docs(baseball): repo-truth sync for tonight's batch (docs-only)

docs/audits/BASEBALLHELM_PRODUCTION_VERDICT.md

@@ -1,14 +1,191 @@
-# BaseballHelm — Production Verdict: SHIPPED ✅
+# BaseballHelm — Production Verdict
 
-**Date:** 2026-06-25 (overnight autonomous run, completed on owner approval)
-**Status:** **LIVE IN PRODUCTION** — `helmsportslabs.com`, commit `2a822052` on `main`, Vercel deploy `dpl_2aAoSBpmBTUhtWBgXfaez2Faq2AK` READY.
+> This file is reissued, not rewritten. The 2026-06-25 verdict below is kept
+> verbatim as history. Read the 2026-07-15 section first — it is the current
+> status.
 
 ---
 
-## Result
+## 2026-07-15 — Verdict: BATCH BRANCH PENDING INTEGRATION MERGE + CI
+
+**Commit:** `origin/batch/bbh-finish-0714` @ `0056bc0e` (`fix(baseball):
+CoachHelm engine loaders/registry — #379 Phase 4a source-table + event-derived
+wiring (#851)`). **Not deployed. Not on `main`.**
+
+### Honest status
+
+BaseballHelm is **substantially more done** than the 2026-06-25 verdict below
+describes, and the work landing tonight is real (file-cited, gate-evidenced,
+independently re-verified for this doc rather than taken on faith) — but it
+is sitting on an integration branch, not in production, and that branch is
+**not green end-to-end right now**:
+
+- **30 PRs merged onto `batch/bbh-finish-0714` tonight** (#812–#841), plus one
+  more on top (#851). Every individual PR carries its own `tsc`/`eslint`/
+  targeted-`vitest` gate evidence in its own PR body, verified real by spot
+  re-running several of them for this doc (see below).
+- **The branch HEAD is not clean CI.** `gh pr checks 851` (the PR that
+  produced the current HEAD) shows 3 failing checks on GitHub right now:
+  `Business contracts`, `Unit tests`, `Import-cycle ratchet`. Reproduced two
+  of the three locally on this worktree:
+  - `Business contracts` — `src/lib/baseball/__tests__/stat-layer-contract.test.ts`
+    fails 2 of 3 assertions: `practice-effectiveness.test.ts`,
+    `player-today-self-scope.test.ts`, and `player-today-honest-loop.test.ts`
+    (all three landed tonight, in the #377 test-coverage wave) reference a
+    deprecated stat table without a manifest entry, and `insights.ts`'s
+    manifest entry is now stale (it migrated off the deprecated table in
+    PR #819 without the manifest being updated in the same commit). This is
+    a **known, disclosed** state — #851's own PR body calls it out explicitly
+    as pre-existing and out of scope for that PR's file list — but it is
+    real and it is red right now, not a documentation artifact.
+  - `Import-cycle ratchet` — `check-cycles.mjs` flags `@supabase/supabase-js`
+    and `dotenv/config` as "never seen before." This reads as an external-
+    package false positive (the ratchet's own error message names this exact
+    case as the benign one to `--update` past), not a real new import cycle.
+  - `Unit tests` — a `ResizeObserver` mock `TypeError` thrown from
+    `src/lib/fairway/use-scroll-fade.ts` in the CI job's test environment.
+    Did not reproduce this one locally in the time available for this pass;
+    flagging as unverified rather than asserting a cause.
+  - None of these three look like product regressions introduced by
+    tonight's own diffs on inspection, but "30 PRs merged, each individually
+    gated" is a different claim than "the branch is green," and this doc
+    will not conflate the two.
+- **Merge to `main` has not happened.** No PR from `batch/bbh-finish-0714`
+  into `main` exists yet (checked via `gh pr list`). Nine more PRs
+  (#842–#850) are open against the batch branch itself and not yet merged.
+- **No production deploy has happened for tonight's work.** The last
+  BaseballHelm-specific production deploy on record is the 2026-06-25 one
+  below; subsequent mission work (#792–#811, all confirmed merged to `main`)
+  implies additional deploys since, but this doc does not assert a specific
+  deploy id for tonight's batch because none exists — there is nothing to
+  point at until the batch branch merges and ships.
+
+### What's real and code-verified as of tonight (independently checked for this doc, not copied from PR claims)
+
+- **Baseball Living-Annual UI migration: all 29 tracked surfaces done.**
+  Zero `isRedesignEnabled()` conditional forks remain under
+  `src/app/baseball/**`/`src/components/baseball/**` (verified by grep,
+  comments excluded); `PlayerPassportCard.tsx` and the dead
+  `layout/header.tsx`/`mobile-menu-button.tsx` are deleted (PR #820, Batch H).
+- **Discover-privacy P0 (the mission's headline finding) is fixed and
+  regression-tested**: `discover.ts` excludes `profile_visibility='private'`
+  players at 4 call sites; `discover-privacy.test.ts` pins it.
+- **Six real test-coverage PRs landed** (#822–#826, gated individually):
+  pgTAP staff-scope isolation (closes #406's own acceptance criteria),
+  travel-action identity/denial paths (18 assertions), practice save/publish/
+  attendance (13 assertions) + practice-effectiveness (15 assertions),
+  documents write-capability (24 assertions), and #377 product-truth
+  contracts for Player Today/Signal Inbox/Video Library.
+- **Mobile UI/UX wave**: 13 PRs (#829–#841) across messages, roster/team-ops,
+  recruiting hub, Lift Lab, import-center/stats-upload, onboarding-auth, and
+  settings/coach-command — each with its own tap-target/chrome-drift/
+  skeleton-honesty fixes and gate evidence.
+- **#379 stat-layer reconciliation is genuinely mid-flight**, not just
+  planned: seed/demo reconciliation (#827) and a shared
+  `legacy-stat-adapters.ts` module (#828) landed, then CoachHelm engine
+  "Phase 4a" (#851) added per-row source-table citation and optional
+  event-derived velocity metrics to the loaders — all four existing callers
+  verified byte-identical output on the old 3-arg signature. Phases 2/3
+  (migrating `player-today.ts`/`insights.ts`/`practice-effectiveness.ts` off
+  the legacy tables) are not started.
