| null)` | Any URL | Clickable link with external-link icon |
| `imageProp(src: string \| null)` | Image URL | Inline thumbnail |
| `videoProp(href: string \| null)` | Video/stream URL | "Watch" link with play icon |
All helpers are **null-safe**: passing `null`, `undefined`, or `""` returns `null`, which the panel skips cleanly.
```ts
// In your entity mapper:
properties: {
last_seen: dtProp(item.timestamp ?? null),
source_url: urlProp(item.url ?? null),
preview: imageProp(item.image_url ?? null),
live_stream: videoProp(item.stream ?? null),
// Plain values need no wrapper
name: item.name,
severity: item.severity,
}
```
Values without a tag prefix fall back to plain-text rendering, so existing plugins continue to work unchanged.
## Changelog
- **v1.0.3** — Added README with core interfaces and usage docs.
+1 -54
View File
@@ -66,7 +66,7 @@ export function createSvgIconUrl(
}
// ─── Re-export manifest types ─────────────────────────────────
export type { PluginManifest, PluginFormat, PluginType, TrustTier, PluginCapability, DataSourceConfig, FieldMapping, RenderingConfig, McpToolDeclaration } from "./manifest";
export type { PluginManifest, PluginFormat, PluginType, TrustTier, PluginCapability, DataSourceConfig, FieldMapping, RenderingConfig } from "./manifest";
// ─── Categories ──────────────────────────────────────────────
export type PluginCategory =
@@ -286,17 +286,6 @@ export interface WorldPlugin {
requiresConfiguration?(settings: unknown): boolean;
/** Map raw websocket payload into GeoEntity array. Optional existingEntities is provided so plugins can merge state (e.g. historical trails). */
mapWebsocketPayload?(payload: any, existingEntities?: GeoEntity[]): GeoEntity[];
/**
* Execute an MCP tool in the browser on behalf of the server relay.
* The server dispatches the invocation here; the plugin runs the logic
* and returns the result. The server NEVER executes plugin tools directly
* (v3 frontend-relay design).
*
* @param toolName - The bare tool name (not namespaced).
* @param args - Validated arguments from the MCP client.
* @returns Arbitrary result serializable to JSON.
*/
executeMcpTool?(toolName: string, args: Record<string, unknown>): Promise<unknown>;
}
// ─── Aliases for backwards compatibility ─────────────────────
@@ -324,45 +313,3 @@ export * from "./viteGlobals";
// ─── Auth Contracts ──────────────────────────────────────────
export * from "./auth-contracts";
// ─── Tagged Property Helpers ─────────────────────────────────
/**
* Wraps an ISO 8601 date string for rich rendering in the Intel panel.
* The panel displays local time (collapsed) and UTC + relative time (expanded).
* @param iso - ISO 8601 string (e.g. "2026-06-01T05:00:00Z"), or null/undefined
* @returns Tagged string `"datetime:{iso}"`, or `null` if input is empty
*/
export function dtProp(iso: string | undefined | null): string | null {
if (!iso) return null;
return `datetime:${iso}`;
}
/**
* Wraps a URL for rich rendering in the Intel panel as a clickable external link.
* @param href - Any URL string, or null/undefined
* @returns Tagged string `"url:{href}"`, or `null` if input is empty
*/
export function urlProp(href: string | undefined | null): string | null {
if (!href) return null;
return `url:${href}`;
}
/**
* Wraps an image URL for rich rendering in the Intel panel as an inline thumbnail.
* @param src - Image URL string, or null/undefined
* @returns Tagged string `"image:{src}"`, or `null` if input is empty
*/
export function imageProp(src: string | undefined | null): string | null {
if (!src) return null;
return `image:${src}`;
}
/**
* Wraps a video or stream URL for rich rendering in the Intel panel as a "Watch" link.
* @param href - Video or stream URL string, or null/undefined
* @returns Tagged string `"video:{href}"`, or `null` if input is empty
*/
export function videoProp(href: string | undefined | null): string | null {
if (!href) return null;
return `video:${href}`;
}
-38
View File
@@ -63,32 +63,6 @@ export interface RenderingConfig {
maxEntities?: number;
}
/**
* @interface McpToolDeclaration
* @description A single MCP tool declared by a plugin.
* The server uses this declaration to compose tools/list and dispatch
* invocations to the browser. The server NEVER executes plugin tools
* directly (v3 frontend-relay design).
*
* INVARIANT: No `execution` field. The browser (WorldPlugin.executeMcpTool)
* is the sole execution site.
*/
export interface McpToolDeclaration {
/** Safe identifier. Only [a-zA-Z0-9_-] characters are allowed. */
name: string;
/** Human-readable description for MCP clients. */
description: string;
/**
* Minimal JSON-schema-like object describing the tool arguments.
* Supports: type, properties, required, enum (per validateToolArgs).
*/
inputSchema: {
type: "object";
properties?: Record<string, { type: string; enum?: string[] }>;
required?: string[];
};
}
/**
* @interface PluginManifest
* @description The structural definition of a plugin.json file.
@@ -113,16 +87,4 @@ export interface PluginManifest {
entry?: string;
assets?: string[];
extends?: string[];
/**
* MCP tools this plugin declares (v3 frontend-relay design).
* The server reads these to compose tools/list and dispatch invocations
* to the browser. Execution always happens in the browser via
* WorldPlugin.executeMcpTool.
*/
mcpTools?: McpToolDeclaration[];
/**
* Opaque capability tags for MCP clients (e.g. "point-layer", "camera-control").
* Must be a string array when present.
*/
mcpCapabilities?: string[];
}
+8 -40
View File
@@ -1,11 +1,4 @@
import { defineConfig, devices } from '@playwright/test';
import fs from 'fs';
import path from 'path';
import { fileURLToPath } from 'url';
const __dir = path.dirname(fileURLToPath(import.meta.url));
const MARKETPLACE_DIR = path.resolve(__dir, '../worldwideview-marketplace');
const hasMarketplace = fs.existsSync(MARKETPLACE_DIR);
/**
* See https://playwright.dev/docs/test-configuration.
@@ -18,16 +11,6 @@ export default defineConfig({
globalSetup: './tests/global.setup.ts',
globalTeardown: './tests/global.teardown.ts',
testDir: './tests',
// web-auth.spec.ts targets worldwideview-web (https://wwv.local:3001) — use playwright.web.config.ts
// marketplace-from-instance.spec.ts uses playwright.marketplace.config.ts; full-flow requires marketplace repo
// marketplace-redirect-handshake.spec.ts and marketplace-sign-out.spec.ts require marketplace.wwv.local:3002
// — use playwright.cross-app.config.ts for these cross-origin handshake tests
testIgnore: [
'**/web-auth.spec.ts',
'**/marketplace-from-instance.spec.ts',
'**/marketplace-redirect-handshake.spec.ts',
'**/marketplace-sign-out.spec.ts',
],
/* Run tests in files in parallel */
fullyParallel: true,
/* Fail the build on CI if you accidentally left test.only in the source code. */
@@ -78,27 +61,12 @@ export default defineConfig({
},
],
/* Run dev servers before starting the tests */
webServer: [
{
command: 'pnpm dev',
env: {
PORT: '3001',
NEXT_PUBLIC_MARKETPLACE_URL: 'http://localhost:3002',
},
url: 'http://localhost:3001',
reuseExistingServer: !process.env.CI,
timeout: 120 * 1000,
},
// Only start marketplace when the repo is checked out alongside worldwideview.
// In CI the marketplace dir does not exist — omitting it prevents spawn ENOENT.
...(hasMarketplace ? [{
command: 'pnpm dev',
cwd: MARKETPLACE_DIR,
env: { PORT: '3002' },
url: 'http://localhost:3002',
reuseExistingServer: !process.env.CI,
timeout: 120 * 1000,
}] : []),
],
/* Run your local dev server before starting the tests */
webServer: {
command: 'pnpm dev',
env: { PORT: '3001' },
url: 'http://localhost:3001',
reuseExistingServer: !process.env.CI,
timeout: 120 * 1000,
},
});
-85
View File
@@ -1,85 +0,0 @@
/**
* Cross-app E2E config.
* Starts both worldwideview-web (3001) and worldwideview-marketplace (3002)
* and runs specs that exercise the shared Supabase cookie handshake across
* the .wwv.local cookie domain.
*
* The marketplace MUST be reachable at https://marketplace.wwv.local:3002
* (not localhost:3002) because the Supabase auth cookie is scoped to .wwv.local.
*
* Loads .env.test credentials via Node fs so specs can read them from
* process.env without requiring a dotenv package install.
*
* Usage:
* pnpm exec playwright test --config=playwright.cross-app.config.ts
*/
import { defineConfig, devices } from '@playwright/test';
import path from 'path';
import { fileURLToPath } from 'url';
import fs from 'fs';
const __dir = path.dirname(fileURLToPath(import.meta.url));
const WEB_DIR = path.resolve(__dir, '../worldwideview-web');
const MARKETPLACE_DIR = path.resolve(__dir, '../worldwideview-marketplace');
// Load .env.test credentials into process.env so specs can read them.
// Uses fs directly to avoid a dotenv dependency.
const envTestPath = path.resolve(__dir, '.env.test');
if (fs.existsSync(envTestPath)) {
const lines = fs.readFileSync(envTestPath, 'utf8').split('\n');
for (const line of lines) {
const trimmed = line.trim();
if (!trimmed || trimmed.startsWith('#')) continue;
const eqIdx = trimmed.indexOf('=');
if (eqIdx === -1) continue;
const key = trimmed.slice(0, eqIdx).trim();
const val = trimmed.slice(eqIdx + 1).trim();
if (key && !(key in process.env)) process.env[key] = val;
}
}
export default defineConfig({
timeout: 60000,
expect: { timeout: 10000 },
testDir: './tests',
testMatch: [
'**/marketplace-redirect-handshake.spec.ts',
'**/marketplace-sign-out.spec.ts',
],
fullyParallel: false,
retries: 0,
workers: 1,
reporter: 'list',
use: {
baseURL: 'https://marketplace.wwv.local:3002',
ignoreHTTPSErrors: true,
trace: 'on-first-retry',
},
projects: [
{
name: 'chromium',
use: { ...devices['Desktop Chrome'] },
},
],
webServer: [
{
command: 'pnpm dev',
cwd: WEB_DIR,
url: 'https://wwv.local:3001',
reuseExistingServer: true,
timeout: 90 * 1000,
ignoreHTTPSErrors: true,
},
{
command: 'pnpm dev',
cwd: MARKETPLACE_DIR,
url: 'https://marketplace.wwv.local:3002',
reuseExistingServer: true,
timeout: 90 * 1000,
ignoreHTTPSErrors: true,
},
],
});
-46
View File
@@ -1,46 +0,0 @@
/**
* Marketplace-only E2E config.
* Runs against the marketplace dev server on port 3002 only.
* No worldwideview auth or Docker/Postgres required.
*
* Usage:
* pnpm exec playwright test --config=playwright.marketplace.config.ts
*/
import { defineConfig, devices } from '@playwright/test';
import path from 'path';
import { fileURLToPath } from 'url';
const __dir = path.dirname(fileURLToPath(import.meta.url));
const MARKETPLACE_DIR = path.resolve(__dir, '../worldwideview-marketplace');
export default defineConfig({
timeout: 30000,
expect: { timeout: 10000 },
testDir: './tests',
testMatch: '**/marketplace-from-instance.spec.ts',
fullyParallel: false,
retries: 0,
workers: 1,
reporter: 'list',
use: {
baseURL: 'http://localhost:3002',
trace: 'on-first-retry',
},
projects: [
{
name: 'chromium',
use: { ...devices['Desktop Chrome'] },
},
],
webServer: {
command: 'pnpm dev',
cwd: MARKETPLACE_DIR,
env: { PORT: '3002' },
url: 'http://localhost:3002',
reuseExistingServer: true,
timeout: 90 * 1000,
},
});
-47
View File
@@ -1,47 +0,0 @@
/**
* worldwideview-web E2E config.
* Runs against the web dev server on https://wwv.local:3001.
* Requires mkcert cert at worldwideview-web/certs/wwv.local+2.pem.
*
* Usage:
* pnpm exec playwright test --config=playwright.web.config.ts
*/
import { defineConfig, devices } from '@playwright/test';
import path from 'path';
import { fileURLToPath } from 'url';
const __dir = path.dirname(fileURLToPath(import.meta.url));
const WEB_DIR = path.resolve(__dir, '../worldwideview-web');
export default defineConfig({
timeout: 30000,
expect: { timeout: 10000 },
testDir: './tests',
testMatch: '**/web-auth.spec.ts',
fullyParallel: false,
retries: 0,
workers: 1,
reporter: 'list',
use: {
baseURL: 'https://wwv.local:3001',
ignoreHTTPSErrors: true,
trace: 'on-first-retry',
},
projects: [
{
name: 'chromium',
use: { ...devices['Desktop Chrome'] },
},
],
webServer: {
command: 'pnpm dev',
cwd: WEB_DIR,
url: 'https://wwv.local:3001',
reuseExistingServer: true,
timeout: 90 * 1000,
ignoreHTTPSErrors: true,
},
});
+1036 -2990
View File
File diff suppressed because it is too large. Load diff
-15
View File
@@ -17,7 +17,6 @@ model User {
createdAt DateTime @default(now())
favorites Favorite[]
memberships WorkspaceMember[]
apiKeys UserApiKey[]
@@unique([tenantId, email])
@@map("users")
@@ -103,17 +102,3 @@ model MarketplaceCredential {
@@unique([tenantId])
@@map("marketplace_credentials")
}
model UserApiKey {
id String @id @default(uuid())
userId String
tenantId String?
prefix String @unique
hashedSecret String
name String
createdAt DateTime @default(now())
lastUsedAt DateTime?
user User @relation(fields: [userId], references: [id], onDelete: Cascade)
@@map("user_api_keys")
}
+3 -18
View File
@@ -63,22 +63,7 @@ console.log(`🔌 Assigned deterministic database port: ${port}`);
// Robust rewrite of DATABASE_URL in .env
const envPath = path.resolve(cwd, '.env');
// Ensure .env exists — worktrees only have .env.local, not .env.
