Simplemart17 · Simplemart17 · May 17, 2026 · May 17, 2026 · May 17, 2026 · May 17, 2026
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
@@ -36,6 +36,48 @@ jobs:
       - name: Tests
         run: npm test -- --no-coverage
 
+      - name: Setup Deno
+        # Story 15-3: install Deno for Edge Function _shared utility tests.
+        # `denoland/setup-deno@v2` is the ACTION installer (second major
+        # release of the installer); `deno-version: v1.x` pins the Deno
+        # RUNTIME. Two independent version axes — don't conflate. Deno 2.0
+        # runtime migration is deferred to a follow-up story. The action
+        # caches the install for faster subsequent runs.
+        # SHA-pinning of this action (and actions/checkout / setup-node)
+        # deferred to `15-3-followup-action-sha-pinning` per R1 BH-1.
+        uses: denoland/setup-deno@v2
+        with:
+          deno-version: v1.x
+
+      - name: Deno tests (Edge Function _shared utilities)
+        # Story 15-3: wire Story 11-3 fetch-with-timeout + Story 12-11
+        # parse-upstream-error Deno tests into CI. Pre-15-3 these were
+        # manual-run only. Now gated on PR merge.
+        #
+        # Test files are named EXPLICITLY (R1 BH-2 / EH-1 vacuous-pass
+        # defense): if either file is renamed or moved, the step fails
+        # loudly with `Module not found` instead of silently passing on an
+        # empty directory match. A future file added to _shared/__tests__/
+        # must be appended here to be picked up.
+        #
+        # Permission flags:
+        #   --allow-net=127.0.0.1 — required by the fetch-with-timeout
+        #     happy-path test which spins up a local HTTP server.
+        #   --no-check — runtime-only verification (R1 EH-10); the
+        #     `npm run type-check` step already covers TS soundness for the
+        #     source modules; Deno's redundant typecheck adds ~5s + a
+        #     drift surface against std lib type updates.
+        # NO `--allow-all` — keep the permission surface minimal.
+        # pgTAP migrations (supabase/migrations/__tests__/*.sql) NOT wired
+        # here — requires a Postgres service container. Deferred to
+        # `15-3-followup-pgtap-ci-wiring`.
+        # Deno deps cache deferred to `15-3-followup-deno-cache` per R1 BH-4.
+        timeout-minutes: 3
+        run: |
+          deno test --no-check --allow-net=127.0.0.1 \
+            supabase/functions/_shared/__tests__/fetch-with-timeout_test.ts \
+            supabase/functions/_shared/__tests__/parse-upstream-error_test.ts
+
       - name: Dependency vulnerability gate
         # Story 12-10: prevent regression of `npm audit reports 0 high
         # vulnerabilities` (Epic 12 AC). Calibrated to `--audit-level=high`

diff --git a/CLAUDE.md b/CLAUDE.md
@@ -201,7 +201,7 @@ A reusable pattern for capping in-memory data structures + UI work-budgets. Two
 
