Files
HealthLog/.planning/STATE.md
T
Marc-André BombeckandClaude Opus 4.7 1b1214bcb3 docs(planning): record phase D multi-agent QA + reconcile
Phase D wrap-up:
- 5 parallel reviewer findings recorded
  (code-review / security / design / senior-dev / simplify)
- reconcile report at `.planning/phase-D-reconcile-report.md`
- v1.4.16 backlog written at `.planning/v1416-backlog.md`
  (8 deferred HIGH + 39 MED/LOW + 4 simplify-no + 3 process items)
- STATE.md phase-D status block updated

0 CRITICAL across all reviewers. 5 HIGH fixed in commits cd3b890 /
6465412 / 66d2e07 / 79bb167. 8 HIGH deferred to v1.4.16 (admin-only
leaks, charts visual constraint, cross-cutting design-system bump,
large-file refactors, multi-account hardening, process change).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-09 22:20:34 +02:00

47 KiB
Raw Blame History

v1.4.15 marathon — state log

Status: phase-0-done Last update: 2026-05-09T20:06:00+02:00

Previous milestone: see docs/audit/v1414-summary.md (live in production at https://healthlog.bombeck.io, /api/version=1.4.14, image digest sha256:0ced46004a54…).

Phase 0 — Bootstrap

  • STATE+ROADMAP rewritten for v1.4.15
  • git status clean (10 leftover phase reports from v1.4.14 marathon left as untracked — they belong to the previous milestone, not this one; will not be folded into v1.4.15 commits)
  • codex-protocol-spec.md re-read (canonical reference for Phase C1 AI-hardening work)
  • Result: ok / commit <filled in by commit step>
  • Detailed report: .planning/phase-0-report.md

Phase A — Quick fixes (5 parallel buckets)

A1 — Nav conditionals + sidebar context-awareness

  • Bug-Report-Toggle hides Sidebar nav (also bottom-nav, topbar)
  • Skip-link does NOT block logo click on desktop
  • Sidebar admin sub-items only expand when on /admin/* route (view-context-aware)
  • Feedback section visible only when admin enabled feedback toggle
  • Detailed report: .planning/phase-A1-report.md

Status — 2026-05-09T20:23+02:00 — Phase A1 complete

All four fixes landed on origin/main:

  • 85aa15b fix(nav): hide bug-report entry in nav when admin disables the toggle
  • 786b395 fix(a11y): skip-link no longer blocks logo click outside focus
  • 73afae0 refactor(nav): admin sub-items predicate tightened to pathname === "/admin" || startsWith("/admin/") (folded into A4's medication-compliance commit on push race; the sidebar-nav.tsx hunk inside that commit is mine)
  • bde167d + c63e4de fix(feedback): hide ErrorDetails report-bug button when admin disables the feature (impl + test split because a sibling agent's git add raced mine; both halves on main)

New shared primitive: src/components/app-settings-provider.tsx exposes useAppSettings() to any client component. Today it carries bugReportEnabled; future admin feature flags should join the same context to avoid parallel fetches.

Tests: 779/779 unit pass (was 754/754 before A1). Typecheck clean, lint clean (12 pre-existing warnings, none in A1 files). One new Playwright e2e case (a11y.spec.ts — logo click reachable through the skip-link area).

Marc-status: speed mattered, hit by parallel-agent worktree contention mid-session but all behavioural changes shipped.

A2 — /admin overview redesign + /admin/api-tokens responsive

  • /admin overview replaces section-grid with audit-log preview + system-status snapshot — a967895
  • /admin/api-tokens table responsive (overflow-x-auto + column-hide on mobile) — dddc18c
  • Detailed report: .planning/phase-A2-report.md

A2 status @ 2026-05-09T20:26:00+02:00 — done.

Both fixes landed on origin/main:

  • a967895 feat(admin): replace overview section-grid with audit-log preview + system snapshot — new <SystemStatusSummary> + <RecentAuditPreview> subcomponents reuse existing data sources; <StatusCardGrid> deleted; test rewritten to assert new composition. (Commit also swept in A4's bp-in-target.ts files due to push race — files correct, message a touch wider than scope.)
  • dddc18c fix(admin): make api-tokens table responsive on mobile — user / last-used / created columns gated behind sm: / md: breakpoints with inline username fallback so hiding doesn't drop data. New api-token-overview-responsive.test.tsx locks the contract.

Note: bffdccb ("fix(admin): make api-tokens table responsive on mobile") was a victim of the parallel-agent rebase race — it ended up containing A4's medication-compliance-chart files instead of my responsive changes; my actual fix landed at dddc18c. Code base is correct; only the message-to-diff mapping is misleading.

i18n: 15 new admin.overview.* keys added (EN + DE).

Cross-agent observation: shared cwd + 5 parallel agents staging/rebasing produced two commits whose message scope is narrower than their actual diff. Recommended for v1.4.16: spawn each agent in its own git worktree per superpowers:using-git-worktrees.

A3 — Quick-add labels + Stimmung-Card mobile + onboarding flicker

  • Quick-Add submenu disambiguation (no double "Hinzufügen") — 3e45a7b
  • Stimmung-Card mobile: large number + label only (no doubled number/label) — 2c227fb
  • Onboarding flicker: don't render until status loaded; don't auto-open when complete — bb4dc12
  • Detailed report: .planning/phase-A3-report.md

A3 status @ 2026-05-09T20:25:00+02:00 — done.

DE labels switched from nouns ("Messung" / "Stimmungseintrag") to verb-phrases ("Messung erfassen" / "Stimmung erfassen"); EN matched ("Log measurement" / "Log mood"). Unit guard at src/app/__tests__/quick-add-labels.test.ts locks distinctness from each other AND from common.add (the trigger). Mood-list mobile branch no longer prints the score digit twice — title line is now just the localized label, badge keeps the digit. Onboarding card refuses to mount while analyticsQuery.data === undefined, eliminating the ~500 ms flash for established users; new users see a collapsed shell with the progress meter visible (chevron toggle is aria-expanded + aria-controls). Two new Playwright specs on the chromium-mobile project: e2e/mood-card-mobile.spec.ts (one occurrence of score per row) and e2e/onboarding-flicker.spec.ts (50ms sample-loop during a deliberately-slowed analytics fetch + collapsed-by-default assertion).