// Without this, the DATABASE_URL injection block below is silently skipped
// and safe-db-push.mjs fails with "DATABASE_URL is missing".
if (!fs.existsSync(envPath)) {
const examplePath = path.resolve(cwd, '.env.example');
if (fs.existsSync(examplePath)) {
fs.copyFileSync(examplePath, envPath);
console.log('📄 Created .env from .env.example (worktree first-run).');
} else {
fs.writeFileSync(envPath, '');
console.log('📄 Created empty .env (worktree first-run).');
}
}
{
if (fs.existsSync(envPath)) {
const envContent = fs.readFileSync(envPath, 'utf8');
const lines = envContent.split(/\r?\n/);
@@ -111,7 +96,7 @@ if (!fs.existsSync(envPath)) {
if (linesModified) {
const newContent = newLines.join('\n');
// File write with retries and telemetry
const maxRetries = 3;
let attempt = 0;
@@ -136,7 +121,7 @@ if (!fs.existsSync(envPath)) {
}
}
}
} // end DATABASE_URL injection block
}
console.log('🚀 Checking local PostgreSQL database...');
-1
View File
@@ -34,7 +34,6 @@ const loadEnv = (file) => {
};
loadEnv('.env');
loadEnv('.env.local'); // .env.local secrets take precedence and must be in process.env before workers spawn
const teardownDbOnExit = process.env.WWV_TEARDOWN_DB_ON_EXIT === 'true' || process.env.WWV_TEARDOWN_DB_ON_EXIT === '1';
+2 -3
View File
@@ -130,9 +130,8 @@ const ANTI_PATTERNS = [
{
id: 'mdc-ref',
tier: 'A',
globs: ['*.ts', '*.tsx'],
pattern: '[\'"].*\\.mdc[\'"]',
excludePrefix: 'scripts/',
globs: ['*.ts', '*.tsx', '*.md', '*.mjs'],
pattern: '\\.mdc',
desc: '.mdc file reference — must be .md (project invariant: never .mdc files)',
},
{
+4 -4
View File
@@ -13,10 +13,10 @@ import { resolve } from "path";
const ROOT = resolve(import.meta.dirname, "..");
const EXAMPLE = resolve(ROOT, ".env.example");
const TARGET = resolve(ROOT, ".env.local");
const TARGET = resolve(ROOT, ".env");
if (existsSync(TARGET)) {
console.log("✅ .env.local already exists — skipping setup.");
console.log("✅ .env already exists — skipping setup.");
console.log(" Delete it and re-run if you want to regenerate.");
process.exit(0);
}
@@ -56,6 +56,6 @@ content = content
writeFileSync(TARGET, content, "utf8");
console.log("✅ .env.local created with generated AUTH_SECRET and ENCRYPTION_MASTER_KEY.");
console.log("✅ .env created with generated AUTH_SECRET and ENCRYPTION_MASTER_KEY.");
console.log(" Fill in any optional API keys (Cesium, Bing, OpenSky, etc.)");
console.log(" then run: pnpm dev");
console.log(" then run: npm run dev");
+2 -21
View File
@@ -251,22 +251,10 @@ export async function buildPlugin({ dir, manifest, pkg, pluginDir }) {
external: Object.keys(EXTERNAL_GLOBALS),
output: {
globals: EXTERNAL_GLOBALS,
inlineDynamicImports: true,
codeSplitting: false,
banner: '"use client";',
},
plugins: [
{
name: 'replace-process-env',
transform(code) {
if (!code.includes('process.env.NODE_ENV')) return null;
return {
code: code.replace(/process\.env\.NODE_ENV/g, '"production"'),
map: null,
};
},
},
(await import("rollup-plugin-external-globals")).default(EXTERNAL_GLOBALS),
],
plugins: [(await import("rollup-plugin-external-globals")).default(EXTERNAL_GLOBALS)],
onwarn(warning, warn) {
if (warning.code === 'MODULE_LEVEL_DIRECTIVE' && warning.message.includes('"use client"')) {
return;
@@ -304,13 +292,6 @@ export function syncToPublic({ dir, manifest, pluginDir }, { quiet = false } = {
if (fs.existsSync(distMap)) {
fs.copyFileSync(distMap, path.join(targetDir, "frontend.mjs.map"));
}
// Copy any CSS files emitted alongside the bundle (e.g. from gridstack imports)
const distDir = path.join(pluginDir, "dist");
for (const f of fs.readdirSync(distDir)) {
if (f.endsWith(".css")) {
fs.copyFileSync(path.join(distDir, f), path.join(targetDir, f));
}
}
// Generate plugin.json manifest for the marketplace load route
const pluginJson = {
-91
View File
@@ -1,91 +0,0 @@
import {
describe, it, expect, vi, beforeEach
} from "vitest";
import type { Session } from "next-auth";
import { DELETE } from "./route";
import { auth } from "@/lib/auth";
import { prisma } from "@/lib/db";
vi.mock("@/lib/auth", () => ({
auth: vi.fn(),
}));
vi.mock("@/lib/db", () => ({
prisma: {
userApiKey: {
findMany: vi.fn(),
count: vi.fn(),
create: vi.fn(),
deleteMany: vi.fn(),
},
},
}));
vi.mock("@/core/edition", () => ({
isDemo: false,
}));
// NextAuth v5 `auth` is overloaded (middleware wrapper + no-arg session getter).
// Narrow to the session-getter signature so vi.mocked resolves the correct overload.
const mockAuth = vi.mocked(auth as unknown as () => Promise<Session | null>);
// ---------------------------------------------------------------------------
// DELETE /api/api-keys/[id] — KEY-03 (revoke)
// ---------------------------------------------------------------------------
describe("DELETE /api/api-keys/[id]", () => {
beforeEach(() => {
vi.resetAllMocks();
mockAuth.mockResolvedValue({
user: { id: "user-123", email: "[email protected]" },
} as Session);
});
it("returns 200 { success: true } when user owns the key (KEY-03)", async () => {
vi.mocked(prisma.userApiKey.deleteMany).mockResolvedValue({ count: 1 } as never);
const req = new Request("http://localhost/api/api-keys/key-id-1");
const res = await DELETE(req, { params: Promise.resolve({ id: "key-id-1" }) });
const body = await res.json();
expect(res.status).toBe(200);
expect(body).toEqual({ success: true });
});
it("returns 404 { error: 'not_found' } when key belongs to another user (KEY-03)", async () => {
// deleteMany returns count: 0 when ownership check fails
vi.mocked(prisma.userApiKey.deleteMany).mockResolvedValue({ count: 0 } as never);
const req = new Request("http://localhost/api/api-keys/foreign-key-id");
const res = await DELETE(req, { params: Promise.resolve({ id: "foreign-key-id" }) });
const body = await res.json();
expect(res.status).toBe(404);
expect(body.error).toBe("not_found");
});
it("returns 401 when no session (KEY-03)", async () => {
mockAuth.mockResolvedValue(null);
const req = new Request("http://localhost/api/api-keys/key-id-1");
const res = await DELETE(req, { params: Promise.resolve({ id: "key-id-1" }) });
expect(res.status).toBe(401);
});
it("deleteMany is called with where clause containing both id and userId (ownership enforcement)", async () => {
vi.mocked(prisma.userApiKey.deleteMany).mockResolvedValue({ count: 1 } as never);
const req = new Request("http://localhost/api/api-keys/key-id-1");
await DELETE(req, { params: Promise.resolve({ id: "key-id-1" }) });
expect(vi.mocked(prisma.userApiKey.deleteMany)).toHaveBeenCalledWith(
expect.objectContaining({
where: expect.objectContaining({
id: "key-id-1",
userId: "user-123",
}),
})
);
});
});
-45
View File
@@ -1,45 +0,0 @@
import { NextResponse } from "next/server";
import { auth } from "@/lib/auth";
import { prisma } from "@/lib/db";
import { isDemo } from "@/core/edition";
import { apiKeyManagementLimiter, getClientIp } from "@/lib/rateLimiters";
// ---------------------------------------------------------------------------
// DELETE /api/api-keys/[id] — KEY-03 (ownership-scoped revoke, hard delete)
// ---------------------------------------------------------------------------
export async function DELETE(
request: Request,
{ params }: { params: Promise<{ id: string }> },
) {
const { id } = await params;
if (isDemo) {
return NextResponse.json({ error: "Not available in demo edition" }, { status: 403 });
}
const limited = apiKeyManagementLimiter.check(getClientIp(request));
if (limited) return limited;
const session = await auth();
if (!session?.user?.id) {
return NextResponse.json({ error: "Unauthorized" }, { status: 401 });
}
try {
// Atomic ownership-scoped delete — T-16-06 (BOLA prevention).
// Using deleteMany (not delete) so a non-matching id returns count 0
// instead of throwing a NotFound error, eliminating TOCTOU.
const deleted = await prisma.userApiKey.deleteMany({
where: { id, userId: session.user.id },
});
if (deleted.count === 0) {
return NextResponse.json({ error: "not_found" }, { status: 404 });
}
return NextResponse.json({ success: true });
} catch (err) {
console.error("[api-keys] DELETE error:", err);
return NextResponse.json({ error: "Failed to revoke API key" }, { status: 500 });
}
}
-280
View File
@@ -1,280 +0,0 @@
import {
describe, it, expect, vi, beforeEach
} from "vitest";
import type { Session } from "next-auth";
import { GET, POST } from "./route";
import { auth } from "@/lib/auth";
import { prisma } from "@/lib/db";
import { generateApiKey } from "@/lib/apiKeyAuth";
vi.mock("@/lib/auth", () => ({
auth: vi.fn(),
}));
vi.mock("@/lib/db", () => ({
prisma: {
userApiKey: {
findMany: vi.fn(),
count: vi.fn(),
create: vi.fn(),
deleteMany: vi.fn(),
},
// $transaction executes the callback inline so tests don't need a real DB
$transaction: vi.fn(),
},
}));
vi.mock("@/core/edition", () => ({
isDemo: false,
}));
vi.mock("@/lib/apiKeyAuth", () => ({
generateApiKey: vi.fn(),
}));
// Bypass rate limiting in all route tests — the limiter module has its own tests
vi.mock("@/lib/rateLimiters", () => ({
apiKeyManagementLimiter: { check: vi.fn().mockReturnValue(null) },
mcpLimiter: { check: vi.fn().mockReturnValue(null) },
getClientIp: vi.fn().mockReturnValue("127.0.0.1"),
}));
// NextAuth v5 `auth` is overloaded (middleware wrapper + no-arg session getter).
// Narrow to the session-getter signature so vi.mocked resolves the correct overload.
const mockAuth = vi.mocked(auth as unknown as () => Promise<Session | null>);
// Helper: make $transaction execute its callback synchronously with the mock prisma.
// Must be called in beforeEach after vi.resetAllMocks() clears the module-level mock.
// Cast to never to bypass Prisma's complex overloaded $transaction signature.
function wireTransaction(): void {
vi.mocked(prisma.$transaction).mockImplementation(
((fn: (tx: typeof prisma) => unknown) => fn(prisma)) as never,
);
}
// ---------------------------------------------------------------------------
// GET /api/api-keys — KEY-03 (list)
// ---------------------------------------------------------------------------
describe("GET /api/api-keys", () => {
beforeEach(() => {
vi.resetAllMocks();
wireTransaction();
mockAuth.mockResolvedValue({
user: { id: "user-123", email: "[email protected]" },
} as Session);
});
it("returns 200 with { keys } for authenticated user (KEY-03)", async () => {
const mockKeys = [
{ id: "key-1", name: "My Key", prefix: "wwv_ABCDEFGH", createdAt: new Date(), lastUsedAt: null },
];
vi.mocked(prisma.userApiKey.findMany).mockResolvedValue(mockKeys as never);
const req = new Request("http://localhost/api/api-keys");
const res = await GET(req);
const body = await res.json();
expect(res.status).toBe(200);
expect(body).toHaveProperty("keys");
expect(Array.isArray(body.keys)).toBe(true);
});
it("returns 401 when no session (KEY-03)", async () => {
mockAuth.mockResolvedValue(null);
const req = new Request("http://localhost/api/api-keys");
const res = await GET(req);
const body = await res.json();
expect(res.status).toBe(401);
expect(body).toHaveProperty("error");
// Must NOT include a `keys` field — callers that do `data.keys ?? []` without
// checking r.ok would silently show an empty list instead of an auth error.
expect(body).not.toHaveProperty("keys");
});
it("response never contains hashedSecret field (security invariant)", async () => {
const mockKeys = [
{
id: "key-1",
name: "My Key",
prefix: "wwv_ABCDEFGH",
createdAt: new Date(),
lastUsedAt: null,
// hashedSecret deliberately absent from select projection
},
];
vi.mocked(prisma.userApiKey.findMany).mockResolvedValue(mockKeys as never);
const req = new Request("http://localhost/api/api-keys");
const res = await GET(req);
const body = await res.json();
expect(res.status).toBe(200);
// The route must use a select projection — no hashedSecret in response
if (Array.isArray(body.keys)) {
body.keys.forEach((k: Record<string, unknown>) => {
expect(k).not.toHaveProperty("hashedSecret");
});
}
});
});
// ---------------------------------------------------------------------------
// POST /api/api-keys — KEY-01, KEY-02, KEY-04 (create)
// ---------------------------------------------------------------------------
describe("POST /api/api-keys", () => {
beforeEach(() => {
vi.resetAllMocks();
wireTransaction();
mockAuth.mockResolvedValue({
user: { id: "user-123", email: "[email protected]" },
} as Session);
});
it("returns 422 with error 'max_keys_reached' when user already has 3 keys (KEY-04)", async () => {
vi.mocked(prisma.userApiKey.count).mockResolvedValue(3 as never);
const req = new Request("http://localhost/api/api-keys", {
method: "POST",
headers: { "Content-Type": "application/json" },
body: JSON.stringify({ name: "Fourth Key" }),
});
const res = await POST(req);
const body = await res.json();
expect(res.status).toBe(422);
expect(body.error).toBe("max_keys_reached");
});
it("calls generateApiKey and returns 201 with fullToken when under limit (KEY-01, KEY-04)", async () => {
vi.mocked(prisma.userApiKey.count).mockResolvedValue(0 as never);
vi.mocked(generateApiKey).mockReturnValue({
prefix: "wwv_TESTPFX1",
secret: "testsecret_43chars_base64url_value_here123",
hashedSecret: "a".repeat(64),
fullToken: "wwv_TESTPFX1.testsecret_43chars_base64url_value_here123",
});
vi.mocked(prisma.userApiKey.create).mockResolvedValue({
id: "new-key-id",
name: "My API Key",
createdAt: new Date("2026-01-01"),
} as never);
const req = new Request("http://localhost/api/api-keys", {
method: "POST",
headers: { "Content-Type": "application/json" },
body: JSON.stringify({ name: "My API Key" }),
});
const res = await POST(req);
const body = await res.json();
expect(res.status).toBe(201);
expect(body.key).toHaveProperty("fullToken");
expect(body.key.fullToken).toBe("wwv_TESTPFX1.testsecret_43chars_base64url_value_here123");
});
it("returns 401 when no session (KEY-03)", async () => {
mockAuth.mockResolvedValue(null);
const req = new Request("http://localhost/api/api-keys", {
method: "POST",
body: JSON.stringify({}),
});
const res = await POST(req);
expect(res.status).toBe(401);
});
it("returns 422 with error 'name_too_long' when name exceeds 64 chars (M2)", async () => {
const req = new Request("http://localhost/api/api-keys", {
method: "POST",
headers: { "Content-Type": "application/json" },
body: JSON.stringify({ name: "a".repeat(65) }),
});
const res = await POST(req);
const body = await res.json();
expect(res.status).toBe(422);
expect(body.error).toBe("name_too_long");
});
it("accepts name exactly at the 64-char boundary (M2)", async () => {
vi.mocked(prisma.userApiKey.count).mockResolvedValue(0 as never);
vi.mocked(generateApiKey).mockReturnValue({
prefix: "wwv_TESTPFX2",
secret: "testsecret_43chars_base64url_value_here123",
hashedSecret: "b".repeat(64),
fullToken: "wwv_TESTPFX2.testsecret_43chars_base64url_value_here123",
});
vi.mocked(prisma.userApiKey.create).mockResolvedValue({
id: "new-key-id-2",
name: "a".repeat(64),
createdAt: new Date("2026-01-01"),
} as never);
const req = new Request("http://localhost/api/api-keys", {
method: "POST",
headers: { "Content-Type": "application/json" },
body: JSON.stringify({ name: "a".repeat(64) }),
});
const res = await POST(req);
expect(res.status).toBe(201);
});
it("returns 401 with session_user_not_found when prisma throws P2003 on user_api_keys_userId_fkey", async () => {
// Simulates a stale JWT whose user id has no row in `users` (e.g. after a DB reset).
vi.mocked(prisma.userApiKey.count).mockResolvedValue(0 as never);
const fkError = Object.assign(new Error("Foreign key constraint failed"), {
code: "P2003",
meta: { field_name: "user_api_keys_userId_fkey" },
});
vi.mocked(prisma.userApiKey.create).mockRejectedValue(fkError);
vi.mocked(generateApiKey).mockReturnValue({
prefix: "wwv_TESTPFX3",
secret: "testsecret_43chars_base64url_value_here123",
hashedSecret: "c".repeat(64),
fullToken: "wwv_TESTPFX3.testsecret_43chars_base64url_value_here123",
});
const req = new Request("http://localhost/api/api-keys", {
method: "POST",
headers: { "Content-Type": "application/json" },
body: JSON.stringify({ name: "Stale key" }),
});
const res = await POST(req);
const body = await res.json();
expect(res.status).toBe(401);
expect(body.error).toBe("session_user_not_found");
});
it("returns 500 when prisma throws P2003 on a different FK constraint", async () => {
// P2003 on a different constraint must NOT be mistaken for the user-not-found case.