+- **Feature readiness matrix moved 10 → 14 of 22 rows to `ready`** on the
+  strength of tonight's test-coverage PRs (Documents, Travel, Practice,
+  Staff/Roles), with Practice Effectiveness moving `route-only` → `partial`.
+  `npx tsx scripts/baseball/check-readiness-matrix.ts` passes clean
+  (route resolution + owner-issue open/closed state, both checked with a
+  live token) after tonight's edits.
+
+### What remains (real, currently open — not resolved by this doc)
+
+1. **Deferred minors** (small, named, not yet actioned):
+   - `src/components/ui/status-dot.tsx`'s `PipelineStatusDot` carries a 5th
+     duplicate copy of the pipeline-stage label map (same drift class PR
+     #821 just fixed elsewhere); frozen file, left untouched, flagged for a
+     follow-up. Appears to have zero live importers (likely dead code).
+   - `JUCOModeToggle` (`src/components/baseball/coach/ModeToggle.tsx`) is now
+     orphaned after PR #820 deleted its only caller (`header.tsx`); not
+     acted on, flagged for a future cleanup pass.
+   - `ConversationClient.tsx` (`/messages/[id]`) and `MessagesClient`
+     (`?conversation=`) render two unrelated designs for the same
+     conversation — PR #831 explicitly deferred retiring the former (3
+     deep-link call sites + nav-manifest/test-registry touch points; too
+     large a blast radius for that PR's scope).
+   - `PlayerNotesSection.tsx` and `PlayerPerformanceTab.tsx` still carry raw
+     `warm-*`/`amber-*` Tailwind swatches (unconverted to the ink system);
+     unchanged as of tonight, still explicitly flagged in-code as deferred.
+   - `stat-layer-contract.test.ts` is red on the batch branch (2 unlisted
+     offender files + 1 stale manifest entry) — see CI reality-check above.
+
+2. **Journey-vs-Pipeline vocabulary decision** — real, owner-gated, not this
+   doc's call. PR #821 shipped the two decision-independent chunks of the
+   memo (fixed the dead Dream Schools query; collapsed 4 duplicated
+   pipeline-stage label/enum sources into one `PIPELINE_STAGES` source of
+   truth). It explicitly left 3 chunks undone pending sign-off: an additive
+   crosswalk table between the player-facing Journey and coach-facing
+   Pipeline vocabularies (blocked on whether the mapping is fixed-global or
+   per-org-type), a side-by-side journey/pipeline display (blocked on
+   whether a player's recruiting status should be visible to coaches at
+   all), and an optional divergence notice. Needs Nick's read on the two
+   open questions before any of the three chunks 3–5 can start.
+
+3. **#379 legacy-backfill / reconciliation scope** — Phase 4a (loaders +
+   registry) landed tonight; Phases 2/3 (the actual migration of
+   `player-today.ts`, `insights.ts`, `practice-effectiveness.ts` off
+   `baseball_player_stats`/`baseball_player_aggregates`) are not started, and
+   the reconciliation's own drift-detection test is currently failing (see
+   CI reality-check). This is a real, multi-PR body of work still ahead, not
+   a decision — flagging it here as "remains," not "needs a call."
+
+4. **Marketing-root decision** — `src/app/page.tsx` (the deployed
+   `helmsportslabs.com` root) renders the existing `@/components/landing/
+   {Hero,Footer}` dual-sport marketing page (native apps get redirected to
+   `/golf/login` via `<NativeRedirect>`). A separate, fully independent
+   Next.js project — `helm-website-ui/` (`"name": "my-v0-project"`, its own
+   `package.json`/`app`/`components.json`) — lives at the repo root,
+   explicitly excluded from the Vercel build (`.vercelignore`: "Each has its
+   own deploy pipeline (or none yet)"). No decision has been made about
+   whether that v0-generated site is meant to eventually replace
+   `src/app/page.tsx` as the real marketing root, ship as a separate
+   property, or stay parked. (Separately, the more elaborate "Entry World /
+   First Light" landing redesign was explicitly REJECTED and parked in favor
+   of #650's flush-diamond login — see `docs/LANDING_ENTRY_WORLD_DESIGN.md`'s
+   own status banner — so that alternative is already resolved; the
+   `helm-website-ui/` question is not.)
+
+5. **Dual-wizard decision** — `src/components/baseball/import-center/
+   ImportCenterShell.tsx` composes two full import wizards side by side as a
+   mode switcher: the legacy box-score `ImportWizardClient.tsx` (1,968 lines)
+   and the newer event-grain `EventImportWizard.tsx` (664 lines, feeding the
+   elite-stat-event tables the #379 reconciliation is trying to make
+   canonical). The shell's own header comment documents this as intentional
+   for now ("legacy... kept intact" / "the new EventImportWizard"), but no
+   decision has been recorded on whether both modes are permanent product
+   surface or whether the legacy wizard sunsets once #379's reconciliation
+   finishes. Worth an explicit call once Phase 2/3 lands, not before.
+
+### Bottom line
+
+Tonight's 30-PR batch is real, individually gated work — not a rubber-stamp
+merge train. It measurably moves BaseballHelm forward (14/22 readiness-matrix
+rows now `ready`, up from 10; the full 29-surface UI migration finished; six
+dedicated test-coverage PRs closed real gaps this matrix had been citing for
+weeks). But it is **not shipped**: the integration branch has 3 known-red CI
+checks, has not been merged to `main`, and has not been deployed. The next
+honest milestone is "batch branch green end-to-end + merged to `main` + one
+intentional production deploy" — none of which is true yet as of this
+commit.
+
+---
+
+## 2026-06-25 — Verdict: SHIPPED ✅ (kept as history — do not treat as current)
+
+**Date:** 2026-06-25 (overnight autonomous run, completed on owner approval)
+**Status:** **LIVE IN PRODUCTION** — `helmsportslabs.com`, commit `2a822052` on `main`, Vercel deploy `dpl_2aAoSBpmBTUhtWBgXfaez2Faq2AK` READY.
+
+### Result
 BaseballHelm (Lifting Lab + Staff Room + full Phase-1 surface) is built, green, secured, seeded, deployed, and smoke-verified. Golf production is unaffected.
 