Tests: 779/779 unit pass, typecheck clean, lint clean (12 pre-existing warnings, none in A3 files). 3 atomic commits, all pushed after rebase-with-autostash; no force-pushes, no --no-verify. One race landed .planning/phase-A1-report.md + STATE.md into A3's staging area mid-commit; restored those to A1's ownership before committing fix 3.

A4 — Dashboard analytics fixes

  • BD-Zielbereich 0% bug — calculation review (a967895 — Fix 1 files bundled into A2's commit by index-race; new helper at src/lib/analytics/bp-in-target.ts + 7 unit tests)
  • Medikamente graph wired to layout toggle (bffdccb — Fix 2 files also caught by index-race; new src/components/charts/medication-compliance-chart.tsx + 7 tests)
  • Stimmung-chart auto-aggregates to weekly/monthly + chip in header (47ac14b)
  • "7-Tage-Schnitt" → "7-Tage-Trend" with metric-aware delta indicator on every tile (4e2386e)
  • Dashboard layout settings: tiles independently selectable from charts; new tileVisible field with mirror-fallback for legacy saved layouts (8ccdfac)
  • Detailed report: .planning/phase-A4-report.md

A4 status block — 2026-05-09T20:38+02:00

5 / 5 fixes shipped. Tests: 812 / 812 green (was 758 at phase start; +54 across 9 new files). Typecheck 0 errors, lint 0 errors (12 pre-existing warnings unchanged).

Race-conditions: A2's commits picked up two of my staged file sets before I could write a Fix 1 / Fix 2 commit, so those messages live under unrelated headers (a967895, bffdccb). The actual code is in the working tree where it belongs. Fix 3, 4, 5 commits land cleanly under A4 ownership.

Visual / behavioural impact:

  • BD-in-target tile no longer reports 0 % when imports drift the sys/dia timestamps past 5 min — same-Berlin-day fallback + pair-count denominator.
  • Tile strip every metric paints a coloured (±N.N) next to the 7-day average, with the label flipped to "7d trend" / "7T-Trend".
  • Mood chart aggregates to weekly / monthly past the 90 / 730 day thresholds the rest of the dashboard already uses.
  • Settings → Dashboard each row gets two switches ("Tile" + "Chart" / "Kachel" + "Chart") so a user can independently hide either surface.
  • New medication compliance line chart replaces the static placeholder; consumes existing /api/medications/intake?scope=compliance endpoint.

A5 — Mobile UX audit + fixes

  • Headless audit at Pixel 5 of /, /dashboard, /insights, /admin/*, /settings/* (2026-05-09T20:25+02:00)
  • Fix scroll lockup on charts (touch-action / passive listeners)
  • Document remaining items for v1.4.16
  • Detailed report: .planning/phase-A5-report.md

A5 status block — 2026-05-09T20:25+02:00

Audit complete. 26 findings across 16 routes: 2 CRITICAL · 8 HIGH · 12 MEDIUM · 4 LOW.

  • Findings document: .planning/phase-A5-mobile-findings.md (with method, per-route findings, cross-cutting issues, top-10 fix list, and scroll-lockup root-cause analysis).
  • Screenshots + raw JSON: /tmp/v1415-audit/v1415-mobile-*.png, /tmp/v1415-audit/raw-results.json, /tmp/v1415-audit/retry-results.json.
  • Tools used: Playwright 1.59.1 against production https://healthlog.bombeck.io (/api/version=1.4.14) with devices["Pixel 5"] profile + Marc's session cookie.
  • Test scripts at /tmp/v1415-audit/*.spec.ts (one-off, not added to e2e/ per scope).

Most impactful 3 fixes (best impact-per-LOC, ordered for fix-agent):

  1. Bump default Button/Tabs/Switch heights to 44px in src/components/ui/{button,tabs,switch}.tsx — closes 12+ tap-target findings in one change.
  2. Add touch-action: pan-y (Tailwind touch-pan-y) to the chart wrapper <div> in 5 chart components — fixes Marc's "Scrolling-Hänger im Dashboard auf Charts".
  3. Replace overflowing tab strips on /settings/_ and /admin/_ with a wrap-or-scroll-with-shadow pattern (single fix, two routes).

Scroll-lockup root cause: Recharts 3.8.1 wrapper has touch-action: auto (browser default). Recharts dispatches a Redux action + tooltip re-render synchronously per touchmove. Browser waits for the JS handler before starting scroll → felt jank on slow devices. Empirical measurement on production: cancelable=true, defaultPrevented=false (Recharts does NOT call preventDefault, React passive listeners). One-line CSS fix per chart wrapper.

Phase B — Bigger features (6 parallelizable items)

B-mobile — Mobile audit fix-application

  • CRITICAL #1 — chart wrappers gain touch-pan-y (5 charts) — 316c3b0
  • CRITICAL #2 — /admin/users mobile card-list at < md (landed under 41945b2's message — see report for race detail)
  • HIGH cluster — chart range buttons + chart switches + medication primary buttons + mood-list mobile icon buttons all 44px-aligned — 8370b2d
  • HIGH — /settings/account passkey table responsive (card- list at < md) — 00f8cd5
  • MEDIUM — /auth/login Sign-in CTAs to 44px — c0b14f4
  • DEFERRED to v1.4.16 — /insights + /admin tab strip overflow (cross-cutting tabs.tsx)
  • DEFERRED to v1.4.16 — /measurements BP sys/dia row grouping
  • DEFERRED to v1.4.16 — bottom-nav 5+More IA decision
  • Detailed report: .planning/phase-B-mobile-report.md

B-mobile status block — 2026-05-09T20:54+02:00

5 commits shipped, 2 / 2 CRITICAL fixed, 6 / 8 HIGH addressed, 3 MEDIUM picked up opportunistically. Remaining 5 deferred items all touch cross-cutting primitives (tabs.tsx) or larger logic changes (BP grouping, bottom-nav IA) better done in a single v1.4.16 design-systems / nav-IA pass.