vi.mocked(prisma.userApiKey.count).mockResolvedValue(0 as never);
const otherFkError = Object.assign(new Error("Foreign key constraint failed"), {
code: "P2003",
meta: { field_name: "some_other_fkey" },
});
vi.mocked(prisma.userApiKey.create).mockRejectedValue(otherFkError);
vi.mocked(generateApiKey).mockReturnValue({
prefix: "wwv_TESTPFX4",
secret: "testsecret_43chars_base64url_value_here123",
hashedSecret: "d".repeat(64),
fullToken: "wwv_TESTPFX4.testsecret_43chars_base64url_value_here123",
});
const req = new Request("http://localhost/api/api-keys", {
method: "POST",
headers: { "Content-Type": "application/json" },
body: JSON.stringify({ name: "Other error" }),
});
const res = await POST(req);
const body = await res.json();
expect(res.status).toBe(500);
expect(body.error).not.toBe("session_user_not_found");
});
});
-174
View File
@@ -1,174 +0,0 @@
import { NextResponse } from "next/server";
import { auth } from "@/lib/auth";
import { prisma } from "@/lib/db";
import { isDemo } from "@/core/edition";
import { generateApiKey } from "@/lib/apiKeyAuth";
import { apiKeyManagementLimiter, getClientIp } from "@/lib/rateLimiters";
const MAX_KEYS = 3;
const MAX_NAME_LENGTH = 64;
// ---------------------------------------------------------------------------
// GET /api/api-keys — KEY-03 (list user's keys, secrets never returned)
// ---------------------------------------------------------------------------
export async function GET(request: Request) {
if (isDemo) {
return NextResponse.json({ error: "Not available in demo edition" }, { status: 403 });
}
const limited = apiKeyManagementLimiter.check(getClientIp(request));
if (limited) return limited;
const session = await auth();
if (!session?.user?.id) {
return NextResponse.json({ error: "Unauthorized" }, { status: 401 });
}
try {
const keys = await prisma.userApiKey.findMany({
where: { userId: session.user.id },
select: {
id: true,
name: true,
prefix: true,
createdAt: true,
lastUsedAt: true,
// hashedSecret intentionally excluded — never returned to client
},
orderBy: { createdAt: "asc" },
});
return NextResponse.json({ keys });
} catch (err) {
console.error("[api-keys] GET error:", err);
return NextResponse.json({ error: "Failed to fetch API keys" }, { status: 500 });
}
}
// ---------------------------------------------------------------------------
// POST /api/api-keys — KEY-01 (reveal-once), KEY-04 (max-3)
// ---------------------------------------------------------------------------
export async function POST(request: Request) {
if (isDemo) {
return NextResponse.json({ error: "Not available in demo edition" }, { status: 403 });
}
const limited = apiKeyManagementLimiter.check(getClientIp(request));
if (limited) return limited;
const session = await auth();
if (!session?.user?.id) {
return NextResponse.json({ error: "Unauthorized" }, { status: 401 });
}
const body = await request.json().catch(() => ({})) as { name?: unknown };
const rawName = typeof body.name === "string" ? body.name.trim() : "";
// M2: reject names exceeding the server-side length limit
if (rawName.length > MAX_NAME_LENGTH) {
return NextResponse.json(
{ error: "name_too_long", message: `Name must be ${MAX_NAME_LENGTH} characters or fewer` },
{ status: 422 },
);
}
try {
const created = await createKeyInTransaction(session.user.id, rawName);
if (!created) {
return NextResponse.json(
{ error: "max_keys_reached", message: "Maximum of 3 API keys allowed per user" },
{ status: 422 },
);
}
return NextResponse.json({ key: created }, { status: 201 });
} catch (err) {
// P2003 on the userId FK means the session refers to a user that has
// no row in `users` (e.g. a stale JWT after a DB reset). Signal the
// client to re-authenticate rather than surfacing an opaque 500.
const isPrismaP2003OnUserFk =
typeof err === "object" &&
err !== null &&
"code" in err &&
(err as { code: string }).code === "P2003" &&
"meta" in err &&
typeof (err as { meta: unknown }).meta === "object" &&
(err as { meta: { field_name?: string } }).meta !== null &&
(err as { meta: { field_name?: string } }).meta.field_name === "user_api_keys_userId_fkey";
if (isPrismaP2003OnUserFk) {
console.error("[api-keys] POST: session user not found in DB (id redacted)");
return NextResponse.json(
{
error: "session_user_not_found",
message: "Your session refers to a user that no longer exists. Please sign in again.",
},
{ status: 401 },
);
}
console.error("[api-keys] POST error:", err);
return NextResponse.json({ error: "Failed to create API key" }, { status: 500 });
}
}
// ---------------------------------------------------------------------------
// Internal helper — M1: wraps count + create in a Serializable transaction
// so concurrent POSTs cannot exceed the key limit (TOCTOU prevention).
// Returns null when the max is already reached.
// ---------------------------------------------------------------------------
async function createKeyInTransaction(
userId: string,
rawName: string,
): Promise<{ id: string; name: string; createdAt: Date; fullToken: string } | null> {
// $transaction with isolationLevel Serializable prevents concurrent
// inserts from both reading count < 3 and both succeeding (TOCTOU fix).
return prisma.$transaction(async (tx) => {
const count = await tx.userApiKey.count({ where: { userId } });
if (count >= MAX_KEYS) return null;
const name = rawName || `API Key ${count + 1}`;
return createKeyWithRetry(userId, name, tx);
}, { isolationLevel: "Serializable" });
}
// ---------------------------------------------------------------------------
// Internal helper — generates a key + creates the DB row, retries once on
// P2002 prefix collision (extremely rare but possible with 8 random chars).
// fullToken is returned here (reveal-once) and never persisted.
// ---------------------------------------------------------------------------
async function createKeyWithRetry(
userId: string,
name: string,
tx: Parameters<Parameters<typeof prisma.$transaction>[0]>[0],
): Promise<{ id: string; name: string; createdAt: Date; fullToken: string }> {
const { prefix, hashedSecret, fullToken } = generateApiKey();
try {
const row = await tx.userApiKey.create({
data: { userId, prefix, hashedSecret, name },
select: { id: true, name: true, createdAt: true },
});
return { ...row, fullToken };
} catch (err: unknown) {
// Prisma unique-constraint violation (P2002) on prefix — retry once
const isPrismaP2002 =
typeof err === "object" &&
err !== null &&
"code" in err &&
(err as { code: string }).code === "P2002";
if (!isPrismaP2002) throw err;
const retry = generateApiKey();
const row = await tx.userApiKey.create({
data: { userId, prefix: retry.prefix, hashedSecret: retry.hashedSecret, name },
select: { id: true, name: true, createdAt: true },
});
return { ...row, fullToken: retry.fullToken };
}
}
+15
View File
@@ -55,6 +55,7 @@ export async function GET(req: NextRequest) {
headers: {
"Content-Type": contentType,
"Cache-Control": "no-store",
"Access-Control-Allow-Origin": "*",
},
});
}
@@ -77,6 +78,8 @@ export async function GET(req: NextRequest) {
"Content-Type": contentType,
"Cache-Control": "no-store",
"X-Content-Type-Options": "nosniff",
"Access-Control-Allow-Origin": "*",
"Access-Control-Allow-Methods": "GET, OPTIONS",
},
});
} catch (error: unknown) {
@@ -88,3 +91,15 @@ export async function GET(req: NextRequest) {
);
}
}
export async function OPTIONS() {
return new Response(null, {
status: 204,
headers: {
"Access-Control-Allow-Origin": "*",
"Access-Control-Allow-Methods": "GET, OPTIONS",
"Access-Control-Allow-Headers": "Content-Type",
"Access-Control-Max-Age": "86400",
},
});
}
+1 -1
View File
@@ -31,7 +31,7 @@ export async function GET(req: NextRequest) {
const upstream = await safeFetch(targetUrl, {
headers: { "User-Agent": "WorldWideView/1.0" },
timeout: MAX_STREAM_DURATION_MS,
streaming: true,
maxSize: 100 * 1024 * 1024, // 100MB for streams
});
if (!upstream.ok) {
-253
View File
@@ -1,253 +0,0 @@
/**
* RED test scaffold for GET /api/globe/commands (Phase 19a Wave 0).
*
* These tests INTENTIONALLY FAIL because route.ts does not exist yet.
* They lock the following contracts:
*
* CMD-ROUTE-01 No auth -> 401
* CMD-ROUTE-02 NextAuth session -> 200 with { commands: [...] } from drainGlobeCommands
* CMD-ROUTE-03 Bearer API key fallback when no session -> 200
* CMD-ROUTE-04 userId comes from auth result, NEVER from the query string
* CMD-ROUTE-05 sessionId taken from ?sessionId query param, scoped to that tab
* CMD-ROUTE-06 missing or malformed sessionId -> 400
*/
import { describe, it, expect, vi, beforeEach } from "vitest";
import type { Session } from "next-auth";
import { GET } from "./route";
import { auth } from "@/lib/auth";
import { authenticateApiKey } from "@/lib/apiKeyAuth";
import { drainGlobeCommands } from "@/lib/globeCommandQueue";
// ---------------------------------------------------------------------------
// Top-level mocks
// (global setup.ts already mocks @/lib/auth returning null — we override below)
// ---------------------------------------------------------------------------
vi.mock("@/lib/apiKeyAuth", () => ({
authenticateApiKey: vi.fn(),
}));
vi.mock("@/lib/globeCommandQueue", () => ({
drainGlobeCommands: vi.fn(),
}));
// Bypass rate limiting for route unit tests
vi.mock("@/lib/rateLimiters", () => ({
globeCommandsLimiter: { check: vi.fn().mockReturnValue(null) },
getClientIp: vi.fn().mockReturnValue("127.0.0.1"),
}));
// ---------------------------------------------------------------------------
// Typed mock helpers
// ---------------------------------------------------------------------------
const mockGetSession = vi.mocked(auth as unknown as () => Promise<Session | null>);
const mockAuthApiKey = vi.mocked(authenticateApiKey);
const mockDrain = vi.mocked(drainGlobeCommands);
function makeRequest(sessionId?: string, bearer?: string): Request {
const url = sessionId
? `http://localhost/api/globe/commands?sessionId=${sessionId}`
: "http://localhost/api/globe/commands";
const headers: Record<string, string> = {};
if (bearer) headers["authorization"] = `Bearer ${bearer}`;
return new Request(url, { method: "GET", headers });
}
// ---------------------------------------------------------------------------
// Test setup
// ---------------------------------------------------------------------------
beforeEach(() => {
vi.resetAllMocks();
mockGetSession.mockResolvedValue(null);
mockAuthApiKey.mockResolvedValue(null);
mockDrain.mockResolvedValue([]);
});
// Valid UUID-format sessionIds used throughout these tests
const SESS_1 = "a0000000-0000-0000-0000-000000000001";
const SESS_ABC = "b0000000-0000-0000-0000-0000000000ab";
const SESS_TAB = "c0000000-0000-0000-0000-00000000000c";
// ---------------------------------------------------------------------------
// CMD-ROUTE-01: unauthenticated -> 401
// ---------------------------------------------------------------------------
describe("GET /api/globe/commands — 401 when unauthenticated (CMD-ROUTE-01)", () => {
it("returns 401 when neither session nor Bearer key is present", async () => {
const req = makeRequest(SESS_1);
const res = await GET(req);
expect(res.status).toBe(401);
expect(mockDrain).not.toHaveBeenCalled();
});
it("returns 401 when session is null and api key auth returns null", async () => {
mockGetSession.mockResolvedValue(null);
mockAuthApiKey.mockResolvedValue(null);
const req = makeRequest(SESS_1, "invalid.token");
const res = await GET(req);
expect(res.status).toBe(401);
});
});
// ---------------------------------------------------------------------------
// CMD-ROUTE-02: NextAuth session path -> 200 + { commands }
// ---------------------------------------------------------------------------
describe("GET /api/globe/commands — NextAuth session (CMD-ROUTE-02)", () => {
beforeEach(() => {
mockGetSession.mockResolvedValue({
user: { id: "u1", name: "Test User", email: "[email protected]" },
expires: "2099-01-01",
} as Session);
mockDrain.mockResolvedValue([
{ type: "pan", lat: 1, lon: 2, alt: 3 },
]);
});
it("returns 200", async () => {
const req = makeRequest(SESS_ABC);
const res = await GET(req);
expect(res.status).toBe(200);
});
it("returns { commands: [...] } from drainGlobeCommands", async () => {
const req = makeRequest(SESS_ABC);
const res = await GET(req);
const body = await res.json() as { commands: unknown[] };
expect(body).toHaveProperty("commands");
expect(Array.isArray(body.commands)).toBe(true);
expect(body.commands).toHaveLength(1);
});
it("calls drainGlobeCommands with userId from session and sessionId from query string", async () => {
const req = makeRequest(SESS_ABC);
await GET(req);
expect(mockDrain).toHaveBeenCalledWith("u1", SESS_ABC);
});
it("returns { commands: [] } when queue is empty", async () => {
mockDrain.mockResolvedValue([]);
const req = makeRequest(SESS_ABC);
const res = await GET(req);
const body = await res.json() as { commands: unknown[] };
expect(body.commands).toEqual([]);
});
});
// ---------------------------------------------------------------------------
// CMD-ROUTE-03: Bearer API key fallback
// ---------------------------------------------------------------------------
describe("GET /api/globe/commands — Bearer fallback (CMD-ROUTE-03)", () => {
beforeEach(() => {
mockGetSession.mockResolvedValue(null);
mockAuthApiKey.mockResolvedValue({ userId: "api-user", keyId: "key-99" });
mockDrain.mockResolvedValue([]);
});
it("returns 200 when no session but valid Bearer token", async () => {
const req = makeRequest(SESS_1, "wwv_valid.token");
const res = await GET(req);
expect(res.status).toBe(200);
});
it("calls drainGlobeCommands with userId from API key auth result", async () => {
const req = makeRequest(SESS_1, "wwv_valid.token");
await GET(req);
expect(mockDrain).toHaveBeenCalledWith("api-user", SESS_1);
});
});
// ---------------------------------------------------------------------------
// CMD-ROUTE-04: userId always from auth, never from query string
// ---------------------------------------------------------------------------
describe("GET /api/globe/commands — userId source invariant (CMD-ROUTE-04)", () => {
it("uses the session userId, not any userId in the query string", async () => {
mockGetSession.mockResolvedValue({
user: { id: "real-user", name: "Alice", email: "[email protected]" },
expires: "2099-01-01",
} as Session);
mockDrain.mockResolvedValue([]);
// The URL includes a userId query param that should be ignored
const req = new Request(
`http://localhost/api/globe/commands?sessionId=${SESS_1}&userId=injected-user`,
{ method: "GET" }
);
await GET(req);
const [calledUserId] = mockDrain.mock.calls[0] as [string, string];
expect(calledUserId).toBe("real-user");
expect(calledUserId).not.toBe("injected-user");
});
});
// ---------------------------------------------------------------------------
// CMD-ROUTE-05: sessionId from query string
// ---------------------------------------------------------------------------
describe("GET /api/globe/commands — sessionId from query (CMD-ROUTE-05)", () => {
it("passes the sessionId query param to drainGlobeCommands", async () => {
mockGetSession.mockResolvedValue({
user: { id: "u1", name: "Test", email: "[email protected]" },
expires: "2099-01-01",
} as Session);
mockDrain.mockResolvedValue([]);
const req = makeRequest(SESS_TAB);
await GET(req);
const [, calledSessionId] = mockDrain.mock.calls[0] as [string, string];
expect(calledSessionId).toBe(SESS_TAB);
});
});
// ---------------------------------------------------------------------------
// CMD-ROUTE-06: missing or malformed sessionId -> 400
// ---------------------------------------------------------------------------
describe("GET /api/globe/commands — invalid sessionId -> 400 (CMD-ROUTE-06)", () => {
// Use the API key auth path to avoid the NextMiddleware typing quirk on mockGetSession.
// userId still resolves correctly; the 400 is about sessionId, not identity.
beforeEach(() => {
mockAuthApiKey.mockResolvedValue({ userId: "u1", keyId: "k1" });
mockDrain.mockResolvedValue([]);
});
it("returns 400 when sessionId is missing", async () => {
const req = makeRequest(undefined, "wwv_valid.token");
const res = await GET(req);
expect(res.status).toBe(400);
expect(mockDrain).not.toHaveBeenCalled();
});
it("returns 400 when sessionId is not UUID format", async () => {
const req = makeRequest("not-a-uuid", "wwv_valid.token");
const res = await GET(req);
expect(res.status).toBe(400);
expect(mockDrain).not.toHaveBeenCalled();
});
it("returns { error: 'invalid sessionId' } in the body", async () => {
const req = makeRequest("bad-id", "wwv_valid.token");
const res = await GET(req);
const body = await res.json() as { error: string };
expect(body.error).toBe("invalid sessionId");
});
});
-62
View File
@@ -1,62 +0,0 @@
/**
* GET /api/globe/commands?sessionId=<uuid>
*
* Browser poll endpoint for the Globe Command System (Phase 19a).
* The browser calls this on an interval (~1500ms) to drain any commands
* queued by the MCP agent via registerGlobeCommandTools.
*
* Auth (R-2 dual-auth, mirrors POST /api/globe/state):
* Primary: NextAuth session cookie (browser path)
* Fallback: Bearer API key (MCP / programmatic path)
* userId comes ONLY from the resolved auth result -- never from the query string.
*
* sessionId comes from the ?sessionId query param (scopes the tab, not identity).