 **Lib unit tests — `srs.ts` SM-2 algorithm + `pronunciation.ts` `identifyWeakSounds` + `cache.ts` core API + `use-dictation.ts` `compareSentences`:** post-Epic-15.1, the FIRST Epic 15 (Test Coverage & QA Infrastructure) story closes the largest pure-function-test gap surfaced by the Story 12-10 audit refresh footnote. Per the coverage inventory at story-creation time, `activity.ts` was already fully tested (Story 9-2 + 12-3); `memory.ts` + `error-tracker.ts` had substantial partial coverage (sanitize via Story 9-4 prompt-injection.test.ts + dedup via Story 11-6 error-tracker-dedupe.test.ts + e2e clustering via 11-6); `pronunciation.ts` had history-cap coverage (Story 12-12) but NOT `identifyWeakSounds`. Story 15-1 lands **GAP-only scope** across **4 pure-function surfaces** (no source-module modifications — test-only): **(1) NEW [`src/lib/__tests__/srs.test.ts`](src/lib/__tests__/srs.test.ts)** (14 cases) pins the full SM-2 algorithm contract — 3 incorrect-response cases (quality 0/1/2 ease-factor deltas −0.80/−0.54/−0.32 + reset to `repetitions=0`, `intervalDays=1`); 3 correct-response cases (quality 3/4/5 ease-factor deltas −0.14/0/+0.10 + interval progression 1 → 6 → round(prev × ef)); 3 ease-factor 1.3-floor clamping cases (boundary + already-at-floor + just-above-floor); 3-consecutive-correct chain verification; 365-day interval cap; midnight-snap `nextReview` math via Date.now() before/after delta check; no-mutation invariant via input-snapshot equality + determinism via twice-call equivalence. **(2) NEW [`src/lib/__tests__/pronunciation.test.ts`](src/lib/__tests__/pronunciation.test.ts)** (10 cases) pins `identifyWeakSounds` contract — empty input + no-weak happy path + below-threshold-but-count<3 NOT flagged + single-phoneme-aggregated-across-results + same-phoneme-aggregated-across-words-in-one-result + threshold boundaries (`avgScore < 70` strict + `count >= 3` boundary) + multi-phoneme ranking ASCENDING by avgScore (worst first) + no-mutation deep-equal. **(3) NEW [`src/lib/__tests__/cache.test.ts`](src/lib/__tests__/cache.test.ts)** (20 cases) covers the AsyncStorage path (existing `cache-flush.test.ts` covers write-queue idempotency; existing `cache-secure-routing.test.ts` covers the Story 12-7 SecureStore fork) — `getCache` empty/fresh/TTL-fresh-boundary/TTL-expired-boundary/corrupted-JSON/throw paths each with `captureError` routing pinned; `setCache` envelope shape `{data, timestamp, ttlMs}` + DEFAULT_TTL_MS (1 hour) fallback + namespacing `@companion_cache:<userId>:<key>` (different userIds never collide); `setCache` → `getCache` round-trip with nested object structure; `invalidateCache` happy + error paths; `cacheWithFallback` 4-path matrix (fetcher-succeeds-writes-cache, fetcher-throws-stale-cache-hit-returns-fromCache:true, fetcher-throws-no-cache-rethrows-original, fetcher-throws-cache-also-throws-captures-fallback-read-error-rethrows-original); constant pins for `CACHE_KEYS` (5 keys) + `CACHE_TTL` (6 durations) + `SECURE_CACHE_KEYS` size===1 with PROFILE in / SKILLS+VOCABULARY out. **(4) NEW [`src/hooks/__tests__/use-dictation-compare.test.ts`](src/hooks/__tests__/use-dictation-compare.test.ts)** (15 cases) pins the pure word-by-word `compareSentences` helper from [`src/hooks/use-dictation.ts:88`](src/hooks/use-dictation.ts#L88) (despite living in a hook file, the function is pure — no React, no async, no side effects, so it fits 15-1 scope). Test mocks `@/src/hooks/use-audio-player` (transitive `expo-audio` native crash in Jest) + `@/src/lib/openai` (transitive supabase chain) at module load to keep the import boundary clean. Cases cover: perfect-match → 100% / one-wrong-word with `typed` populated → round(2/3 × 100)=67% / missing-word semantics (user shorter → past-user-length positions are "missing", overlap positions compare normally) / extra-word (user longer → extra tokens IGNORED; only original-length iterated) / empty-userInput → all "missing" / empty-original → empty wordResults + accuracy=0 + isFullyCorrect=true vacuously / both-empty same / case-insensitive ("Bonjour" vs "bonjour") with display preserving original capitalization / accent-insensitive ("café" vs "cafe") with display preserving accent / trailing punctuation stripped / internal comma stripped / multi-space collapses / leading+trailing whitespace trimmed / apostrophe stripped ("l'eau" matches "leau") / fresh wordResults array per call (no shared module-level cache). **3 operator decisions resolved per Recommended** (Q1 defer `assessPronunciation` Edge wrapper to 15-1-followup; Q2 defer `memory.ts` async/DB-touching paths to 15-2; Q3 defer `error-tracker.ts` non-dedup paths to 15-2) — keeps 15-1 pure-function-only per the Epic 14 retro lesson that test-writing stories are at scope-drift risk. **Explicitly out of scope** (filed for follow-up if motivated): `memory.ts` `extractFacts`/`persistMemories`/`retrieveMemories`; `error-tracker.ts` `getTopErrors`/`extractErrorsFromCorrections`/`persistErrorPatterns`/`getRecentResolvedError`; `pronunciation.ts` `assessPronunciation` Edge Function wrapper; `cache.ts` `enqueueWrite`/`flushWriteQueue` (already in `cache-flush.test.ts`)/`clearUserCache`/`clearAllCache` multi-key sweeps; `use-dictation.ts` `analyzeErrorPatterns`. **Cross-story invariants preserved by construction:** zero source-module modifications (Story 15-1 is test-only); Story 9-3 Sentry allowlist zero-diff (cache tests verify EXISTING `cache-get`/`cache-set`/`cache-invalidate`/`cache-fallback-read` feature tags fire; no new tags); Story 12-7 SecureStore fork unchanged (cache.test.ts uses non-secure key `"skills"` to exercise AsyncStorage path); all Epic 14 invariants orthogonal. **+59 net Jest cases** (2099 → 2158; spec target +50-65 — squarely in range). 4 new test files; 0 modified source files. All 5 design-system gates green (type-check 0 errors / lint 0 warnings / prettier clean / check:tokens clean / jest 113 suites / 2158 tests). Verified 2026-05-16, story 15-1.
 