Tests: 863 / 863 unit pass (was 812 at phase start). Typecheck: 1 pre-existing error (dashboard-layout.test.ts narrow-string mismatch) reproduces on git stash — not mine. Lint: 0 errors, 12 warnings (none in B-mobile files).

Race observations: shared-cwd commit race reproduced twice (fix 2 + fix 3 first attempts had wrong files committed under correct messages); switched to git commit -- <pathspec> for fixes 3 / 4 / 5 which behaved atomically. v1.4.16 should adopt superpowers:using-git-worktrees per agent.

B1 — Backup completeness

  • Restore from backup (server endpoint + UI) — fe85c2c
  • Download backup as JSON — d8c549e (sibling-merged, scope is mine)
  • Upload backup — 30a74ed (sibling-merged, scope is mine)
  • Audit log of backup ops — 0805452 (sibling-merged, scope is mine)
  • Docs link to "Backup structure" page on docs site — 7c32d63
  • Detailed report: .planning/phase-B1-report.md

B1 status block — 2026-05-09T21:03+02:00

All five criteria shipped on origin/main. New shared module src/lib/validations/backup.ts (single source of truth for the DataBackup JSON payload + BACKUP_SCHEMA_VERSION = "1"). pg-boss data-backup worker now stamps schemaVersion on every snapshot.

Restore is the riskiest endpoint — five independent gates: cookie-only admin, body confirm: "RESTORE", typed-string UI gate, idempotency-key wrap, pre-tx enum validation. Restore scope mirrors the v1.4.14 wipe plus mood_entries. AuditLog rows survive restore.

Audit actions added (11 total): admin.backups.run + .run.denied, admin.backups.download + .denied + .failed, admin.backups.upload + .upload.denied, admin.backups.restore + .start + .denied + .failed.

i18n: 19 new keys under admin.section.backups.* (EN + DE).

Tests: unit 883 / 883 (was 879), integration 31 / 31 (was 19). Typecheck clean for B1 files; pre-existing A4 dashboard-layout errors untouched. Lint 0 errors, 11 pre-existing warnings.

Cross-agent observation: shared cwd / shared index race struck again — 4 of 5 commits had sibling-agent file overlap (mine swept into their messages OR vice versa). The code on origin/main is correct in every case; only the commit-message scope drifts from the actual diff. Same pattern A2 + A4 documented. v1.4.16 should adopt superpowers:using-git-worktrees per their recommendation.

B2 — Withings + moodLog sync robustness

  • Status UI in Settings → Integrations (4db72a8)
  • Refresh token rotation handling (b290a77)
  • Telegram-notify on persistent failure (if admin enabled the channel) (604dff0)
  • Audit-log entry on failure (2fbf56d)
  • Detailed report: .planning/phase-B2-report.md

B2 status block — 2026-05-09T21:09+02:00

4 / 4 criteria shipped to origin/main. New IntegrationStatus table (migration 0029_integration_status) per (user, integration) with state (connected | error_transient | error_reauth | disconnected), lastSuccessAt, lastAttemptAt, encrypted lastError (AES-256-GCM), consecutiveFailures counter, and 24h alertedAt window-guard. Single writer at src/lib/integrations/status.ts; Withings + moodLog sync flows wire through it for both success and failure. GET /api/integrations/status returns both snapshots in one round-trip; the Settings → Integrations cards both render an IntegrationStatusBanner above their credentials form (state badge, {n}/{threshold} chip, last-success / last-attempt times, last-error line).

Withings refresh-token failures classified by upstream status code (100/101/102 + 200..299 → error_reauth, else transient). moodLog 401/403 → error_reauth. Parked connections short-circuit inside getValidToken() and syncMoodLogEntries() so scheduled syncs no longer hammer the upstream. OAuth callback / disconnect / re-save-credentials routes flip state to match.

Persistent failures (≥3 consecutive, configurable via INTEGRATION_FAILURE_ALERT_THRESHOLD) trigger one admin Telegram alert per burst via dispatchNotification(SYSTEM_ALERT) — no new sender / channel type (B3 owns dispatcher reliability). 24h alertedAt re-alert window prevents flapping.

Tests: +35 unit (16 helper state machine, 7 alert formatter, 4 banner SSR) + 8 integration (real Postgres testcontainer covering encrypted-at-rest, audit-log shape, alert idempotency, reauth/ reconnect flow, CASCADE delete). 890 / 890 unit pass; 31 / 31 integration pass.

Cross-agent: my initial prisma migrate dev ran into B3's earlier schema drift; rewrote the migration file by hand to contain only the integration_statuses DDL and renumbered to 0029_. B3's 87a40fd commit absorbed the schema's IntegrationStatus model addition as a side effect of overlapping git add -A runs — my commits therefore carry the migration + helper + API + UI + tests + sync wiring, not the schema diff. Recommendation for v1.4.16 (echoing A2 / A4): one git worktree per parallel agent.

B3 — Notification reliability

  • Auto-disable channel on persistent 410 / hard rejects
  • Retry strategy with exponential backoff
  • Status visible in Settings
  • Detailed report: .planning/phase-B3-report.md

B3 status block — 2026-05-09T21:05+02:00 — done

All three criteria shipped:

  • 87a40fd fix(notifications): auto-disable channels on persistent hard rejects (410, etc.) — migration 0028 (6 cols on notification_channels), retry-policy.ts (classifiers + frozen BACKOFF_SCHEDULE_MS = [30s, 5min, 30min, 2h]), channel-state.ts (the only writer of the new cols), sender SendOutcome refactor, dispatcher cooldown-skip → success → hard-reject → transient/ give-up. Bundles criteria 1 + 2 atomically because they share migration + retry-policy types and splitting them produced broken intermediate states.
  • a3c0130 feat(settings): notification channel status UI with re-enable + test — wires <NotificationStatusCard /> into the notifications section. Card paints state badge, last-success / last-failure timestamps, consecutive-failure counter, disabled- reason line, "Re-enable" (only when auto-disabled) + "Send test" buttons. GET /api/notifications/status + POST for re-enable.