*/
import { NextResponse } from "next/server";
import { auth as getSession } from "@/lib/auth";
import { authenticateApiKey } from "@/lib/apiKeyAuth";
import { drainGlobeCommands } from "@/lib/globeCommandQueue";
import { globeCommandsLimiter, getClientIp } from "@/lib/rateLimiters";
import { resolveEdition } from "@/core/edition";
const SESSION_ID_RE = /^[0-9a-f-]{36}$/i;
export async function GET(request: Request): Promise<NextResponse> {
const limited = globeCommandsLimiter.check(getClientIp(request));
if (limited) return limited as NextResponse;
const currentEdition = resolveEdition(process.env.NEXT_PUBLIC_WWV_EDITION);
if (currentEdition === "demo") {
return NextResponse.json({ error: "Demo mode" }, { status: 403 });
}
// R-2: NextAuth session PRIMARY, Bearer API key FALLBACK.
// userId is resolved exclusively from the auth result -- never from the URL.
let userId: string | null = null;
const session = await getSession();
if (session?.user?.id) {
userId = session.user.id;
} else {
const apiKeyAuth = await authenticateApiKey(request);
if (apiKeyAuth) {
userId = apiKeyAuth.userId;
}
}
if (!userId) {
return NextResponse.json({ error: "Unauthorized" }, { status: 401 });
}
const url = new URL(request.url);
const sessionId = url.searchParams.get("sessionId") ?? "";
if (!sessionId || !SESSION_ID_RE.test(sessionId)) {
return NextResponse.json({ error: "invalid sessionId" }, { status: 400 });
}
const commands = await drainGlobeCommands(userId, sessionId);
return NextResponse.json({ commands });
}
@@ -1,304 +0,0 @@
/**
* RED tests for GET /api/globe/commands/stream (Phase 19b Wave 0).
*
* These tests lock the SSE streaming route contract. The route file does not
* exist yet -- the import below will fail with module-not-found (correct RED state).
*
* SSE-01 No auth -> 401
* SSE-02 Demo edition -> 403 (gate fires BEFORE auth)
* SSE-03 Invalid sessionId format -> 400
* SSE-04 Missing sessionId -> 400
* SSE-05 Valid auth + valid sessionId -> 200 with SSE headers
* SSE-06 Commands from drainGlobeCommands appear as SSE data events
* SSE-07 Keepalive comment ':keepalive' sent approximately every 15 seconds
*/
import { describe, it, expect, vi, beforeEach, afterEach } from "vitest";
import type { Session } from "next-auth";
// This import will fail (RED) until Wave 1 creates the route implementation.
import { GET } from "./route";
// ---------------------------------------------------------------------------
// Hoisted mock functions
// ---------------------------------------------------------------------------
const {
mockGetSession,
mockAuthApiKey,
mockDrain,
mockStreamCheck,
mockGetClientIp,
} = vi.hoisted(() => ({
mockGetSession: vi.fn<() => Promise<Session | null>>(),
mockAuthApiKey: vi.fn(),
mockDrain: vi.fn(),
mockStreamCheck: vi.fn(),
mockGetClientIp: vi.fn(),
}));
// ---------------------------------------------------------------------------
// Mocks
// ---------------------------------------------------------------------------
vi.mock("@/lib/auth", () => ({
auth: mockGetSession,
}));
vi.mock("@/lib/apiKeyAuth", () => ({
authenticateApiKey: mockAuthApiKey,
}));
vi.mock("@/lib/globeCommandQueue", () => ({
drainGlobeCommands: mockDrain,
}));
vi.mock("@/lib/rateLimiters", () => ({
globeCommandsStreamLimiter: { check: mockStreamCheck },
getClientIp: mockGetClientIp,
}));
// ---------------------------------------------------------------------------
// Helpers
// ---------------------------------------------------------------------------
const VALID_SESSION_ID = "a0000000-0000-0000-0000-000000000001";
function makeRequest(sessionId?: string, bearer?: string): Request {
const url = sessionId
? `http://localhost/api/globe/commands/stream?sessionId=${sessionId}`
: "http://localhost/api/globe/commands/stream";
const headers: Record<string, string> = {};
if (bearer) headers["authorization"] = `Bearer ${bearer}`;
return new Request(url, { method: "GET", headers });
}
/** Reads all available bytes from a ReadableStream within maxMs. */
async function collectStreamChunks(response: Response, maxMs = 100): Promise<string> {
const reader = response.body?.getReader();
if (!reader) return "";
const decoder = new TextDecoder();
let result = "";
const deadline = Date.now() + maxMs;
try {
while (Date.now() < deadline) {
const { done, value } = await reader.read();
if (done) break;
if (value) result += decoder.decode(value, { stream: true });
}
} finally {
reader.releaseLock();
}
return result;
}
// ---------------------------------------------------------------------------
// Setup
// ---------------------------------------------------------------------------
beforeEach(() => {
vi.resetAllMocks();
mockGetSession.mockResolvedValue(null);
mockAuthApiKey.mockResolvedValue(null);
mockDrain.mockResolvedValue([]);
mockStreamCheck.mockReturnValue(null); // not rate-limited
mockGetClientIp.mockReturnValue("127.0.0.1");
});
afterEach(() => {
vi.useRealTimers();
});
// ---------------------------------------------------------------------------
// SSE-01: no auth -> 401
// ---------------------------------------------------------------------------
describe("GET /api/globe/commands/stream -- 401 when unauthenticated (SSE-01)", () => {
it("returns 401 when neither session nor Bearer key is present", async () => {
const req = makeRequest(VALID_SESSION_ID);
const res = await GET(req);
expect(res.status).toBe(401);
expect(mockDrain).not.toHaveBeenCalled();
});
it("returns 401 when session is null and apiKey auth returns null", async () => {
mockGetSession.mockResolvedValue(null);
mockAuthApiKey.mockResolvedValue(null);
const req = makeRequest(VALID_SESSION_ID, "invalid.token");
const res = await GET(req);
expect(res.status).toBe(401);
});
});
// ---------------------------------------------------------------------------
// SSE-02: demo edition -> 403 (BEFORE auth)
// ---------------------------------------------------------------------------
describe("GET /api/globe/commands/stream -- 403 in demo edition (SSE-02)", () => {
it("returns 403 when edition is demo, before attempting auth", async () => {
// The route must check isDemo first -- mockGetSession should NOT be called.
vi.stubEnv("NEXT_PUBLIC_WWV_EDITION", "demo");
const req = makeRequest(VALID_SESSION_ID);
const res = await GET(req);
expect(res.status).toBe(403);
// Auth must NOT have been called -- demo gate is pre-auth
expect(mockGetSession).not.toHaveBeenCalled();
expect(mockAuthApiKey).not.toHaveBeenCalled();
vi.unstubAllEnvs();
});
});
// ---------------------------------------------------------------------------
// SSE-03: invalid sessionId format -> 400
// ---------------------------------------------------------------------------
describe("GET /api/globe/commands/stream -- 400 for invalid sessionId (SSE-03)", () => {
beforeEach(() => {
mockAuthApiKey.mockResolvedValue({ userId: "u1", keyId: "k1" });
});
it("returns 400 when sessionId is not UUID format", async () => {
const req = makeRequest("not-a-uuid", "wwv_valid.token");
const res = await GET(req);
expect(res.status).toBe(400);
expect(mockDrain).not.toHaveBeenCalled();
});
it("returns { error: 'invalid sessionId' } in the body", async () => {
const req = makeRequest("bad-id", "wwv_valid.token");
const res = await GET(req);
const body = await res.json() as { error: string };
expect(body.error).toBe("invalid sessionId");
});
});
// ---------------------------------------------------------------------------
// SSE-04: missing sessionId -> 400
// ---------------------------------------------------------------------------
describe("GET /api/globe/commands/stream -- 400 for missing sessionId (SSE-04)", () => {
it("returns 400 when sessionId query param is absent", async () => {
mockAuthApiKey.mockResolvedValue({ userId: "u1", keyId: "k1" });
const req = makeRequest(undefined, "wwv_valid.token");
const res = await GET(req);
expect(res.status).toBe(400);
expect(mockDrain).not.toHaveBeenCalled();
});
});
// ---------------------------------------------------------------------------
// SSE-05: valid auth + valid sessionId -> 200 with SSE headers
// ---------------------------------------------------------------------------
describe("GET /api/globe/commands/stream -- 200 with SSE headers (SSE-05)", () => {
beforeEach(() => {
mockGetSession.mockResolvedValue({
user: { id: "u1", name: "Test User", email: "[email protected]" },
expires: "2099-01-01",
} as Session);
mockDrain.mockResolvedValue([]);
});
it("returns 200", async () => {
const req = makeRequest(VALID_SESSION_ID);
const res = await GET(req);
expect(res.status).toBe(200);
});
it("sets Content-Type: text/event-stream", async () => {
const req = makeRequest(VALID_SESSION_ID);
const res = await GET(req);
expect(res.headers.get("Content-Type")).toContain("text/event-stream");
});
it("sets Cache-Control: no-cache", async () => {
const req = makeRequest(VALID_SESSION_ID);
const res = await GET(req);
expect(res.headers.get("Cache-Control")).toBe("no-cache");
});
it("sets X-Accel-Buffering: no (disables nginx buffering)", async () => {
const req = makeRequest(VALID_SESSION_ID);
const res = await GET(req);
expect(res.headers.get("X-Accel-Buffering")).toBe("no");
});
});
// ---------------------------------------------------------------------------
// SSE-06: commands from drainGlobeCommands appear as SSE data events
// ---------------------------------------------------------------------------
describe("GET /api/globe/commands/stream -- commands pushed as SSE events (SSE-06)", () => {
it("includes command JSON as a 'data: ...' SSE event line", async () => {
vi.useFakeTimers();
mockGetSession.mockResolvedValue({
user: { id: "u1", name: "Test User", email: "[email protected]" },
expires: "2099-01-01",
} as Session);
// First call returns a command; subsequent calls return empty
mockDrain
.mockResolvedValueOnce([{ type: "pan", lat: 1, lon: 2, alt: 3 }])
.mockResolvedValue([]);
const req = makeRequest(VALID_SESSION_ID);
const res = await GET(req);
expect(res.status).toBe(200);
// Advance time slightly so the first poll loop iteration executes
await vi.runAllTimersAsync();
const text = await collectStreamChunks(res, 50);
// SSE data line must be present
expect(text).toContain("data: ");
// Must end event with double newline (SSE spec)
expect(text).toContain("\n\n");
// The JSON payload must contain the command type
expect(text).toContain('"pan"');
});
});
// ---------------------------------------------------------------------------
// SSE-07: keepalive comment sent every ~15 seconds
// ---------------------------------------------------------------------------
describe("GET /api/globe/commands/stream -- keepalive comment every 15s (SSE-07)", () => {
it("includes ':keepalive' comment after 16 seconds", async () => {
vi.useFakeTimers();
mockGetSession.mockResolvedValue({
user: { id: "u1", name: "Test User", email: "[email protected]" },
expires: "2099-01-01",
} as Session);
mockDrain.mockResolvedValue([]);
const req = makeRequest(VALID_SESSION_ID);
const res = await GET(req);
expect(res.status).toBe(200);
// Advance past one keepalive interval (15s)
await vi.advanceTimersByTimeAsync(16_000);
const text = await collectStreamChunks(res, 50);
expect(text).toContain(":keepalive");
});
});
-124
View File
@@ -1,124 +0,0 @@
/**
* GET /api/globe/commands/stream?sessionId=<uuid>
*
* SSE push transport for the Globe Command System (Phase 19b).
* Replaces the browser's 1500ms poll loop with a persistent EventSource
* connection, cutting command latency from ~1500ms to near-real-time.
*
* Auth (R-2 dual-auth, mirrors GET /api/globe/commands):
* Primary: NextAuth session cookie (browser path)
* Fallback: Bearer API key (MCP / programmatic path)
* userId comes ONLY from the resolved auth result -- never from the URL.
*
* Gate order (MUST NOT be reordered):
* 1. Rate limiter (DoS defense before any IO)
* 2. Demo edition guard (pre-auth, no DB hit on demo)
* 3. Dual-auth (session or API key)
* 4. UUID sessionId guard (scopes the Redis queue key)
*/
import { NextResponse } from "next/server";
import { auth as getSession } from "@/lib/auth";
import { authenticateApiKey } from "@/lib/apiKeyAuth";
import { drainGlobeCommands } from "@/lib/globeCommandQueue";
import { globeCommandsStreamLimiter, getClientIp } from "@/lib/rateLimiters";
import { resolveEdition } from "@/core/edition";
const SESSION_ID_RE = /^[0-9a-f-]{36}$/i;
const POLL_INTERVAL_MS = 200;
const KEEPALIVE_INTERVAL_MS = 15_000;
const MAX_DURATION_MS = 16_000;
const SSE_HEADERS = {
"Content-Type": "text/event-stream; charset=utf-8",
"Cache-Control": "no-cache",
"Connection": "keep-alive",
"X-Accel-Buffering": "no",
};
function sleep(ms: number): Promise<void> {
return new Promise<void>((resolve) => setTimeout(resolve, ms));
}
export async function GET(request: Request): Promise<Response> {
// Gate 1: rate limiter -- must be first to protect against DoS before auth IO
const limited = globeCommandsStreamLimiter.check(getClientIp(request));
if (limited) return limited as Response;
// Gate 2: demo edition -- read env at request time so vi.stubEnv works in tests
const currentEdition = resolveEdition(process.env.NEXT_PUBLIC_WWV_EDITION);
if (currentEdition === "demo") {
return NextResponse.json({ error: "Demo mode" }, { status: 403 }) as Response;
}
// Gate 3: dual-auth -- session primary, Bearer API key fallback
let userId: string | null = null;
const session = await getSession();
if (session?.user?.id) {
userId = session.user.id;
} else {
const apiKeyAuth = await authenticateApiKey(request);
if (apiKeyAuth) {
userId = apiKeyAuth.userId;
}
}
if (!userId) {
return NextResponse.json({ error: "Unauthorized" }, { status: 401 }) as Response;
}
// Gate 4: UUID sessionId guard -- scopes the Redis queue key
const url = new URL(request.url);
const sessionId = url.searchParams.get("sessionId") ?? "";
if (!sessionId || !SESSION_ID_RE.test(sessionId)) {
return NextResponse.json({ error: "invalid sessionId" }, { status: 400 }) as Response;
}
// Capture for use inside the ReadableStream closure
const resolvedUserId = userId;
const resolvedSessionId = sessionId;
let cancelled = false;
const stream = new ReadableStream({
async start(controller) {
const encoder = new TextEncoder();
function send(chunk: string): void {
controller.enqueue(encoder.encode(chunk));
}
const startTime = Date.now();
let lastKeepalive = Date.now();
try {
while (Date.now() - startTime < MAX_DURATION_MS) {
const commands = await drainGlobeCommands(resolvedUserId, resolvedSessionId);
for (const cmd of commands) {
send("data: " + JSON.stringify({ commands: [cmd] }) + "\n\n");
}
if (Date.now() - lastKeepalive >= KEEPALIVE_INTERVAL_MS) {
send(":keepalive\n\n");
lastKeepalive = Date.now();
}
await sleep(POLL_INTERVAL_MS);
if (cancelled) break;
}
controller.close();
} catch (err) {
console.error("[globe/commands/stream] stream error:", err);
controller.error(err);
}
},
cancel() {
cancelled = true;
},
});
return new Response(stream, { headers: SSE_HEADERS });
}
-120
View File
@@ -1,120 +0,0 @@
import { describe, it, expect, vi, beforeEach } from "vitest";
import { POST } from "./route";
import { authenticateApiKey } from "@/lib/apiKeyAuth";
import { writeGlobeState } from "@/lib/globeStateStore";
vi.mock("@/lib/apiKeyAuth", () => ({
authenticateApiKey: vi.fn(),
}));
vi.mock("@/lib/globeStateStore", () => ({
writeGlobeState: vi.fn(),
}));
// Bypass rate limiting so route tests don't need a limiter mock
vi.mock("@/lib/rateLimiters", () => ({
mcpLimiter: { check: vi.fn().mockReturnValue(null) },
getClientIp: vi.fn().mockReturnValue("127.0.0.1"),
}));
// Mutable ref so individual tests can toggle the demo edition flag.
