suaveplan/openspec/archive/2026-05-26-api-sandbox-permissions/tasks.md

Tasks — @suaveplan/api-sandbox-permissions

Every checkbox MUST reflect work actually done in this change.

0. Pre-flight (non-negotiable)

• 0.1 Confirm clean working tree (git status reports nothing) and no in-flight worktrees with uncommitted work (git worktree list then git status in each)
• 0.2 If any worktrees exist under .claude/worktrees/, account for each one: still-active agent, work-pending-merge, or stale-and-prune. Stale worktrees MUST be removed (git worktree remove) before opening a new one — code that has been lost twice on main is the failure mode this guards against.
• 0.3 Confirm the package set in MANIFEST.md matches find packages -maxdepth 3 -name package.json count. Drift here means a prior change archived without regenerating the manifest.

1. Spec authoring

• 1.1 Read design doc docs/superpowers/specs/2026-05-13-api-plugin-foundation-design.md (§2.2, §3, §5.2b, §12.4)
• 1.2 Read openspec/AGENTS.md Sections 13 and 14
• 1.3 Draft proposal.md
• 1.4 Draft specs/api-sandbox-permissions/spec.md
• 1.5 Every requirement carries a **Verified by:** line naming concrete it(...) titles

2. Implementation

• 2.1 Scaffold packages/api/api-sandbox-permissions/ ("version": "0.1.0", "sideEffects": false)
• 2.2 src/policy/ — PermissionPolicy value object built from PermissionsManifest
• 2.3 src/checker/fs/ — path-allowlist glob matching, traversal rejection
• 2.4 src/checker/fetch/ — host-allowlist matching (exact + *.suffix), scheme gate
• 2.5 src/checker/capability/ — db scope / secrets name / events topic gating
• 2.6 src/budget/ — BudgetTracker for CPU sample / memory sample / wall-clock
• 2.7 src/error-codes.ts — PERMISSION_DENIED, BUDGET_EXCEEDED via defineCodes
• 2.8 src/instruments/ — plugin.permission_denied_total, plugin.budget_exceeded_total descriptors
• 2.9 src/testing/run-permissions-conformance.ts — shared harness
• 2.10 src/index.ts — explicit named re-exports (no export *)
• 2.11 package.json exports — ./testing subpath

3. Tests

• 3.1 src/policy/policy.test.ts
• 3.2 src/checker/fs/fs.test.ts, src/checker/fetch/fetch.test.ts, src/checker/capability/capability.test.ts
• 3.3 src/budget/budget.test.ts
• 3.4 src/checker/checker.spec-conformance.test.ts — denial metric + error code with real in-memory MeterProvider
• 3.5 src/testing/run-permissions-conformance.test.ts — harness self-test
• 3.6 tests/dep-graph.test.ts
• 3.7 Every it(...) carries { timeout: 5000 }

4. Documentation

• 4.1 Sibling .md for every implementation .ts (≥200 words, basic + advanced examples, full API reference)
• 4.2 README.md links every co-located .md under ## Modules
• 4.3 Update CLAUDE.md package listing

5. Pre-archive verification (the 5-gate triple)

• 5.1 bun run check — Biome 0/0/0
• 5.2 bunx turbo lint typecheck test --filter=@suaveplan/api-sandbox-permissions — green
• 5.3 bunx turbo test:coverage --filter=@suaveplan/api-sandbox-permissions — 100/100/100/100
• 5.4 bun scripts/openspec-conformance.ts @suaveplan/api-sandbox-permissions — exit 0
• 5.5 bunx turbo build --filter=@suaveplan/api-sandbox-permissions — green

6. Pre-archive 7-row checklist (AGENTS.md Section 13)

• 6.1 Generics threaded; expectTypeOf on PermissionPolicy<Manifest> literal-type claims
• 6.2 Zod schemas wired at PermissionsManifest ingestion boundary
• 6.3 PERMISSION_DENIED and BUDGET_EXCEEDED both thrown by at least one production path
• 6.4 Strict-mode behaviour — both enforce and warn deny modes tested
• 6.5 OTel instruments emitted and asserted against a real in-memory MeterProvider
• 6.6 Lifecycle dedup — BudgetTracker is idempotent across repeat samples
• 6.7 No retroactive tick

7. Archive

• 7.1 Fold spec delta into openspec/specs/api-sandbox-permissions/spec.md
• 7.2 Move folder to openspec/archive/<yyyy-mm-dd>-api-sandbox-permissions/
• 7.3 Tick 8.1 in umbrella tasks
• 7.4 Refresh MANIFEST.md