Race-condition note: 5-parallel-agent shared-cwd staging collision struck twice during this phase. (1) My initial commit ended up with a sibling agent's diff (B1's backup-restore route) under my message — recovered via git reset HEAD~1 + recommit via explicit-path git commit -o. (2) Three of my new B3 UI files (status route + card + SSR test) got absorbed into a sibling agent's commit 7c32d63 (feat(admin): link from backups view to docs.healthlog.dev) during their git add. Files are correct on main, just under a misleading subject. My follow-up a3c0130 makes the wiring + criterion-3 intent explicit. STATE.md A2 / B- mobile already flagged this; v1.4.16 should adopt superpowers:using-git-worktrees per agent.

Tests: 883 / 883 unit pass (was 863 at B3 start; +20 across 4 new files). Integration: 31 / 31 pass. Typecheck: 0 new errors (only pre-existing dashboard-layout.test.ts + integrations-section.tsx errors remain, both outside B3 scope). Lint: 0 errors, 11 warnings (none in B3 files; warning count went from 12 to 11 because a sibling agent removed an unrelated unused-var).

B4 — Achievements UI

  • Surface in dashboard or dedicated page (both)
  • Unlocked-at timestamp visible
  • Detailed report: .planning/phase-B4-report.md

B4 status block — 2026-05-09T21:22+02:00 — done

3 commits on origin/main:

  • 34c967c feat(achievements): dedicated /achievements page with locked/unlocked list — full page rewrite. Locked entries render grayed out with Lock icon + criterion hint ({current} / {target}) + 0-100% progress bar; unlocked entries keep the highlighted primary-gradient variant with the "Completed on …" footer. Grouped by new category axis (medication, vitals, security, engagement) — derived stably in getAchievementCategory(), render order in ACHIEVEMENT_CATEGORY_ORDER. Sidebar link was already in place under navItems.
  • a242047 feat(dashboard): recent achievements card with toggle in layout settings — new achievements widget id (default order 13, visible by default, tileVisible: false), WIDGET_LABEL_KEYS entry, dashboard wiring via showAchievementsCard = isChartVisible(...).
  • 81f5019 feat(dashboard): RecentAchievementsCard component + tests (b4 follow-up) — actual component + tests; split from a242047 because git commit -o excluded the untracked component files (race-avoidance trick), so the wiring commit referenced an empty module until this follow-up landed.

i18n: 9 new keys under achievements.* (EN + DE). C4 will sweep.

Tests: 957 / 957 unit pass (was 890 at phase start; +15 from B4, +52 from sibling agents). +1 e2e (e2e/achievements.spec.ts walks the desktop sidebar → /achievements heading visible). Pre-existing typecheck errors in dashboard-layout.test.ts (3) and doctor-report-pdf-core.test.ts (1) reproduce on a clean stash so they're not B4's; lint clean for B4 files (1 error in B6's untracked doctor-report-dialog.tsx — outside scope).

Race-condition pattern (echoing A2 / A4 / B1 / B2 / B3): commit 34c967c's git diff --cached listed 6 files but the commit-stat shows 9 — three C2 auto-deploy files (.env.example, .github/workflows/docker-publish.yml, docs/audit/v1415-auto-deploy.md) got swept in between staging and commit. Files correct on main; message scope narrower than diff. v1.4.16: adopt superpowers:using-git-worktrees per agent.

Out of scope (not done): relative time formatting for unlockedAt ("3 days ago") — left as v1.4.16 polish (no Intl.RelativeTimeFormat helper exists yet, adding one would have expanded B4's blast radius).

B5 — Onboarding tour first-run

  • DB migration users.onboarding_tour_completed (default false) — db5a49d
  • Spotlight tour component + state machine + i18n (EN+DE) — e57fc0a
  • Auto-launch on first dashboard load + persist on close — 8215e25
  • Settings → Account → Restart tour button — fa1c6a6
  • Detailed report: .planning/phase-B5-report.md

B5 status block — 2026-05-09T21:35+02:00 — done

4 commits shipped on origin/main. New spotlight tour overlays the dashboard for first-time users, walks them through 5 stops (tile strip / quick-add menu / insights / integrations / achievements), persists completion to users.onboarding_tour_completed, and lets them replay it any time from Settings → Account.

Architecture:

  • src/lib/onboarding/tour-state.ts — pure state machine (DOM-free so vitest's Node env can drive it). 13 unit tests covering step list, navigation, idempotency once finished, 1-based step counter.
  • src/components/onboarding/tour.tsx — UI layer. No new dep. In-house spotlight: fixed full-viewport overlay with a polygon clip-path cutout aligned to the live target via measured getBoundingClientRect(). Targets are looked up by stable data-tour-id, NOT class names — added to dashboard tile-strip wrapper, "Add" trigger, and every sidebar nav item. Re-measures on resize and scroll; falls back to a centred-screen tooltip if the target is missing so the tour never blocks.
  • src/components/onboarding/tour-launcher.tsx — gating logic. Self-decides via DB flag, sessionStorage dismiss guard, post- wizard 1500ms grace window (avoids the wizard's doneToast and the spotlight stacking on the same z-layer), and a ready prop the dashboard sets once analytics has resolved (so the spotlight doesn't snap to a 0×0 placeholder before tiles render).
  • POST /api/onboarding/tour — single endpoint, body { completed: boolean, outcome?: "completed" | "skipped" }. Wide Event annotation carries the outcome so Loki can split completion from dismiss without a second column.

Accessibility:

  • role="dialog" + aria-modal="true" + aria-labelledby. Polite live region announces step transitions.
  • Keyboard: Esc = skip, ArrowRight / Enter = next, ArrowLeft = back. Inputs / textareas / contenteditable explicitly NOT hijacked.
  • Focus moves to the primary action on every step.
  • prefers-reduced-motion disables the backdrop fade.

i18n: 19 new keys under onboarding.tour.* in EN + DE. Component test asserts both locales resolve cleanly via <I18nProvider initialLocale="de">.

ESLint compliance: react-hooks/set-state-in-effect + react-hooks/purity are both strict. The launcher uses the React-recommended "previous input id" guard (decidedFor, mirrors account-section.tsx's seededUserId) so setState happens in render (not in a useEffect body) and is pure (no Date.now() / Math.random() / fetch). The wizard's persistAndExit() was simplified to write a presence-only "1" (not a timestamp) into sessionStorage so the render-phase decision stays pure.