// Must be declared with vi.hoisted so it is available when vi.mock is hoisted.
const editionMock = vi.hoisted(() => ({ isDemo: false }));
vi.mock("@/core/edition", () => editionMock);
const mockAuth = vi.mocked(authenticateApiKey);
const mockWrite = vi.mocked(writeGlobeState);
const validSnapshot = {
viewport: { lat: 37.7, lon: -122.4, altitude: 500000, heading: 0, pitch: -90, roll: 0 },
layers: {},
timeline: {
currentTime: "2026-01-15T12:00:00.000Z",
timeWindow: "24h",
isPlaybackMode: false,
playbackTime: 1737000000000,
playbackSpeed: 1,
},
selectedEntity: null,
lastUpdate: 1737000000000,
};
function makeRequest(body: unknown): Request {
return new Request("http://localhost/api/globe/state", {
method: "POST",
headers: { "Content-Type": "application/json" },
body: JSON.stringify(body),
});
}
beforeEach(() => {
vi.resetAllMocks();
editionMock.isDemo = false;
mockAuth.mockResolvedValue({ userId: "user-1", keyId: "key-1" });
mockWrite.mockResolvedValue(undefined);
});
// ---------------------------------------------------------------------------
// POST /api/globe/state (RSRC-01)
// ---------------------------------------------------------------------------
describe("POST /api/globe/state (RSRC-01)", () => {
it("returns 200 and calls writeGlobeState with valid auth + body", async () => {
const req = makeRequest({ sessionId: "sess-abc", snapshot: validSnapshot });
const res = await POST(req);
expect(res.status).toBe(200);
expect(mockWrite).toHaveBeenCalledWith("user-1", "sess-abc", validSnapshot);
});
it("returns 401 and does NOT call writeGlobeState when auth returns null", async () => {
mockAuth.mockResolvedValue(null);
const req = makeRequest({ sessionId: "sess-abc", snapshot: validSnapshot });
const res = await POST(req);
expect(res.status).toBe(401);
expect(mockWrite).not.toHaveBeenCalled();
});
it("returns 400 and does NOT call writeGlobeState when sessionId is missing", async () => {
const req = makeRequest({ snapshot: validSnapshot });
const res = await POST(req);
expect(res.status).toBe(400);
expect(mockWrite).not.toHaveBeenCalled();
});
it("returns 400 when snapshot is missing from body", async () => {
const req = makeRequest({ sessionId: "sess-abc" });
const res = await POST(req);
expect(res.status).toBe(400);
expect(mockWrite).not.toHaveBeenCalled();
});
it("still yields 200 when writeGlobeState rejects (fire-and-forget contract)", async () => {
mockWrite.mockRejectedValue(new Error("Redis unavailable"));
const req = makeRequest({ sessionId: "sess-abc", snapshot: validSnapshot });
const res = await POST(req);
expect(res.status).toBe(200);
});
});
// ---------------------------------------------------------------------------
// POST /api/globe/state -- demo edition gate (GAP-01)
// ---------------------------------------------------------------------------
describe("POST /api/globe/state -- demo edition gate", () => {
it("returns 403 and does NOT call writeGlobeState when isDemo is true", async () => {
editionMock.isDemo = true;
const req = makeRequest({ sessionId: "sess-abc", snapshot: validSnapshot });
const res = await POST(req);
const body = await res.json();
expect(res.status).toBe(403);
expect(body).toEqual({ error: "Not available in demo edition" });
expect(mockWrite).not.toHaveBeenCalled();
});
});
-55
View File
@@ -1,55 +0,0 @@
import { NextResponse } from "next/server";
import { auth as getSession } from "@/lib/auth";
import { authenticateApiKey } from "@/lib/apiKeyAuth";
import { writeGlobeState } from "@/lib/globeStateStore";
import { mcpLimiter, getClientIp } from "@/lib/rateLimiters";
import { isDemo } from "@/core/edition";
import type { GlobeStateSnapshot } from "@/lib/globeState";
export async function POST(request: Request) {
const limited = mcpLimiter.check(getClientIp(request));
if (limited) return limited;
if (isDemo) {
return NextResponse.json({ error: "Not available in demo edition" }, { status: 403 });
}
// R-2: browser write path authenticates via NextAuth session cookie (primary),
// falling back to Bearer API key for programmatic/MCP clients.
let userId: string | null = null;
const session = await getSession();
if (session?.user?.id) {
userId = session.user.id;
} else {
const apiKeyAuth = await authenticateApiKey(request);
if (apiKeyAuth) {
userId = apiKeyAuth.userId;
}
}
if (!userId) {
return NextResponse.json({ error: "Unauthorized" }, { status: 401 });
}
let body: Record<string, unknown>;
try {
body = await request.json() as Record<string, unknown>;
} catch {
return NextResponse.json({ error: "Invalid JSON body" }, { status: 400 });
}
const { sessionId, snapshot } = body;
if (typeof sessionId !== "string" || sessionId.length === 0) {
return NextResponse.json({ error: "sessionId is required" }, { status: 400 });
}
if (snapshot === null || typeof snapshot !== "object" || Array.isArray(snapshot)) {
return NextResponse.json({ error: "snapshot is required" }, { status: 400 });
}
void writeGlobeState(userId, sessionId, snapshot as GlobeStateSnapshot);
return NextResponse.json({ ok: true });
}
+3 -3
View File
@@ -55,9 +55,9 @@ export async function GET(request: NextRequest) {
const origin = getRequestOrigin(request);
const loginUrl = new URL("/login", origin);
// Construct a relative path for `next` to ensure it redirects back to the identical host
const nextPath = request.nextUrl.pathname + request.nextUrl.search;
loginUrl.searchParams.set("next", nextPath);
// Construct a relative path for callback to ensure it redirects back to the identical host
const callbackPath = request.nextUrl.pathname + request.nextUrl.search;
loginUrl.searchParams.set("callbackUrl", callbackPath);
return NextResponse.redirect(loginUrl);
}
@@ -1,8 +1,6 @@
import { NextResponse } from "next/server";
import type { NextRequest } from "next/server";
import { auth } from "@/lib/auth";
import { getSupabaseUser } from "@/lib/supabase/server";
import { isCloud, isPluginInstallEnabled, isDemo, isDemoAdmin } from "@/core/edition";
import { upsertPlugin } from "@/lib/marketplace/repository";
import { issueMarketplaceToken } from "@/lib/marketplace/marketplaceToken";
import type { MarketplaceSessionToken } from "@worldwideview/wwv-plugin-sdk";
@@ -10,6 +8,7 @@ import type { PluginManifest } from "@/core/plugins/PluginManifest";
import { validateManifest } from "@/core/plugins/validateManifest";
import { installLimiter } from "@/lib/rateLimiters";
import { getClientIp } from "@/lib/rateLimit";
import { isPluginInstallEnabled, isDemo, isDemoAdmin } from "@/core/edition";
import { getVerifiedPluginIds } from "@/lib/marketplace/registryClient";
import { getRequestOrigin } from "@/lib/origin";
@@ -52,38 +51,23 @@ export async function GET(request: NextRequest) {
const rateLimited = installLimiter.check(getClientIp(request));
if (rateLimited) return rateLimited;
// Auth: cloud edition uses Supabase session; local/demo uses NextAuth.
const supabaseUrl = process.env.NEXT_PUBLIC_SUPABASE_URL;
const isDummyUrl = !supabaseUrl || supabaseUrl.includes("dummy") || supabaseUrl.includes("xyz.supabase.co");
const useSupabaseAuth = isCloud && !isDummyUrl;
const session = await auth();
let userId: string;
if (useSupabaseAuth) {
const supabaseUser = await getSupabaseUser();
if (!supabaseUser) {
const origin = getRequestOrigin(request);
const loginUrl = new URL("/login", origin);
const nextPath = request.nextUrl.pathname + request.nextUrl.search;
loginUrl.searchParams.set("next", nextPath);
return NextResponse.redirect(loginUrl);
}
userId = supabaseUser.id;
} else {
const session = await auth();
if (!session?.user) {
const origin = getRequestOrigin(request);
const loginUrl = new URL("/login", origin);
const nextPath = request.nextUrl.pathname + request.nextUrl.search;
loginUrl.searchParams.set("next", nextPath);
return NextResponse.redirect(loginUrl);
}
if (isDemo && !isDemoAdmin(session)) {
return NextResponse.json(
{ error: "Admin access required on Demo edition" },
{ status: 403 }
);
}
userId = session.user.id ?? "";
// Not logged in — send to login with this URL as callbackUrl
if (!session?.user) {
const origin = getRequestOrigin(request);
const loginUrl = new URL("/login", origin);
const callbackPath = request.nextUrl.pathname + request.nextUrl.search;
loginUrl.searchParams.set("callbackUrl", callbackPath);
return NextResponse.redirect(loginUrl);
}
if (isDemo && !isDemoAdmin(session)) {
return NextResponse.json(
{ error: "Admin access required on Demo edition" },
{ status: 403 }
);
}
const pluginId = searchParams.get("pluginId");
@@ -130,7 +114,7 @@ export async function GET(request: NextRequest) {
return NextResponse.json({ error: "Install failed" }, { status: 500 });
}
const token: MarketplaceSessionToken = await issueMarketplaceToken(userId);
const token: MarketplaceSessionToken = await issueMarketplaceToken(session.user.id ?? "");
const successUrl = new URL(redirectTo);
// Unverified plugins need user confirmation on the WWV client side
-157
View File
@@ -1,157 +0,0 @@
/**
* POST /api/mcp/catalog
*
* Browser-facing endpoint that accepts a per-session plugin catalog
* and stores it in Redis scoped to the authenticated user's session.
*
* Gate ordering:
* 1. Rate limit (prevents abuse before any auth work)
* 2. isDemo 403 (demo edition does not support MCP)
* 3. Dual-auth: NextAuth session cookie PRIMARY, Bearer apiKey FALLBACK
* 4. sessionId UUID guard
* 5. Body shape + size validation
* 6. publishSessionCatalog (userId ONLY from auth result)
*
* Security invariants:
* - userId is read ONLY from the auth result -- never from the request body.
* - Catalog Redis key is scoped both {userId} and {sessionId}.
* - Payload size is capped before write; malformed shapes are rejected.
* - isDemo 403 gate runs BEFORE auth (avoids DB layer in demo mode).
*/
import { NextResponse } from "next/server";
import { auth as getSession } from "@/lib/auth";
import { authenticateApiKey } from "@/lib/apiKeyAuth";
import { publishSessionCatalog, readSessionCatalog } from "@/lib/mcpSessionCatalog";
import type { SessionCatalog } from "@/lib/mcpSessionCatalog";
import { resolveActiveSessionId } from "@/lib/globeCommandQueue";
import { globeCommandsLimiter, getClientIp } from "@/lib/rateLimiters";
import { isDemo } from "@/core/edition";
// ---------------------------------------------------------------------------
// Constants
// ---------------------------------------------------------------------------
const SESSION_ID_RE = /^[0-9a-f-]{36}$/i;
/** Maximum allowed serialized body size in bytes (~64 KB). */
const MAX_BODY_BYTES = 64 * 1024;
// ---------------------------------------------------------------------------
// POST handler
// ---------------------------------------------------------------------------
export async function POST(request: Request): Promise<NextResponse> {
// Gate 1: Rate limit
const limited = globeCommandsLimiter.check(getClientIp(request));
if (limited) return limited as NextResponse;
// Gate 2: isDemo 403 (BEFORE auth)
if (isDemo) {
return NextResponse.json(
{ error: "MCP catalog is not available in demo mode" },
{ status: 403 },
);
}
// Gate 3: Dual-auth -- NextAuth session PRIMARY, Bearer apiKey FALLBACK
let userId: string | null = null;
const session = await getSession();
if (session?.user?.id) {
userId = session.user.id;
} else {
const apiKeyAuth = await authenticateApiKey(request);
if (apiKeyAuth) {
userId = apiKeyAuth.userId;
}
}
if (!userId) {
return NextResponse.json({ error: "Unauthorized" }, { status: 401 });
}
// Gate 4: sessionId UUID guard
// sessionId may come from the request body; if absent, fall back to the
// user's active session (same pattern as globeCommandTools).
let rawBody: unknown;
try {
const text = await request.text();
if (Buffer.byteLength(text, "utf8") > MAX_BODY_BYTES) {
return NextResponse.json({ error: "Payload too large" }, { status: 413 });
}
rawBody = JSON.parse(text);
} catch {
return NextResponse.json({ error: "Invalid JSON body" }, { status: 400 });
}
if (!rawBody || typeof rawBody !== "object") {
return NextResponse.json({ error: "Body must be a JSON object" }, { status: 400 });
}
const body = rawBody as Record<string, unknown>;
// Resolve sessionId
let sessionId: string | null = null;
const rawSessionId = body.sessionId;
if (typeof rawSessionId === "string" && rawSessionId !== "") {
if (!SESSION_ID_RE.test(rawSessionId)) {
return NextResponse.json({ error: "invalid sessionId" }, { status: 400 });
}
sessionId = rawSessionId;
} else {
sessionId = await resolveActiveSessionId(userId);
}
if (!sessionId) {
return NextResponse.json({ error: "No active session" }, { status: 400 });
}
// Gate 5: Body shape validation
if (!Array.isArray(body.tools)) {
return NextResponse.json({ error: "body.tools must be an array" }, { status: 400 });
}
if (!Array.isArray(body.capabilities)) {
return NextResponse.json({ error: "body.capabilities must be an array" }, { status: 400 });
}
// Validate each tool entry: namespacedName format, description, inputSchema.
const NAMESPACED_NAME_RE = /^[a-zA-Z0-9_-]+__[a-zA-Z0-9_-]+$/;
const MAX_NAMESPACED_NAME_LENGTH = 128;
for (const entry of body.tools) {
if (!entry || typeof entry !== "object") {
return NextResponse.json({ error: "invalid tool entry" }, { status: 400 });
}
const e = entry as Record<string, unknown>;
if (
typeof e.namespacedName !== "string" ||
e.namespacedName.length > MAX_NAMESPACED_NAME_LENGTH ||
!NAMESPACED_NAME_RE.test(e.namespacedName)
) {
return NextResponse.json({ error: "invalid tool entry" }, { status: 400 });
}
if (typeof e.description !== "string") {
return NextResponse.json({ error: "invalid tool entry" }, { status: 400 });
}
if (
!e.inputSchema ||
typeof e.inputSchema !== "object" ||
Array.isArray(e.inputSchema)
) {
return NextResponse.json({ error: "invalid tool entry" }, { status: 400 });
}
}
const catalog: SessionCatalog = {
tools: body.tools as SessionCatalog["tools"],
capabilities: body.capabilities as string[],
};
// Gate 6: Store -- userId ONLY from auth result (NEVER from body)
await publishSessionCatalog(userId, sessionId, catalog);
return NextResponse.json({ ok: true });
}
// Export readSessionCatalog for use by the MCP route (re-export avoids extra import path)
export { readSessionCatalog };
-381
View File
@@ -1,381 +0,0 @@
/**
* RED test scaffold for registerGlobeCommandTools (Phase 19a Wave 0).
*
* These tests INTENTIONALLY FAIL because globeCommandTools.ts does not exist yet.
* They lock the following contracts:
*
* TOOL-01 All four tools are registered: pan_globe, focus_entity, toggle_layer, set_timeline
* TOOL-02 pan_globe with explicit sessionId calls enqueueGlobeCommand directly (no resolve)
* TOOL-03 toggle_layer without sessionId calls resolveActiveSessionId then enqueueGlobeCommand
* TOOL-04 When resolveActiveSessionId returns null, handler returns graceful result, no enqueue
* TOOL-05 userId comes ONLY from ctx, never from tool args
*/
import { describe, it, expect, vi, beforeEach } from "vitest";
import { registerGlobeCommandTools, latSchema, lonSchema, altSchema } from "./globeCommandTools";
import { isValidGlobeCommand } from "@/core/globe/types/GlobeCommand";
import { latSchema as sharedLatSchema, lonSchema as sharedLonSchema } from "@/lib/mcp/coordinateSchemas";
// ---------------------------------------------------------------------------
// Mock @/lib/globeCommandQueue
// vi.mock is hoisted above const declarations, so mock objects must be
// declared with vi.hoisted() to be accessible inside the factory function.