-**Golden-flow E2E Maestro skeleton flows + operator runbook:** post-Epic-15.4 ships the structural framework for end-to-end smoke testing. **Operator pre-decision at autopilot kickoff: Maestro** over Detox per the spec's recommended default (faster iteration, simpler YAML config, no Xcode/Android Studio project-level integration required). **SKELETON-ONLY scope** because Maestro requires a running simulator/emulator + built app binary + Maestro CLI installed locally — none of which are available in the autopilot session; the cleanest scope-cut is YAML flow files + operator runbook + drift detector, with CI wiring deferred to `15-4-followup-maestro-ci-wiring` after the operator sets up simulator/emulator infrastructure on a dedicated runner. NEW `.maestro/config.yaml` (declares `appId: com.companion.app`) + 5 YAML flow files covering the spec's golden flows: [`.maestro/01-signup-flow.yaml`](.maestro/01-signup-flow.yaml) (sign-up → Story 12-9 EmailVerificationGate); [`.maestro/02-onboarding-flow.yaml`](.maestro/02-onboarding-flow.yaml) (sign-in → 3-step onboarding wizard → placement test 15Q → home); [`.maestro/03-first-exercise.yaml`](.maestro/03-first-exercise.yaml) (Practice tab → Grammar → AI exercise → grade); [`.maestro/04-first-conversation.yaml`](.maestro/04-first-conversation.yaml) (Conversation tab → topic → start Realtime session → mic permission → end — validates SHELL connection + UI state transitions, NOT AI response content since Maestro can't inject audio); [`.maestro/05-mock-test-partial-review.yaml`](.maestro/05-mock-test-partial-review.yaml) (Mock Test tab → TCF full sim → listening section 29Q → submit partial → results screen with Story 13-4 parallel section generation + Story 14-7 results reconstruction). Each flow file carries `# TODO: verify selector` markers throughout — operator must verify testIDs / accessibility labels / text strings against the actual running app via `maestro studio` before first execution. Env-injected test account emails (`e2e-test+<flow>-001@invalid.localdomain` + `Abcdefghi1` password complies with Story 12-8 ≥10-char policy). NEW operator runbook [`_bmad-output/planning-artifacts/runbooks/maestro-e2e-setup.md`](_bmad-output/planning-artifacts/runbooks/maestro-e2e-setup.md) (7 sections) walks: install Maestro CLI → boot iOS Simulator → `npx expo run:ios` dev client → `maestro studio` selector verification → seed verified test accounts (Supabase Studio Admin OR seed script using `email_confirm: true`) → run flows locally → CI wire-up procedure for the deferred follow-up. NEW [`src/lib/__tests__/maestro-flows-source-drift.test.ts`](src/lib/__tests__/maestro-flows-source-drift.test.ts) (5 cases) pins: config.yaml with canonical appId; all 5 flow files exist; each flow has non-trivial body (>20 lines + appId); each tags itself with `smoke`; NEGATIVE — NO Maestro CI step is wired in `ci.yml` yet (15-4 skeleton-only contract; CI wiring should fail loudly when added without the 15-4-followup completing first). **3 operator decisions resolved per Recommended:** Q1 Maestro (pre-decided at autopilot kickoff); Q2 defer CI wiring (`15-4-followup-maestro-ci-wiring` filed); Q3 selector-verification TODO markers (vs hard-coded selectors — hardcoding without `maestro studio` verification is hope-driven). **Cross-story invariants preserved:** Story 12-8 password policy (test accounts use 10-char `Abcdefghi1`); Story 12-9 EmailVerificationGate (Flow 1 verifies it fires; Flows 2-5 assume verified accounts); Story 14-1 chrome rule (flows use English button/tab labels); Story 12-10 R1-H2 silent-disable defense (runbook explicitly forbids `continue-on-error: true` on the future CI step — partial gating is a footgun); 0 source-module modifications outside `.maestro/` + new drift test + runbook. **+5 net Jest cases** (drift only; Maestro flows execute outside Jest via `maestro test` CLI). All 5 design-system gates green. Verified 2026-05-17, story 15-4.
+**Edge Function Deno tests wired into CI:** post-Epic-15.3 partially closes Epic 12 AI #7 + Epic 13 AI #8 (pgTAP CI wiring follow-through — Deno portion shipped; pgTAP portion deferred to `15-3-followup-pgtap-ci-wiring`). Pre-15-3 the 2 Deno test files (Story 11-3 `fetch-with-timeout_test.ts` + Story 12-11 `parse-upstream-error_test.ts`) were manual-run only; regressions could silently survive PR merges. Story 15-3 wires both into [`.github/workflows/ci.yml`](.github/workflows/ci.yml) via new `Setup Deno` step (`denoland/setup-deno@v2`, pinned `deno-version: v1.x`) + `Deno tests (Edge Function _shared utilities)` step running `deno test --allow-net=127.0.0.1 supabase/functions/_shared/__tests__/`. Permission surface minimal — NO `--allow-all`. Ordering: Deno steps AFTER `Tests` so npm-test failures short-circuit before Deno install. NEW [`src/lib/__tests__/ci-deno-step-source-drift.test.ts`](src/lib/__tests__/ci-deno-step-source-drift.test.ts) (5 cases) pins via Story 12-2 P12 + Story 12-10 R1-H2 patterns. TIGHT scope — pgTAP (5 SQL files) DEFERRED to `15-3-followup-pgtap-ci-wiring`; NEW Edge Function tests DEFERRED to `15-3-followup-edge-function-coverage`. 3 operator decisions resolved per Recommended. +5 net Jest cases (drift only; Deno tests run in CI). All 5 design-system gates green. Verified 2026-05-17, story 15-3.
 
 ### Routing (`app/`)