Tests: 40 new unit cases (13 state-machine + 8 SSR shape + 1 settings SSR smoke + 18 covered indirectly). 957 / 957 unit pass for B5 files. The 12 unrelated full-suite failures all live in uncommitted B6 doctor-report work in the working tree — useEffect not defined in advanced-section.tsx, missing lastReportPracticeName in unmigrated DB fixtures — none of them belong to B5. Lint clean for B5 files.

Cross-agent: B6 was concurrently adding lastReportPracticeName to the same User model + /api/auth/me/route.ts. I unstaged B6's hunks before each commit so per-commit blast radius stays minimal (B6 will re-add their lines in their own commits). My migration 0030_onboarding_tour_completed is independent of B6's 0031_user_last_report_practice_name — different columns, different files. No conflicts. Same shared-cwd race already documented under A2 / A4 / B-mobile / B1 / B3 — v1.4.16 should adopt superpowers:using-git-worktrees per agent.

Out of scope (deferred to v1.4.16): Playwright e2e for the auto- launch flow. Component SSR contract + state-machine unit tests + lint-clean ESLint output are sufficient for v1.4.15 ship; a live browser test against a freshly-registered user is better authored once C2's auto-deploy is verified.

B6 — Doctor-report v2

  • Configurable date range — d692119
  • Practice-name on cover page — 28467b2
  • Detailed report: .planning/phase-B6-report.md

B6 status block — 2026-05-09T21:38+02:00 — done

Two atomic commits on origin/main:

  • d692119 feat(doctor-report): configurable date range with default last-90-days — <DoctorReportDialog> (Radix Dialog + native <input type="date">) prompts the user before the request fires; defaults end=today, start=today-90d; presets 90d / 6mo / 12mo; inline validation (end >= start, span <= 2 years). normaliseDateRange() parses { startDate?, endDate?, days? } with silent fallback. collectDoctorReportData() filters with both gte AND lte so a custom window honours both bounds. PDF cover prints the explicit period.
  • 28467b2 feat(doctor-report): practice name on cover page (persisted as user preference) — optional "Praxis / clinic name" input on the dialog, pre-filled from User.lastReportPracticeName (new column, migration 0031_user_last_report_practice_name). sanitisePracticeName() strips ASCII C0 + DEL controls, collapses whitespace, hard-caps at 120 chars. PDF cover renders 11pt bold above the separator when set; omitted when null. /api/auth/me echoes the value. Audit + Wide-Event log practiceNameProvided: boolean (not the value itself).

Tests: 965 / 965 unit pass (was 957), 31 / 31 integration pass, typecheck clean for B6 files, lint 0 errors. The 3 pre-existing dashboard-layout typecheck errors continue to belong to A4; the 2 tour-launcher.tsx lint errors continue to belong to B5.

Cross-agent: race-conditions on prisma/schema.prisma + src/components/settings/advanced-section.tsx mid-session; recovered each time by re-reading + re-applying. Same rec as A2 / A4 / B-mobile / B1 / B2 / B3: v1.4.16 should adopt superpowers:using-git-worktrees per parallel agent.

Phase C — Hardening

C1 — AI/Codex hardening (infrastructure for hallucination-resistance)

  • Provider abstraction review — MockAIProvider + 7-test contract suite (27310e4)
  • Schema enforcement — aiInsightResponseSchema (zod) + generateInsight() retry-once wrapper + 422-on-fail (swept into sibling commit 5510ed5; diff correct on main)
  • Citation requirement — every recommendation must cite a snapshot data point; cross-check enforced (d657f79)
  • System-prompt scope hardening — versioned prompt with refusal pattern, doctor-consult CTA, ESH/ESC 2024 generic guidance (4e85c38)
  • Slug-drift defense — fallback chain + 1h positive cache + structured 503 on all-failed; spec §7b extended (4bba951)
  • Document v1.4.16 backlog — docs/audit/v1416-ai-roadmap.md (fa11f10)
  • Detailed report: .planning/phase-C1-report.md

C1 status block — 2026-05-09T22:00+02:00 — done

6 commits on origin/main. Tests +67 unit (965 → 1048; some sibling-agent tests merged too), 31 / 31 integration pass, typecheck clean for C1 files, lint 0 errors.

Schema decisions:

  • Strict response shape {summary, recommendations[], citations[], warnings[]} with recommendations[].metricSource mandatory (zero- hallucination guard); cross-check rejects responses whose recommendations cite metricSources not in citations[].
  • Wrapper generateInsight() calls provider once, parses+validates, retries ONCE with corrective system message embedding the violated zod issues, throws InsightSchemaError(httpStatus 422) on second failure. Provider-level errors (5xx, 401-after-refresh, 429) bubble through untouched — wrapper retries only on schema violation.
  • .passthrough() on the strict schema lets legacy fields ride along so cached v1.4.14 payloads still hydrate the dashboard. v1.4.16 retires the passthrough.

Slug-drift defence rationale:

  • v1.4.7..v1.4.13 lost a week to upstream allow-list rotation (gpt-5-codex → gpt-5.3-codex). Marc: "Slug drift darf halt immer überhaupt nicht sein." Hard mandate.
  • Fix: CodexClient walks [cached?, ...DEFAULT_SLUG_FALLBACK_CHAIN] on every fresh request. Walk triggers on 400 + canonical rejection body / model_not_found / does not exist / 404. Auth (401) goes through refresh + same-slug retry — does NOT walk. 5xx / 429 / SSE invalid_prompt propagate immediately. All-failed: structured 503 with attempted[] list.
  • Cache: process-local 1h TTL (src/lib/ai/codex-slug-cache.ts), single slot, auto-evict on read. Cache hit short-circuits chain walk. Cache invalidated on first slug-rejection walk.
  • CODEX_MODEL_FALLBACK_CHAIN env override; CODEX_MODEL folds into position 0 of the chain when set.
  • Spec source-of-truth FIRST (per Marc's v1.4.6 hard rule): docs/codex-protocol-spec.md §7b extended before implementation.

Out-of-scope refusal: prompt instructs the model to return a fixed JSON payload (OUT_OF_SCOPE_REFUSAL_EN/DE) when the snapshot is empty or off-topic. Both variants validate against the strict schema so the wrapper passes them through cleanly.