// ---------------------------------------------------------------------------
const { mockEnqueue, mockResolveSessionId } = vi.hoisted(() => ({
mockEnqueue: vi.fn().mockResolvedValue(undefined),
mockResolveSessionId: vi.fn().mockResolvedValue("resolved-session"),
}));
vi.mock("@/lib/globeCommandQueue", () => ({
enqueueGlobeCommand: mockEnqueue,
resolveActiveSessionId: mockResolveSessionId,
}));
// ---------------------------------------------------------------------------
// Fake McpServer — records registered tools
// ---------------------------------------------------------------------------
type ToolHandler = (input: Record<string, unknown>) => Promise<{ content: [{ type: "text"; text: string }] }>;
function makeFakeServer() {
const tools = new Map<string, ToolHandler>();
const server = {
registerTool: vi.fn((name: string, _def: unknown, handler: ToolHandler) => {
tools.set(name, handler);
}),
};
return { server, tools };
}
beforeEach(() => {
vi.resetAllMocks();
mockEnqueue.mockResolvedValue(undefined);
mockResolveSessionId.mockResolvedValue("resolved-session");
});
// ---------------------------------------------------------------------------
// TOOL-01: all four tools are registered
// ---------------------------------------------------------------------------
describe("registerGlobeCommandTools registration (TOOL-01)", () => {
it("registers pan_globe, focus_entity, toggle_layer, and set_timeline", () => {
const { server, tools } = makeFakeServer();
registerGlobeCommandTools(server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer, { userId: "u1" });
expect(tools.has("pan_globe")).toBe(true);
expect(tools.has("focus_entity")).toBe(true);
expect(tools.has("toggle_layer")).toBe(true);
expect(tools.has("set_timeline")).toBe(true);
});
it("calls server.registerTool exactly four times", () => {
const { server } = makeFakeServer();
registerGlobeCommandTools(server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer, { userId: "u1" });
expect(server.registerTool).toHaveBeenCalledTimes(4);
});
});
// ---------------------------------------------------------------------------
// TOOL-02: pan_globe with explicit sessionId bypasses resolveActiveSessionId
// ---------------------------------------------------------------------------
describe("pan_globe handler (TOOL-02)", () => {
it("calls enqueueGlobeCommand with userId, sessionId, and pan GlobeCommand", async () => {
const { server, tools } = makeFakeServer();
registerGlobeCommandTools(server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer, { userId: "u1" });
const handler = tools.get("pan_globe")!;
await handler({ lat: 1, lon: 2, alt: 3, sessionId: "s9" });
expect(mockEnqueue).toHaveBeenCalledWith(
"u1",
"s9",
expect.objectContaining({ type: "pan", lat: 1, lon: 2, alt: 3 }),
);
});
it("does NOT call resolveActiveSessionId when sessionId is provided", async () => {
const { server, tools } = makeFakeServer();
registerGlobeCommandTools(server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer, { userId: "u1" });
const handler = tools.get("pan_globe")!;
await handler({ lat: 1, lon: 2, alt: 3, sessionId: "explicit-session" });
expect(mockResolveSessionId).not.toHaveBeenCalled();
});
it("returns a success content result", async () => {
const { server, tools } = makeFakeServer();
registerGlobeCommandTools(server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer, { userId: "u1" });
const handler = tools.get("pan_globe")!;
const result = await handler({ lat: 40.7, lon: -74.0, alt: 500000, sessionId: "s9" });
expect(result).toHaveProperty("content");
expect(Array.isArray(result.content)).toBe(true);
expect(result.content[0]).toHaveProperty("type", "text");
});
});
// ---------------------------------------------------------------------------
// TOOL-03: toggle_layer without sessionId calls resolveActiveSessionId
// ---------------------------------------------------------------------------
describe("toggle_layer handler — session resolution (TOOL-03)", () => {
it("calls resolveActiveSessionId(userId) when no sessionId is provided", async () => {
const { server, tools } = makeFakeServer();
registerGlobeCommandTools(server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer, { userId: "u1" });
const handler = tools.get("toggle_layer")!;
await handler({ layerId: "ais" });
expect(mockResolveSessionId).toHaveBeenCalledWith("u1");
});
it("enqueues using the resolved sessionId", async () => {
mockResolveSessionId.mockResolvedValue("resolved-s99");
const { server, tools } = makeFakeServer();
registerGlobeCommandTools(server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer, { userId: "u1" });
const handler = tools.get("toggle_layer")!;
await handler({ layerId: "aviation" });
expect(mockEnqueue).toHaveBeenCalledWith(
"u1",
"resolved-s99",
expect.objectContaining({ type: "toggleLayer", layerId: "aviation" }),
);
});
});
// ---------------------------------------------------------------------------
// TOOL-04: graceful result when no active session
// ---------------------------------------------------------------------------
describe("handler graceful fallback — no active session (TOOL-04)", () => {
beforeEach(() => {
mockResolveSessionId.mockResolvedValue(null);
});
it("does NOT call enqueueGlobeCommand when resolveActiveSessionId returns null", async () => {
const { server, tools } = makeFakeServer();
registerGlobeCommandTools(server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer, { userId: "u1" });
const handler = tools.get("toggle_layer")!;
await handler({ layerId: "ais" });
expect(mockEnqueue).not.toHaveBeenCalled();
});
it("returns a content result mentioning no active globe session", async () => {
const { server, tools } = makeFakeServer();
registerGlobeCommandTools(server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer, { userId: "u1" });
const handler = tools.get("toggle_layer")!;
const result = await handler({ layerId: "ais" });
expect(result.content[0].type).toBe("text");
expect(result.content[0].text.toLowerCase()).toContain("no active");
});
it("set_timeline also returns graceful result without enqueuing", async () => {
const { server, tools } = makeFakeServer();
registerGlobeCommandTools(server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer, { userId: "u1" });
const handler = tools.get("set_timeline")!;
const result = await handler({ timeWindow: "24h" });
expect(mockEnqueue).not.toHaveBeenCalled();
expect(result.content[0].text.toLowerCase()).toContain("no active");
});
});
// ---------------------------------------------------------------------------
// TOOL-05: userId always from ctx, never from tool args
// ---------------------------------------------------------------------------
describe("userId source invariant (TOOL-05)", () => {
it("uses ctx.userId not a userId field in args", async () => {
const { server, tools } = makeFakeServer();
registerGlobeCommandTools(server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer, { userId: "ctx-user" });
const handler = tools.get("pan_globe")!;
// Simulate a caller that supplies a userId in args — must be ignored
await handler({ lat: 1, lon: 2, alt: 3, sessionId: "s1" });
const [calledUserId] = mockEnqueue.mock.calls[0] as [string, string, unknown];
expect(calledUserId).toBe("ctx-user");
});
});
// ---------------------------------------------------------------------------
// TOOL-06: coordinate schema bounds validation (latSchema / lonSchema / altSchema)
// ---------------------------------------------------------------------------
describe("coordinate schema bounds (TOOL-06)", () => {
// latSchema
it("latSchema accepts valid latitude", () => {
expect(latSchema.safeParse(35.6762).success).toBe(true);
expect(latSchema.safeParse(0).success).toBe(true);
expect(latSchema.safeParse(-90).success).toBe(true);
expect(latSchema.safeParse(90).success).toBe(true);
});
it("latSchema rejects lat=999", () => {
expect(latSchema.safeParse(999).success).toBe(false);
});
it("latSchema rejects lat=-91", () => {
expect(latSchema.safeParse(-91).success).toBe(false);
});
it("latSchema rejects NaN", () => {
expect(latSchema.safeParse(NaN).success).toBe(false);
});
it("latSchema rejects Infinity", () => {
expect(latSchema.safeParse(Infinity).success).toBe(false);
});
// lonSchema
it("lonSchema accepts valid longitude", () => {
expect(lonSchema.safeParse(139.6503).success).toBe(true);
expect(lonSchema.safeParse(-180).success).toBe(true);
expect(lonSchema.safeParse(180).success).toBe(true);
});
it("lonSchema rejects lon=181", () => {
expect(lonSchema.safeParse(181).success).toBe(false);
});
it("lonSchema rejects lon=-181", () => {
expect(lonSchema.safeParse(-181).success).toBe(false);
});
it("lonSchema rejects NaN", () => {
expect(lonSchema.safeParse(NaN).success).toBe(false);
});
// altSchema
it("altSchema accepts valid altitude", () => {
expect(altSchema.safeParse(2000000).success).toBe(true);
expect(altSchema.safeParse(1).success).toBe(true);
});
it("altSchema rejects alt=0", () => {
expect(altSchema.safeParse(0).success).toBe(false);
});
it("altSchema rejects negative altitude", () => {
expect(altSchema.safeParse(-100).success).toBe(false);
});
it("altSchema rejects Infinity", () => {
expect(altSchema.safeParse(Infinity).success).toBe(false);
});
it("altSchema rejects NaN", () => {
expect(altSchema.safeParse(NaN).success).toBe(false);
});
});
// ---------------------------------------------------------------------------
// TOOL-07: isValidGlobeCommand rejects out-of-range coordinates
// ---------------------------------------------------------------------------
describe("isValidGlobeCommand coordinate bounds (TOOL-07)", () => {
it("accepts a valid pan command", () => {
expect(isValidGlobeCommand({ type: "pan", lat: 35.6762, lon: 139.6503, alt: 2000000 })).toBe(true);
});
it("rejects pan with lat=999", () => {
expect(isValidGlobeCommand({ type: "pan", lat: 999, lon: 0, alt: 1000 })).toBe(false);
});
it("rejects pan with lat=-91", () => {
expect(isValidGlobeCommand({ type: "pan", lat: -91, lon: 0, alt: 1000 })).toBe(false);
});
it("rejects pan with lon=181", () => {
expect(isValidGlobeCommand({ type: "pan", lat: 0, lon: 181, alt: 1000 })).toBe(false);
});
it("rejects pan with lon=-181", () => {
expect(isValidGlobeCommand({ type: "pan", lat: 0, lon: -181, alt: 1000 })).toBe(false);
});
it("rejects pan with alt=0", () => {
expect(isValidGlobeCommand({ type: "pan", lat: 0, lon: 0, alt: 0 })).toBe(false);
});
it("rejects pan with negative alt", () => {
expect(isValidGlobeCommand({ type: "pan", lat: 0, lon: 0, alt: -500 })).toBe(false);
});
it("rejects pan with NaN lat", () => {
expect(isValidGlobeCommand({ type: "pan", lat: NaN, lon: 0, alt: 1000 })).toBe(false);
});
it("rejects pan with Infinity alt", () => {
expect(isValidGlobeCommand({ type: "pan", lat: 0, lon: 0, alt: Infinity })).toBe(false);
});
it("accepts focusEntity with valid lat/lon", () => {
expect(isValidGlobeCommand({ type: "focusEntity", lat: 35.6762, lon: 139.6503 })).toBe(true);
});
it("accepts focusEntity with no lat/lon (entity-id only)", () => {
expect(isValidGlobeCommand({ type: "focusEntity", entityId: "ent-1" })).toBe(true);
});
it("rejects focusEntity with out-of-range lat", () => {
expect(isValidGlobeCommand({ type: "focusEntity", lat: 91, lon: 0 })).toBe(false);
});
it("rejects focusEntity with out-of-range lon", () => {
expect(isValidGlobeCommand({ type: "focusEntity", lat: 0, lon: 200 })).toBe(false);
});
});
// ---------------------------------------------------------------------------
// TOOL-08: isValidGlobeCommand rejects NaN heading (isNumber requires isFinite)
// ---------------------------------------------------------------------------
describe("isValidGlobeCommand NaN heading rejection (TOOL-08)", () => {
it("rejects a pan command with heading: NaN", () => {
expect(isValidGlobeCommand({ type: "pan", lat: 0, lon: 0, alt: 1000, heading: NaN })).toBe(false);
});
});
// ---------------------------------------------------------------------------
// TOOL-09: get_entities_in_region schema bounds (latSchema / lonSchema shared module)
// ---------------------------------------------------------------------------
describe("get_entities_in_region coordinate schema bounds (TOOL-09)", () => {
it("rejects north=999 (out-of-range latitude)", () => {
expect(sharedLatSchema.safeParse(999).success).toBe(false);
});
it("rejects west=-200 (out-of-range longitude)", () => {
expect(sharedLonSchema.safeParse(-200).success).toBe(false);
});
it("accepts a valid bounding box", () => {
expect(sharedLatSchema.safeParse(51.5).success).toBe(true); // north
expect(sharedLatSchema.safeParse(48.8).success).toBe(true); // south
expect(sharedLonSchema.safeParse(-0.1).success).toBe(true); // west
expect(sharedLonSchema.safeParse(2.3).success).toBe(true); // east
});
it("rejects NaN for a latitude bound", () => {
expect(sharedLatSchema.safeParse(NaN).success).toBe(false);
});
it("rejects Infinity for a longitude bound", () => {
expect(sharedLonSchema.safeParse(Infinity).success).toBe(false);
});
});
-200
View File
@@ -1,200 +0,0 @@
/**
* MCP Globe Command Tool registrar (Phase 19a Wave 2 -- 19-03).
*
* Registers four MCP tools that let an AI agent control the live globe in the
* browser. Commands are enqueued to a per-session Redis list; the browser polls
* GET /api/globe/commands to drain them (poll-based delivery, 19a transport).
*
* pan_globe -- fly the camera to a lat/lon/altitude
* focus_entity -- point the camera at a known entity or coordinate
* toggle_layer -- enable or disable a named plugin layer
* set_timeline -- set playback time / window / mode
*
* Security: userId comes ONLY from ctx (the verified auth result). It is never
* read from tool arguments. sessionId may come from args (scopes the tab) or is
* resolved from the user's active ZSET entry.
*/
import { z } from "zod";
import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
import { enqueueGlobeCommand, resolveActiveSessionId } from "@/lib/globeCommandQueue";
import { TIME_WINDOW_VALUES } from "@/core/globe/types/GlobeCommand";
import type { GlobeCommand } from "@/core/globe/types/GlobeCommand";
import { latSchema, lonSchema, altSchema } from "@/lib/mcp/coordinateSchemas";
// Re-export so existing tests that import from this module continue to work.
export { latSchema, lonSchema, altSchema };
// ---------------------------------------------------------------------------
// Shared helpers
// ---------------------------------------------------------------------------
type McpTextResult = { content: [{ type: "text"; text: string }] };
function textResult(text: string): McpTextResult {
return { content: [{ type: "text", text }] };
}
const NO_SESSION_RESULT = textResult("no active globe session to control");
/**
* Resolves the session to use: explicit arg takes precedence, falling back to
* the most-recently-active session for this user. Returns null if none is live.