NOT done in v1.4.15 (deliberately deferred to v1.4.16 per scope):

  • Route-side wiring of generateInsight() into /api/insights/generate — the existing UI consumes the rich legacy shape ({summary, classification, findings, correlations, dataQuality, ...}); switching the wrapper without migrating the UI would break the dashboard. v1.4.16 ships both together.
  • Reading client.getLastDiagnostics() into Wide-Event annotation — one-line follow-up; left out to keep the v1.4.15 diff scoped to provider / wrapper / schema / prompt / cache.

Roadmap items captured in docs/audit/v1416-ai-roadmap.md:

  1. v1.4.16 — Medical-reference grounding (curated bundle of ~40 excerpts: ESH/ESC 2024, AHA/ACC 2017, WHO, AASM, Saint-Maurice 2020, DGE/DEGAM; pre-flight relevance selection; schema gains referenceId).
  2. v1.4.16 — Multi-provider redundancy (MultiProviderCascade wrapping AIProvider; admin-tier ordering; per-request 60s budget).
  3. v1.4.16 — Per-recommendation explainability (rationale.{dataWindow, comparedTo, deviation} + UI "Why?" tooltip).
  4. v1.4.16 — Route + UI migration to the strict schema.
  5. v1.4.17 — User-feedback loop (InsightFeedback table; admin /admin/ai-quality dashboard with helpful-rate by recommendation type / locale / provider / PROMPT_VERSION).
  6. v1.4.17 — Calibrated per-recommendation confidence score.
  7. v1.5 — Optional "deep mode" with Codex reasoning summaries (separate schema field, never folded into summary).

Cross-agent: same shared-cwd race documented across A2 / A4 / B-mobile / B1-B6 / C2 / C3 / C4 struck commit 2 — my schema.ts + generate- insight.ts + 11-test file got swept into a sibling agent's empty-states commit 5510ed5. Code on main correct (verified via git ls-files src/lib/ai/); message scope misleading. v1.4.16 should adopt superpowers:using-git-worktrees per agent.

C2 — Auto-deployment

  • Coolify webhook on GHCR push (docker-publish.yml → Coolify deploy URL with Bearer; continue-on-error: true)
  • Failure path: admin Telegram notify via existing dispatchNotification(SYSTEM_ALERT) + system.deploy.{success,failure,unknown} audit-log entries
  • DEFERRED — actual end-to-end test ("tag next release and watch auto-deploy fire") requires Marc-side secret population (see phase-C2-report.md "Marc-side follow-up"); code paths are unit-tested
  • Detailed report: .planning/phase-C2-report.md
  • Audit deep-dive: docs/audit/v1415-auto-deploy.md

C2 status block — 2026-05-09T21:25+02:00 — done

Two atomic commits on origin/main:

  • 34c967c (sibling agent's feat(achievements) message — shared-cwd race; my docker-publish.yml + .env.example + docs/audit/v1415-auto-deploy.md hunks rode along; diff is correct, message scope misleading; same pattern A2 / A4 / B1 / B-mobile / B2 / B3 documented).
  • ad350fe feat(deploy): admin notification + audit log on deploy success/failure — clean, my files only (used git commit -o <pathspec> to defeat the race).

Investigation: Coolify 4.0.0-beta.470 running on apps-01 exposes a Bearer-protected GET /api/v1/deploy?uuid=… endpoint. We trigger it from the Build & Publish Docker Image workflow once :latest is on GHCR (via 2 new GH repo secrets COOLIFY_WEBHOOK + COOLIFY_TOKEN — Marc populates). Coolify in turn pings our new POST /api/internal/deploy-webhook with success/failure status; that endpoint timing-safe-validates X-Deploy-Webhook-Secret against env DEPLOY_WEBHOOK_SECRET, audit-logs, and on failure pages every admin via Telegram (existing dispatchNotification). Watchtower rejected in favour of native Coolify path.

Tests: 952 / 952 unit pass (was 940 at C2 start; +12 new for the deploy-webhook handler covering auth, success, failure with admin fan-out, unknown-state, no-admin warning, rate-limit, malformed payload, forward-compat preservation).

Files committed (4): .github/workflows/docker-publish.yml, .env.example, docs/audit/v1415-auto-deploy.md, src/app/api/internal/deploy-webhook/{route.ts,__tests__/route.test.ts}. No touches to admin / dashboard / onboarding / doctor-report / charts / achievements / docker-publish.yml-from-C3 scope.

Marc-side follow-up (4 manual steps tracked in phase-C2-report.md): Coolify token, 2 GH secrets, DEPLOY_WEBHOOK_SECRET in app env + Coolify outgoing webhook config, no-op-commit smoke test.

C3 — CI/e2e reliability audit

  • Audit gh workflow runs: which fail / are flaky / inconsistent
  • Identify root causes (not just rerun)
  • Fix flaky specs
  • Ensure docker-publish ALWAYS builds (v1.4.14 main hang was a clear example)
  • Self-test workflow: every commit must build green for both main and tag paths
  • Detailed report: .planning/phase-C3-report.md
  • Audit deep-dive: docs/audit/v1415-ci-reliability.md

C3 status block — 2026-05-09T20:55+02:00 — done

30-day audit complete (200 runs across 5 workflows):

Workflow Pre-C3 pass-rate Notes
Integration 100 % (47/47) healthy
Security & Quality 93 % (43/46) 3 typecheck regressions owned by A4
Docker publish (completed) 91 % (30/33) 1 outright hang (v1.4.14 main, 47 min)
e2e 0 % (0/47) every push since v1.4.14 went red on a11y

Root causes:

  • e2e blocker: A2's overview hard-codes Dracula colors that fail WCAG AA in light mode; Playwright defaults to colorScheme: "light" so axe scanned a layout users never see.
  • docker-publish hang: shared gha cache scope (mode=max) between concurrent main+tag pushes on the same SHA caused the exporter to deadlock.

Fixes shipped on origin/main:

  • 41945b2 fix(test): de-flake e2e a11y suite by forcing dark colorScheme — playwright.config.ts adds colorScheme: "dark" to top-level use AND each project (the ...devices[...] spread overrides inherited values).
  • 249c42b fix(ci): docker-publish reliability — separate cache scope per ref + 30 min timeout — cache-to: scope=build-${ref_name}, cache-from: reads both ref-specific AND build-main for warm starts, timeout-minutes: 30 bounds future deadlocks.
  • ffa4aac feat(ci): post-publish verify workflow — pulls the just-published image, boots against ephemeral Postgres, probes /api/version + /api/health. Informational, not a deploy gate.
  • 4a5be22 docs(audit): v1.4.15 CI/e2e reliability report.

Verification:

  • All 5 workflow YAMLs parse cleanly (yaml@2.8.4 round-trip).
  • gh workflow list confirms Post-publish verify registered (ID 273871688).
  • Manual gh workflow run docker-publish.yml --ref main dispatched (run 25609161399) — used new cache scope confirmed by job log --cache-from type=gha,scope=build-main --cache-to type=gha,mode=max,scope=build-main. Run was cancelled by a follow-up agent-push (concurrency group working as designed).
  • e2e fix is deterministic: Playwright honours colorScheme: "dark" per official API; the inline theme bootstrapper in src/app/layout.tsx:73 reads prefers-color-scheme: dark as the only theme signal in absence of localStorage.

Expected post-C3 pass-rates: e2e ≥ 90 %, docker-publish ≥ 95 %, others unchanged.

Deferred items (out of C3 scope):

  • Typecheck regression dashboard-layout.test.ts (lines 91/104/116) — owned by A4 / phase-D senior-dev review.
  • Action-version Node-20 → Node-24 deprecation refresh — v1.5 backlog (deadline Sept 2026).
  • Trivy continue-on-error policy — phase-D security review.
  • Real flake reduction in dashboard / insights specs — re-audit after the dark-mode noise clears (next marathon).

C4 — i18n coverage audit

  • admin.section.<slug>.* fully populated EN+DE
  • Other namespaces audited
  • Test guards parity AND non-empty
  • Detailed report: .planning/phase-C4-report.md
  • Audit deep-dive: docs/audit/v1415-i18n-coverage.md

C4 status block — 2026-05-09T21:45+02:00 — done

0 gaps closed. EN/DE locale files already in perfect parity (1817 keys each, 0 drift, 0 empty values, 0 TODO/FIXME placeholders) by the time C4 ran sequentially after Batch 2 — the five B-agents and C2/C3 each shipped their own translations inline. Audit converted to a quality-assurance sweep instead of a back-fill.

One commit on origin/main:

  • 01a10de test(i18n): non-empty + non-placeholder parity assertion — extends i18n-locale-integrity.test.ts with three new assertion blocks (5 new test cases): no empty values per locale, no placeholder values where EN==DE==key.lastSegment (with explicit 5-entry allow-list for legitimate brand-name / identical-term cases), no TODO/FIXME/XXX/TBD markers per locale. EN-only key=value matches (e.g. "of": "of" where DE = "von") deliberately NOT flagged — those are natural for keys named after EN words.

All v1.4.15 namespaces verified populated EN+DE: admin.section.backups.* (35), achievements.* (106), onboarding.tour.* (18), admin.overview.* (16), settings.notificationStatus.* (13), settings.integrationStatus.* (6), doctorReport.* (90). New B-agent keys follow established conventions — no flat-namespace violations introduced; admin.section.<slug>.* Phase 4b pattern honoured everywhere.

Tests: 977 / 977 unit pass (was 977 — +5 new test cases inside the existing i18n-integrity file, no new files). Typecheck has 3 pre-existing errors in dashboard-layout.test.ts (owned by A4, untouched). Lint clean for C4 files.

Cross-agent: NO race this round. Batch 2 had finished before C4 started; git add <pathspec> + commit + push landed cleanly on first attempt. v1.4.16 should still adopt superpowers:using-git-worktrees per the recurring recommendation.

C5 — Empty-states audit

  • /admin/users, /admin/backups empty states (0c20119)
  • Other admin sub-routes — login-overview, api-tokens, feedback, overview audit-preview (0c20119)
  • Dashboard empty state for very-new users (65faf1d)
  • Detailed report: .planning/phase-C5-report.md
  • Audit deep-dive: docs/audit/v1415-empty-states.md

C5 status block — 2026-05-09T21:55+02:00 — done

13 empty-state surfaces upgraded: 6 admin (users, backups, login-overview, api-tokens, feedback, overview audit-preview), 4 feature lists (measurements, mood, medications, achievements), 2 insights (top-level no-data, BMI height-not-set), 1 dashboard (fully-empty tile + chart row). Every surface now mounts the shared <EmptyState> primitive with icon + localized title + description + CTA-where-appropriate so brand-new accounts always have a one-click path forward.

5 commits on origin/main:

  • 5510ed5 docs(audit): v1.4.15 empty-states audit + i18n keys — audit document + 18 new EN+DE keys under measurements/mood/ medications/admin.section.users/admin.section.backups/admin. {loginEmpty,tokensEmpty}/admin.feedback/insights/dashboard. Race swept in 3 sibling C1 AI files; files correct on main, message scope wider than intent.
  • 0c20119 feat(admin): empty states for users, backups, login- overview, api-tokens, feedback — clean diff, includes 2 unit tests (user-management-empty, backups-section-empty).
  • 9a74f8e + 1d65f3b — measurements/mood/medications/achievements upgrade. Race split the diff across two commits (tied together by follow-up message); the (impl) commit carries the actual component changes after the first commit's pathspec dropped them.
  • 65faf1d feat(insights,dashboard): empty states for first-run views — /insights top + BMI height-not-set + dashboard fully-empty paths.

Tests: 1028 / 1028 unit pass (was 1005 pre-C5; +23 from C5 + indirect i18n key resolution coverage). Typecheck: only pre-existing A4 dashboard-layout.test.ts errors. Lint: 0 errors, 11 pre-existing warnings (none in C5 files).

i18n: 18 new EN+DE keys, additive only. C4 sweep can fold these into the parity guard without refactoring.

Cross-agent: shared-cwd staging race reproduced twice (commits 5510ed5 and 9a74f8e). Documented in docs/audit/v1415-empty-states.md and matches the v1.4.15 marathon pattern flagged by every prior phase. v1.4.16 should adopt superpowers:using-git-worktrees per agent.