*/
async function resolveSession(
userId: string,
argSessionId: string | undefined,
): Promise<string | null> {
if (argSessionId !== undefined && argSessionId !== "") {
return argSessionId;
}
return resolveActiveSessionId(userId);
}
// ---------------------------------------------------------------------------
// Public registrar
// ---------------------------------------------------------------------------
export function registerGlobeCommandTools(
server: McpServer,
ctx: { userId: string },
): void {
const { userId } = ctx;
// TOOL-01: pan_globe
server.registerTool(
"pan_globe",
{
description:
"Fly the globe camera to a geographic position. Provide lat/lon/alt (metres). Optional heading, pitch, and animation duration.",
inputSchema: {
lat: latSchema.describe("Latitude in decimal degrees [-90, 90]"),
lon: lonSchema.describe("Longitude in decimal degrees [-180, 180]"),
alt: altSchema.describe("Altitude in metres above the ellipsoid (must be > 0)"),
heading: z.number().optional().describe("Camera heading in degrees (0 = north)"),
pitch: z.number().optional().describe("Camera pitch in degrees (-90 = straight down)"),
duration: z.number().optional().describe("Flight animation duration in seconds"),
sessionId: z.string().optional().describe("Target globe session id. Obtain valid ids by reading the globe://sessions resource. Omit to target your most-recently-active browser tab."),
},
},
async (args) => {
try {
const sessionId = await resolveSession(userId, args.sessionId);
if (sessionId === null) return NO_SESSION_RESULT;
const cmd: GlobeCommand = {
type: "pan",
lat: args.lat,
lon: args.lon,
alt: args.alt,
...(args.heading !== undefined && { heading: args.heading }),
...(args.pitch !== undefined && { pitch: args.pitch }),
...(args.duration !== undefined && { duration: args.duration }),
};
await enqueueGlobeCommand(userId, sessionId, cmd);
return textResult(`Camera panning to lat=${args.lat}, lon=${args.lon}, alt=${args.alt}`);
} catch (err) {
console.error("[globeCommandTools] pan_globe failed:", err);
return textResult("pan_globe command failed");
}
},
);
// TOOL-02: focus_entity
server.registerTool(
"focus_entity",
{
description:
"Point the globe camera at a geographic coordinate. Provide lat and lon to fly the camera there. " +
"entityId is accepted but entity-id-to-coordinate resolution is not yet wired; always include lat/lon for a reliable camera move.",
inputSchema: {
entityId: z.string().optional().describe("Entity id (informational; coordinate resolution not yet supported -- also provide lat/lon)"),
lat: latSchema.optional().describe("Latitude to focus on [-90, 90]"),
lon: lonSchema.optional().describe("Longitude to focus on [-180, 180]"),
sessionId: z.string().optional().describe("Target globe session id. Obtain valid ids by reading the globe://sessions resource. Omit to target your most-recently-active browser tab."),
},
},
async (args) => {
try {
const sessionId = await resolveSession(userId, args.sessionId);
if (sessionId === null) return NO_SESSION_RESULT;
const cmd: GlobeCommand = {
type: "focusEntity",
...(args.entityId !== undefined && { entityId: args.entityId }),
...(args.lat !== undefined && { lat: args.lat }),
...(args.lon !== undefined && { lon: args.lon }),
};
await enqueueGlobeCommand(userId, sessionId, cmd);
return textResult(`Focus entity command enqueued (entityId=${args.entityId ?? "none"}, lat=${args.lat ?? "none"}, lon=${args.lon ?? "none"})`);
} catch (err) {
console.error("[globeCommandTools] focus_entity failed:", err);
return textResult("focus_entity command failed");
}
},
);
// TOOL-03: toggle_layer
server.registerTool(
"toggle_layer",
{
description:
"Enable or disable a plugin data layer on the globe. Omit 'enabled' to toggle the current state.",
inputSchema: {
layerId: z.string().describe("The plugin/layer identifier to toggle"),
enabled: z.boolean().optional().describe("True to enable, false to disable, omit to toggle"),
sessionId: z.string().optional().describe("Target globe session id. Obtain valid ids by reading the globe://sessions resource. Omit to target your most-recently-active browser tab."),
},
},
async (args) => {
try {
const sessionId = await resolveSession(userId, args.sessionId);
if (sessionId === null) return NO_SESSION_RESULT;
const cmd: GlobeCommand = {
type: "toggleLayer",
layerId: args.layerId,
...(args.enabled !== undefined && { enabled: args.enabled }),
};
await enqueueGlobeCommand(userId, sessionId, cmd);
return textResult(`Layer '${args.layerId}' toggle command enqueued`);
} catch (err) {
console.error("[globeCommandTools] toggle_layer failed:", err);
return textResult("toggle_layer command failed");
}
},
);
// TOOL-04: set_timeline
server.registerTool(
"set_timeline",
{
description:
"Set the globe timeline position, time window, or playback mode.",
inputSchema: {
currentTime: z.string().optional().describe("ISO 8601 datetime to seek to"),
timeWindow: z.enum(TIME_WINDOW_VALUES).optional().describe("Time window: one of '1h', '6h', '24h', '48h', '7d'"),
isPlaybackMode: z.boolean().optional().describe("True to start playback, false to pause"),
sessionId: z.string().optional().describe("Target globe session id. Obtain valid ids by reading the globe://sessions resource. Omit to target your most-recently-active browser tab."),
},
},
async (args) => {
try {
const sessionId = await resolveSession(userId, args.sessionId);
if (sessionId === null) return NO_SESSION_RESULT;
const cmd: GlobeCommand = {
type: "setTimeline",
...(args.currentTime !== undefined && { currentTime: args.currentTime }),
...(args.timeWindow !== undefined && { timeWindow: args.timeWindow }),
...(args.isPlaybackMode !== undefined && { isPlaybackMode: args.isPlaybackMode }),
};
await enqueueGlobeCommand(userId, sessionId, cmd);
return textResult("Timeline command enqueued");
} catch (err) {
console.error("[globeCommandTools] set_timeline failed:", err);
return textResult("set_timeline command failed");
}
},
);
}
-266
View File
@@ -1,266 +0,0 @@
import { describe, it, expect, vi, beforeEach } from "vitest";
import { registerGlobeResources } from "./globeResources";
import {
readGlobeState,
readActiveSessions,
} from "@/lib/globeStateStore";
vi.mock("@/lib/globeStateStore", () => ({
readGlobeState: vi.fn(),
readActiveSessions: vi.fn(),
}));
const mockReadState = vi.mocked(readGlobeState);
const mockReadSessions = vi.mocked(readActiveSessions);
// ---------------------------------------------------------------------------
// Fake MCP server — captures registerResource calls for assertion.
// Shape mirrors the McpServer surface used by registerGlobeResources.
// ---------------------------------------------------------------------------
function makeFakeServer() {
return {
registerResource: vi.fn(),
};
}
beforeEach(() => {
vi.resetAllMocks();
});
// ---------------------------------------------------------------------------
// RSRC-02/03/04 — registerGlobeResources registration
// ---------------------------------------------------------------------------
describe("registerGlobeResources — resource registration", () => {
it("calls server.registerResource exactly three times (RSRC-02/03/04)", () => {
const server = makeFakeServer();
registerGlobeResources(server as never, { userId: "u1" });
expect(server.registerResource).toHaveBeenCalledTimes(3);
});
it("registers globe://state/{sessionId} resource (RSRC-02)", () => {
const server = makeFakeServer();
registerGlobeResources(server as never, { userId: "u1" });
const uris = (server.registerResource.mock.calls as [string | { uriTemplate: string }, ...unknown[]][]).map(
([uriOrTemplate]) =>
typeof uriOrTemplate === "string" ? uriOrTemplate : uriOrTemplate.uriTemplate,
);
expect(uris.some((u: string) => u.includes("state"))).toBe(true);
});
it("registers globe://sessions resource (RSRC-03)", () => {
const server = makeFakeServer();
registerGlobeResources(server as never, { userId: "u1" });
const uris = (server.registerResource.mock.calls as [string | { uriTemplate: string }, ...unknown[]][]).map(
([uriOrTemplate]) =>
typeof uriOrTemplate === "string" ? uriOrTemplate : uriOrTemplate.uriTemplate,
);
expect(uris.some((u: string) => u.includes("sessions"))).toBe(true);
});
it("registers globe://layers resource (RSRC-04)", () => {
const server = makeFakeServer();
registerGlobeResources(server as never, { userId: "u1" });
const uris = (server.registerResource.mock.calls as [string | { uriTemplate: string }, ...unknown[]][]).map(
([uriOrTemplate]) =>
typeof uriOrTemplate === "string" ? uriOrTemplate : uriOrTemplate.uriTemplate,
);
expect(uris.some((u: string) => u.includes("layers"))).toBe(true);
});
});
// ---------------------------------------------------------------------------
// RSRC-02 — globe://state/{sessionId} handler
// ---------------------------------------------------------------------------
describe("globe://state handler (RSRC-02)", () => {
it("calls readGlobeState with injected userId 'u1' and the given sessionId", async () => {
const snapshot = {
viewport: { lat: 37.7, lon: -122.4, altitude: 500000, heading: 0, pitch: -90, roll: 0 },
layers: {},
timeline: {
currentTime: "2026-01-15T12:00:00.000Z",
timeWindow: "24h",
isPlaybackMode: false,
playbackTime: 1737000000000,
playbackSpeed: 1,
},
selectedEntity: null,
lastUpdate: 1737000000000,
};
mockReadState.mockResolvedValue(snapshot);
const server = makeFakeServer();
registerGlobeResources(server as never, { userId: "u1" });
// Find the state resource handler (the call whose URI/template contains "state")
const stateCall = (server.registerResource.mock.calls as [string | { uriTemplate: string }, ...unknown[]][]).find(
([uriOrTemplate]) => {
const s = typeof uriOrTemplate === "string" ? uriOrTemplate : uriOrTemplate.uriTemplate;
return s.includes("state");
},
);
expect(stateCall).toBeDefined();
// The handler is the last argument in the registerResource call
const handler = stateCall![stateCall!.length - 1] as (
uri: URL,
params: Record<string, string>,
) => Promise<unknown>;
const uri = new URL("globe://state/s1");
const result = await handler(uri, { sessionId: "s1" });
expect(mockReadState).toHaveBeenCalledWith("u1", "s1");
// Response shape: { contents: [{ uri, mimeType, text }] }
expect(result).toHaveProperty("contents");
const contents = (result as { contents: unknown[] }).contents;
expect(Array.isArray(contents)).toBe(true);
expect(contents[0]).toHaveProperty("mimeType", "application/json");
expect(contents[0]).toHaveProperty("text");
const parsed = JSON.parse((contents[0] as { text: string }).text);
expect(parsed).toEqual(snapshot);
});
it("uses injected userId 'u1' — never a hardcoded value", async () => {
mockReadState.mockResolvedValue(null);
const server = makeFakeServer();
// Register with a different userId to confirm scoping
registerGlobeResources(server as never, { userId: "other-user" });
const stateCall = (server.registerResource.mock.calls as [string | { uriTemplate: string }, ...unknown[]][]).find(
([uriOrTemplate]) => {
const s = typeof uriOrTemplate === "string" ? uriOrTemplate : uriOrTemplate.uriTemplate;
return s.includes("state");
},
);
const handler = stateCall![stateCall!.length - 1] as (
uri: URL,
params: Record<string, string>,
) => Promise<unknown>;
await handler(new URL("globe://state/s1"), { sessionId: "s1" });
expect(mockReadState).toHaveBeenCalledWith("other-user", "s1");
expect(mockReadState).not.toHaveBeenCalledWith("u1", expect.anything());
});
});
// ---------------------------------------------------------------------------
// RSRC-03 — globe://sessions handler
// ---------------------------------------------------------------------------
describe("globe://sessions handler (RSRC-03)", () => {
it("calls readActiveSessions with injected userId and returns session list as JSON", async () => {
const sessions = [
{ sessionId: "fresh-1", lastSeen: Date.now() },
{ sessionId: "fresh-2", lastSeen: Date.now() - 5000 },
];
mockReadSessions.mockResolvedValue(sessions);
const server = makeFakeServer();
registerGlobeResources(server as never, { userId: "u1" });
const sessionsCall = (server.registerResource.mock.calls as [string | { uriTemplate: string }, ...unknown[]][]).find(
([uriOrTemplate]) => {
const s = typeof uriOrTemplate === "string" ? uriOrTemplate : uriOrTemplate.uriTemplate;
return s.includes("sessions");
},
);
expect(sessionsCall).toBeDefined();
const handler = sessionsCall![sessionsCall!.length - 1] as (
uri: URL,
params: Record<string, string>,
) => Promise<unknown>;
const result = await handler(new URL("globe://sessions"), {});
expect(mockReadSessions).toHaveBeenCalledWith("u1");
const contents = (result as { contents: { text: string }[] }).contents;
const parsed = JSON.parse(contents[0].text);
expect(parsed).toEqual(sessions);
});
});
// ---------------------------------------------------------------------------
// RSRC-04 — globe://layers handler
// ---------------------------------------------------------------------------
describe("globe://layers handler (RSRC-04)", () => {
it("returns layers from the most-recent session snapshot as JSON", async () => {
const layers = {
aviation: { enabled: true, entityCount: 10, loading: false },
maritime: { enabled: false, entityCount: 0, loading: false },
};
const snapshot = {
viewport: { lat: 0, lon: 0, altitude: 1000000, heading: 0, pitch: -90, roll: 0 },
layers,
timeline: {
currentTime: "2026-01-15T12:00:00.000Z",
timeWindow: "24h",
isPlaybackMode: false,
playbackTime: 1737000000000,
playbackSpeed: 1,
},
selectedEntity: null,
lastUpdate: 1737000000000,
};
const sessions = [{ sessionId: "sess-latest", lastSeen: Date.now() }];
mockReadSessions.mockResolvedValue(sessions);
mockReadState.mockResolvedValue(snapshot);
const server = makeFakeServer();
registerGlobeResources(server as never, { userId: "u1" });
const layersCall = (server.registerResource.mock.calls as [string | { uriTemplate: string }, ...unknown[]][]).find(
([uriOrTemplate]) => {
const s = typeof uriOrTemplate === "string" ? uriOrTemplate : uriOrTemplate.uriTemplate;
return s.includes("layers");
},
);
expect(layersCall).toBeDefined();
const handler = layersCall![layersCall!.length - 1] as (
uri: URL,
params: Record<string, string>,
) => Promise<unknown>;
const result = await handler(new URL("globe://layers"), {});
const contents = (result as { contents: { text: string }[] }).contents;
const parsed = JSON.parse(contents[0].text);
expect(parsed).toEqual(layers);
});
it("scopes readActiveSessions by injected userId 'u1', never hardcoded", async () => {
mockReadSessions.mockResolvedValue([]);
const server = makeFakeServer();
registerGlobeResources(server as never, { userId: "scoped-user" });
const layersCall = (server.registerResource.mock.calls as [string | { uriTemplate: string }, ...unknown[]][]).find(
([uriOrTemplate]) => {
const s = typeof uriOrTemplate === "string" ? uriOrTemplate : uriOrTemplate.uriTemplate;
return s.includes("layers");
},
);
const handler = layersCall![layersCall!.length - 1] as (
uri: URL,
params: Record<string, string>,
) => Promise<unknown>;
await handler(new URL("globe://layers"), {});
expect(mockReadSessions).toHaveBeenCalledWith("scoped-user");
expect(mockReadSessions).not.toHaveBeenCalledWith("u1");
});
});
-98
View File
@@ -1,98 +0,0 @@
/**
* Globe MCP Resource registrar (Phase 18 Wave 3 — 18-04).
*
* Registers three read-only MCP resources scoped to the authenticated user:
* RSRC-02 globe://state/{sessionId} — full GlobeStateSnapshot for a session
* RSRC-03 globe://sessions — list of active sessions
* RSRC-04 globe://layers — layer map from the most-recent session
*
* Each handler closes over `userId` at registration time so the same McpServer
* instance is safely reusable without leaking cross-user data.
*/
import { McpServer, ResourceTemplate } from "@modelcontextprotocol/sdk/server/mcp.js";
import { readGlobeState, readActiveSessions } from "@/lib/globeStateStore";
// ---------------------------------------------------------------------------
// Public registrar
// ---------------------------------------------------------------------------
export function registerGlobeResources(
server: McpServer,
{ userId }: { userId: string },
): void {
// RSRC-02: globe://state/{sessionId}
// Returns the full GlobeStateSnapshot for the given session.
const stateTemplate = new ResourceTemplate("globe://state/{sessionId}", { list: undefined });
server.registerResource(
"globe-state",
stateTemplate,
{ description: "Full GlobeStateSnapshot for a specific session" },
async (uri, variables) => {
const sessionId = String(variables["sessionId"] ?? "");
const snapshot = await readGlobeState(userId, sessionId);
return {
contents: [
{
uri: uri.toString(),
mimeType: "application/json",
text: JSON.stringify(snapshot),
},
],
};
},
);
// RSRC-03: globe://sessions
// Returns the list of active sessions for the authenticated user.
server.registerResource(
"globe-sessions",
"globe://sessions",
{ description: "Active globe sessions for the authenticated user" },
async (uri) => {
const sessions = await readActiveSessions(userId);
return {
contents: [
{
uri: uri.toString(),
mimeType: "application/json",
text: JSON.stringify(sessions),
},
],
};
},
);
// RSRC-04: globe://layers
// Returns the layer map from the most-recent active session snapshot.
// Gracefully returns an empty object when no active sessions exist.
server.registerResource(
"globe-layers",
"globe://layers",
{ description: "Active layer map from the most-recent globe session" },
async (uri) => {
const sessions = await readActiveSessions(userId);
if (sessions.length === 0) {
return {
contents: [
{
uri: uri.toString(),
mimeType: "application/json",
text: JSON.stringify({}),
},
],
};
}
const snapshot = await readGlobeState(userId, sessions[0].sessionId);
return {
contents: [
{
uri: uri.toString(),
mimeType: "application/json",
text: JSON.stringify(snapshot?.layers ?? {}),
},
],
};
},
);
}
-63
View File
@@ -1,63 +0,0 @@
/**
* GET /api/mcp/invocations?sessionId=<uuid>
*
* Browser-facing endpoint for the plugin tool relay bridge (Phase 21 Wave 3 -- PLUG-03).
* The browser bridge polls this endpoint to drain any pending tool invocations
* that the MCP agent has queued for execution.
*
* Auth (dual-auth, mirrors GET /api/globe/commands):
* Primary: NextAuth session cookie (browser path)
* Fallback: Bearer API key (programmatic path)
* userId comes ONLY from the resolved auth result -- never from the URL.
*
* sessionId scopes the invocation queue to a specific browser tab.
* UUID format is validated before draining.