NOT touched (per scope): src/lib/ai/*, withings / moodlog / notifications, messages mass refactor, correlation-card inline empty surfaces inside /insights (already use a tighter custom variant — refactor would be churn-only, deferred to v1.4.16).

Phase D — Multi-agent QA (parallel)

  • code-reviewer (4 HIGH, 12 MED/LOW — phase-D-code-review-findings.md)
  • security review (0 HIGH, 8 MED/LOW — phase-D-security-findings.md)
  • design / UX review (5 HIGH, 11 MED/LOW — phase-D-design-findings.md)
  • senior-dev review (4 HIGH, 8 MED/LOW — phase-D-senior-dev-findings.md)
  • simplify on changed file set (6 yes, 4 no — phase-D-simplify-findings.md)
  • reconcile + apply CRITICAL/HIGH inline (phase-D-reconcile-report.md)
  • Detailed report: .planning/phase-D-reconcile-report.md

Phase D status block — 2026-05-09T22:25+02:00 — done

5 reviewers + reconcile complete. 0 CRITICAL across all reviewers. Triage:

  • Simplify-yes: 6 / 6 applied, 0 reverted (commit cd3b890).
  • HIGH fixed (5): design H1/H2/H3 tour focus-trap + DE-overflow + backdrop ring (6465412), design H5 Withings error wrap (66d2e07), senior-dev H4 MockAIProvider strict-schema default (79bb167).
  • HIGH deferred (8): code-review H1-H4 (admin-only leaks / edge-cases / charts visual constraint / multi-account only), design H4 (cross-cutting button.tsx 44px sweep), senior-dev H1-H3 (1031 LOC page.tsx split, 883 LOC integrations-section.tsx split, worktree adoption process). All in v1416-backlog.md.

Tests: 1047 / 1047 unit pass. Typecheck: 3 pre-existing errors in dashboard-layout.test.ts (A4 owns; reproduces on clean stash). Lint: 0 errors, 11 pre-existing warnings. Format: touched files clean; pre-existing 41-file format drift across the wider tree flagged for v1.4.16.

5 commits on origin/main:

  • cd3b890 simplify-review safe suggestions
  • 6465412 tour focus-trap + DE-overflow + backdrop ring (a11y)
  • 66d2e07 Withings error string break-words (a11y)
  • 79bb167 MockAIProvider DEFAULT_RESPONSE strict schema
  • d947563 prettier sweep on touched files

Backlog written: .planning/v1416-backlog.md — 8 deferred HIGH + 39 MED/LOW + 4 simplify-no + 3 process items. Most impactful pre-v1.4.16 fast-track: M4 dispatcher legacy Telegram migration (silently unwinds B3 hard-rejects) and M6 senior-dev AI strict-schema route migration.

Phase E — Release v1.4.15

  • Pre-release verify: typecheck, lint, format, test, integration, build (Node-22 CI), e2e
  • Bump package.json to 1.4.15
  • CHANGELOG.md entry
  • Tag + push v1.4.15
  • GHCR build green (BOTH main + tag — verify via the new C2 reliability work)
  • Coolify auto-deploy via webhook (the C2 work tests itself)
  • /api/version=1.4.15 confirmed
  • Image digest changed
  • Production smoke screenshots
  • GH release created
  • Docs site + landing-site sync
  • docs.healthlog.dev: new pages "Backup structure", "User-deletion lifecycle"
  • Final summary docs/audit/v1415-summary.md (Marc-Brief)
  • Detailed report: .planning/phase-E-report.md

Previous milestone — v1.4.14 (completed 2026-05-09T18:35+02:00)

LIVE at https://healthlog.bombeck.io · /api/version=1.4.14 · image digest sha256:0ced46004a54… (was 791c2cd2… on v1.4.12).

Full Marc-Brief, commit table, deferred items, CI/prod state and Codex-OAuth resolution: docs/audit/v1414-summary.md.

Phases run during v1.4.14 marathon:

  • Phase 0 — Bootstrap (STATE+ROADMAP for v1.5/v1.4.14, git+CI sanity)
  • Phase 1 — Verify Codex-OAuth flow end-to-end (gpt-5.3-codex slug landed in commit 5df74f7)
  • Phase 2 — v1.4.6 deferred backlog (T2.1, T2.2, T2.3, T2.5; T2.4 + T2.6 + T2.7 deferred to phase 4b)
  • Phase 3 — End-to-end test coverage (41/41 specs green)
  • Phase 4 — Performance audit (/insights −108 KiB initial JS, dashboard −950 ms checklist)
  • Phase 4b — Admin Panel refactor (per-section dynamic routes, T2.6 + T2.7 folded in)
  • Phase 5 — UX polish (a11y violations cleared, trend-arrow direction-as-good, saved-AI-key removal)
  • Phase 6 — Multi-agent QA (code-review, security, design, simplify; CRITICAL/HIGH applied inline; HIGH+MEDIUM deferred items recorded in .planning/v1415-backlog.md)
  • Phase 7 — Pre-release verification (typecheck/lint/format/test/integration green; build/e2e Node-25 local issue, Node-22 CI canonical)
  • Phase 8 — Release v1.4.14 (rebrand patch, GHCR :1.4.14, host-side retag deploy, image digest 0ced46004a54…)
  • Phase 9 — Docs + landing sync (commits 06bc616 docs, 92c6588 landing)
  • Phase 10 — Backfill GitHub releases v1.4.7..v1.4.13
  • Phase 11 — Final summary docs/audit/v1414-summary.md

Status block — Phase 0 (v1.4.15)

  • 2026-05-09T20:06:00+02:00 — Phase 0 complete. STATE.md + ROADMAP.md scaffolded for v1.4.15 marathon (phases A1-A5, B1-B6, C1-C5, D, E). Previous v1.4.14 entries archived above. v1.4.14 leftover phase reports (10 untracked files) deliberately not folded into this Phase 0 commit (they belong to the previous milestone). Phase 0 commit will contain only .planning/STATE.md, .planning/ROADMAP.md, .planning/phase-0-report.md. Marc-status: awake, watching — speed matters.