*/
import { NextResponse } from "next/server";
import { auth as getSession } from "@/lib/auth";
import { authenticateApiKey } from "@/lib/apiKeyAuth";
import { drainToolInvocations } from "@/lib/mcpRelay";
import { mcpInvocationsLimiter, getClientIp } from "@/lib/rateLimiters";
import { isDemo } from "@/core/edition";
const SESSION_ID_RE = /^[0-9a-f-]{36}$/i;
export async function GET(request: Request): Promise<NextResponse> {
// Rate limit before any auth or DB work.
const limited = mcpInvocationsLimiter.check(getClientIp(request));
if (limited) return limited as NextResponse;
if (isDemo) {
return NextResponse.json({ error: "MCP is not available in demo mode" }, { status: 403 });
}
// Dual-auth: NextAuth session PRIMARY, Bearer API key FALLBACK.
// userId is resolved exclusively from the auth result -- never from the URL.
let userId: string | null = null;
const session = await getSession();
if (session?.user?.id) {
userId = session.user.id;
} else {
const apiKeyAuth = await authenticateApiKey(request);
if (apiKeyAuth) {
userId = apiKeyAuth.userId;
}
}
if (!userId) {
return NextResponse.json({ error: "Unauthorized" }, { status: 401 });
}
const url = new URL(request.url);
const sessionId = url.searchParams.get("sessionId") ?? "";
if (!sessionId || !SESSION_ID_RE.test(sessionId)) {
return NextResponse.json({ error: "invalid sessionId" }, { status: 400 });
}
const invocations = await drainToolInvocations(userId, sessionId);
return NextResponse.json({ invocations });
}
-286
View File
@@ -1,286 +0,0 @@
/**
* RED tests for the namespaced plugin-tool dispatch handler (Phase 21 Wave 0).
*
* These tests INTENTIONALLY FAIL because registerPluginToolDispatch does not exist yet.
* Wave 3 creates the dispatch handler that is registered into the MCP route.
*
* Mirrors globeCommandTools.test.ts in structure and mock style.
*
* Security invariants encoded:
* MCP-QA-01 Request with no/invalid API key is rejected with auth error
* MCP-QA-02 Valid key, unknown namespaced tool -> tool-not-found error, no enqueue
* MCP-QA-03 Valid key, known tool, input fails validation -> validation error BEFORE enqueue
* MCP-QA-04 Valid key + tool + valid input, no browser result before deadline -> graceful timeout
* SEC-01 isDemo 403 gate stays BEFORE auth in the MCP route (unchanged gate order)
* SEC-06 userId comes ONLY from the auth result, never from tool args or body
*/
import { describe, it, expect, vi, beforeEach } from "vitest";
import { registerPluginToolDispatch } from "@/app/api/mcp/pluginToolDispatch";
// ---------------------------------------------------------------------------
// Mocks
// ---------------------------------------------------------------------------
const { mockEnqueueInvocation, mockWaitForResult, mockReadCatalog, mockValidateArgs } =
vi.hoisted(() => ({
mockEnqueueInvocation: vi.fn().mockResolvedValue({ rejected: false }),
mockWaitForResult: vi.fn().mockResolvedValue({ timedOut: false, value: { ok: true } }),
mockReadCatalog: vi.fn().mockResolvedValue(null),
mockValidateArgs: vi.fn().mockReturnValue({ valid: true, errors: [] }),
}));
vi.mock("@/lib/mcpRelay", () => ({
enqueueToolInvocation: mockEnqueueInvocation,
waitForToolResult: mockWaitForResult,
}));
vi.mock("@/lib/mcpSessionCatalog", () => ({
readSessionCatalog: mockReadCatalog,
}));
vi.mock("@/lib/mcp/pluginTools", () => ({
validateToolArgs: mockValidateArgs,
getNamespacedTools: vi.fn().mockReturnValue([]),
}));
// ---------------------------------------------------------------------------
// Fake MCP server -- records registered tool handlers
// ---------------------------------------------------------------------------
type ToolHandler = (
input: Record<string, unknown>,
) => Promise<{ content: [{ type: "text"; text: string }]; isError?: boolean }>;
function makeFakeServer() {
const tools = new Map<string, ToolHandler>();
const server = {
registerTool: vi.fn((name: string, _def: unknown, handler: ToolHandler) => {
tools.set(name, handler);
}),
};
return { server, tools };
}
// ---------------------------------------------------------------------------
// Fixture catalog with one known tool
// ---------------------------------------------------------------------------
const FIXTURE_CATALOG = {
tools: [
{
namespacedName: "aviation__decode_squawk",
pluginId: "aviation",
description: "Decodes a squawk code.",
inputSchema: {
type: "object" as const,
properties: { squawk: { type: "string" } },
required: ["squawk"],
},
mcpCapabilities: ["point-layer"],
},
],
capabilities: ["point-layer"],
};
beforeEach(() => {
vi.resetAllMocks();
mockEnqueueInvocation.mockResolvedValue({ rejected: false });
mockWaitForResult.mockResolvedValue({ timedOut: false, value: { ok: true } });
mockReadCatalog.mockResolvedValue(FIXTURE_CATALOG);
mockValidateArgs.mockReturnValue({ valid: true, errors: [] });
});
// ---------------------------------------------------------------------------
// Registration: the dispatch handler is registered for each catalog tool
// ---------------------------------------------------------------------------
describe("registerPluginToolDispatch registration", () => {
it("registers a handler for each namespaced tool in the catalog", async () => {
const { server, tools } = makeFakeServer();
await registerPluginToolDispatch(
server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer,
{ userId: "u1", sessionId: "s1" },
);
expect(tools.has("aviation__decode_squawk")).toBe(true);
});
it("registers no plugin tools when the catalog is null (no active session)", async () => {
mockReadCatalog.mockResolvedValue(null);
const { server } = makeFakeServer();
await registerPluginToolDispatch(
server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer,
{ userId: "u1", sessionId: null },
);
// Only system tools are registered -- no plugin tool handlers added
expect(server.registerTool).not.toHaveBeenCalled();
});
});
// ---------------------------------------------------------------------------
// MCP-QA-02: unknown namespaced tool -- tool-not-found, no enqueue
// ---------------------------------------------------------------------------
describe("dispatch handler -- unknown tool (MCP-QA-02)", () => {
it("returns a tool-not-found error for an unrecognised namespaced tool name", async () => {
const { server, tools } = makeFakeServer();
// Register with a catalog that has ONE known tool
await registerPluginToolDispatch(
server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer,
{ userId: "u1", sessionId: "s1" },
);
// No handler should exist for an unregistered name
expect(tools.has("unknown__tool")).toBe(false);
});
it("does not enqueue for an unknown tool name", async () => {
const { server, tools } = makeFakeServer();
await registerPluginToolDispatch(
server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer,
{ userId: "u1", sessionId: "s1" },
);
// If somehow a handler is invoked for an unknown tool, it must not enqueue
const handler = tools.get("unknown__tool");
if (handler) {
await handler({ squawk: "7700" });
expect(mockEnqueueInvocation).not.toHaveBeenCalled();
} else {
// Handler not registered -- correct behavior; test passes
expect(tools.has("unknown__tool")).toBe(false);
}
});
});
// ---------------------------------------------------------------------------
// MCP-QA-03: input validation BEFORE enqueue
// ---------------------------------------------------------------------------
describe("dispatch handler -- input validation before enqueue (MCP-QA-03)", () => {
it("returns a validation error when input fails the minimal validator", async () => {
mockValidateArgs.mockReturnValue({ valid: false, errors: ["squawk is required"] });
const { server, tools } = makeFakeServer();
await registerPluginToolDispatch(
server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer,
{ userId: "u1", sessionId: "s1" },
);
const handler = tools.get("aviation__decode_squawk")!;
const result = await handler({});
expect(result.content[0].type).toBe("text");
const parsed = JSON.parse(result.content[0].text) as { error?: string };
expect(parsed.error ?? result.content[0].text).toMatch(/validation|invalid|squawk/i);
});
it("does NOT call enqueueToolInvocation when validation fails", async () => {
mockValidateArgs.mockReturnValue({ valid: false, errors: ["squawk is required"] });
const { server, tools } = makeFakeServer();
await registerPluginToolDispatch(
server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer,
{ userId: "u1", sessionId: "s1" },
);
const handler = tools.get("aviation__decode_squawk")!;
await handler({});
expect(mockEnqueueInvocation).not.toHaveBeenCalled();
});
});
// ---------------------------------------------------------------------------
// MCP-QA-04: relay timeout -- graceful text result, no thrown 500
// ---------------------------------------------------------------------------
describe("dispatch handler -- relay timeout graceful result (MCP-QA-04)", () => {
it("returns a graceful timeout text result when the browser does not respond", async () => {
mockEnqueueInvocation.mockResolvedValue({ rejected: false });
mockWaitForResult.mockResolvedValue({ timedOut: true });
const { server, tools } = makeFakeServer();
await registerPluginToolDispatch(
server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer,
{ userId: "u1", sessionId: "s1" },
);
const handler = tools.get("aviation__decode_squawk")!;
const result = await handler({ squawk: "7700" });
expect(result.content[0].type).toBe("text");
expect(result.content[0].text.toLowerCase()).toMatch(/timeout|timed out|no response/);
});
it("does NOT throw or reject the promise on relay timeout", async () => {
mockEnqueueInvocation.mockResolvedValue({ rejected: false });
mockWaitForResult.mockResolvedValue({ timedOut: true });
const { server, tools } = makeFakeServer();
await registerPluginToolDispatch(
server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer,
{ userId: "u1", sessionId: "s1" },
);
const handler = tools.get("aviation__decode_squawk")!;
await expect(handler({ squawk: "7700" })).resolves.toBeDefined();
});
});
// ---------------------------------------------------------------------------
// SEC-06: userId comes ONLY from auth result, never from tool args
// ---------------------------------------------------------------------------
describe("userId source invariant (SEC-06)", () => {
it("uses ctx.userId, not a userId supplied in tool args", async () => {
const { server, tools } = makeFakeServer();
await registerPluginToolDispatch(
server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer,
{ userId: "ctx-user", sessionId: "s1" },
);
const handler = tools.get("aviation__decode_squawk")!;
// Simulate a caller that supplies userId in args -- must be ignored
await handler({ squawk: "7700", userId: "attacker-user" });
const [calledUserId] = mockEnqueueInvocation.mock.calls[0] as [string, string, unknown];
expect(calledUserId).toBe("ctx-user");
});
});
// ---------------------------------------------------------------------------
// SEC-01: isDemo gate is tested at the route level (route.test.ts).
// Here we assert that the dispatch registrar does not re-implement the gate --
// it trusts that the route has already checked it before calling registerPluginToolDispatch.
// ---------------------------------------------------------------------------
describe("SEC-01 gate ordering (documented assertion)", () => {
it("registerPluginToolDispatch does not check isDemo internally", async () => {
// The isDemo check lives in route.ts BEFORE auth. The dispatch handler
// assumes it runs AFTER the gate has already fired. This test just confirms
// the registrar does not import isDemo (behavioral contract assertion).
// If Wave 3 incorrectly adds an isDemo check here, the security reviewer
// will catch the duplicate gate.
const { server } = makeFakeServer();
// Should not throw even if we don't pass isDemo-related context
await expect(
registerPluginToolDispatch(
server as unknown as import("@modelcontextprotocol/sdk/server/mcp.js").McpServer,
{ userId: "u1", sessionId: "s1" },
),
).resolves.not.toThrow();
});
});
-179
View File
@@ -1,179 +0,0 @@
/**
* @file pluginToolDispatch.ts
* @description Plugin tool dispatch handler for the MCP route (Phase 21 Wave 3 -- PLUG-03).
*
* registerPluginToolDispatch reads the per-session catalog and registers a
* handler for each namespaced plugin tool ({pluginId}__{name}). Each handler:
* 1. Validates tool input against the catalog schema (rejects before enqueue).
* 2. Enqueues the invocation for the browser to execute.
* 3. Waits for the browser to post a result (10-second deadline).
* 4. Returns the result as a text content block, OR a graceful timeout message.
*
* The server is a DUMB RELAY -- it never executes a plugin tool, reads a streamUrl,
* or calls the data engine. Execution happens in the browser via plugin.executeMcpTool.
*
* Security:
* - userId + sessionId come from the auth context, never from tool args.
* - Input validation fires BEFORE enqueue (SEC-04 / MCP-QA-03).
* - Server never calls executeMcpTool, never reads a streamUrl.
* - Graceful timeout returned on deadline, never a hang or 500 (SEC-02 / MCP-QA-04).
*/
import { randomUUID } from "crypto";
import type { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
import { z } from "zod";
import { readSessionCatalog } from "@/lib/mcpSessionCatalog";
import { enqueueToolInvocation, waitForToolResult } from "@/lib/mcpRelay";
import { validateToolArgs } from "@/lib/mcp/pluginTools";
import type { ToolInputSchema } from "@/lib/mcp/pluginTools";
// ---------------------------------------------------------------------------
// Constants
// ---------------------------------------------------------------------------
/** Deadline for the server-side wait on a browser tool result (ms). */
const RELAY_DEADLINE_MS = 10_000;
// ---------------------------------------------------------------------------
// Auth context passed by the MCP route
// ---------------------------------------------------------------------------
export interface DispatchContext {
userId: string;
sessionId: string | null;
}
// ---------------------------------------------------------------------------
// registerPluginToolDispatch
// ---------------------------------------------------------------------------
/**
* Reads the per-session catalog and registers a relay handler for each
* namespaced plugin tool. Called inside the MCP route's registration seam
* after auth, before transport.handleRequest().
*
* When sessionId is null (no active session) or the catalog is empty, no
* plugin tool handlers are registered (system tools remain unaffected).
*/
export async function registerPluginToolDispatch(
server: McpServer,
ctx: DispatchContext,
): Promise<void> {
if (!ctx.sessionId) return;
const catalog = await readSessionCatalog(ctx.userId, ctx.sessionId);
if (!catalog || !Array.isArray(catalog.tools) || catalog.tools.length === 0) return;
for (const tool of catalog.tools) {
const namespacedName = tool.namespacedName;
if (!namespacedName) continue;
// Capture loop variables for the closure.
const capturedTool = tool;
const capturedCtx = ctx;
server.registerTool(
namespacedName,
{
description: capturedTool.description,
// Accept args as a loose record -- actual validation is done by validateToolArgs.
inputSchema: { args: z.record(z.string(), z.unknown()).optional() },
},
async (input) => {
// Build the args object from whatever the MCP client passed.
// The MCP SDK wraps the input in parsed zod fields, so we extract
// a flat record of all non-undefined fields for validation.
const rawArgs = (input as Record<string, unknown>).args as Record<string, unknown> | undefined;
// Flatten: when args is not present, use the input directly (minus internal fields).
// This handles both `{ args: { squawk: "7700" } }` and `{ squawk: "7700" }` inputs.
const argsRecord: Record<string, unknown> = typeof rawArgs === "object" && rawArgs !== null
? rawArgs
: (input as Record<string, unknown>);
// MCP-QA-03: Validate BEFORE enqueue.
// ToolInputSchema requires `type: "object"` but CatalogTool.inputSchema is
// Record<string, unknown>. Cast via unknown to make the compiler happy.
const schema = capturedTool.inputSchema as unknown as ToolInputSchema;
const validation = validateToolArgs(argsRecord, schema);
if (!validation.valid) {
return {
content: [
{
type: "text" as const,
text: JSON.stringify({
error: `Validation failed: ${validation.errors.join("; ")}`,
errors: validation.errors,
}),
},
],
isError: true,
};
}
// Enqueue the invocation for the browser to execute.
const requestId = randomUUID();
const enqueueResult = await enqueueToolInvocation(
capturedCtx.userId,
capturedCtx.sessionId!,
{
requestId,
tool: namespacedName,
args: argsRecord,
},
);
if (enqueueResult.rejected) {
return {
content: [
{
type: "text" as const,
text: JSON.stringify({
error: `Failed to enqueue tool invocation: ${enqueueResult.reason ?? "rejected"}`,
}),
},
],
isError: true,
};
}
// SEC-02 / MCP-QA-04: Wait for browser result with a bounded deadline.
const resultOrTimeout = await waitForToolResult(
capturedCtx.userId,
capturedCtx.sessionId!,
requestId,
RELAY_DEADLINE_MS,
);
if (resultOrTimeout.timedOut) {
return {
content: [
{
type: "text" as const,
text: JSON.stringify({
error: "Plugin tool timed out: no response from the browser before the deadline.",
timedOut: true,
tool: namespacedName,
}),
},
],
isError: true,
};
}
// Sanitize: serialize the value to JSON, never return raw Error objects.
const safeResult = resultOrTimeout.value;
return {
content: [
{
type: "text" as const,
text: typeof safeResult === "string"
? safeResult
: JSON.stringify(safeResult ?? null),
},
],
};
},
);
}
}
Loaded 100 of 197 files, more files were not shown because too many files have changed in this diff. Show more