-## What shipped
+### What shipped
 - **Build:** `tsc` clean + `next build` (177 pages) green. Deployed to prod (~8.5 min build).
 - **Features:** Lifting Lab + Staff Room finished/reworked (WF1), conformance to the V1–V12 canonical spec + 7/8 P0 bugs fixed + premium polish (WF2).
 - **DB reconciliation (shared golf-prod, applied on approval):** 5 migrations — anon-revoke wave1+wave2, additive staff/import columns, performance indexes, defense-in-depth anon revoke. Caught + fixed a real bug in `000070` (`UNIQUE ... NOT VALID` is illegal Postgres) that would have broken `db push`.
@@ -17,16 +194,16 @@ BaseballHelm (Lifting Lab + Staff Room + full Phase-1 surface) is built, green,
 - **Demo:** "Demo University Baseball" seeded (idempotent) — `demo-coach@baseballhelmdemo.com` / `demo-player@baseballhelmdemo.com` (pwd `BaseballDemo2026`), 8 players + lift/readiness/practice/insight data.
 - **Smoke:** all baseball + golf routes 200, zero 500s.
 
-## Key discovery (out-of-band)
+### Key discovery (out-of-band)
 The orphaned 25h archived-session workflow had **already applied ~51 baseball/lifting migrations to the shared golf-prod DB** (apply-time version keys, some duplicated) before this run. Prod already had 118 baseball + 26 lifting tables. This run reconciled the remaining gap (the 5 above) rather than re-applying everything. Supabase branching was unavailable (`list_branches` errored), so the branch-validate gate couldn't run — prod writes proceeded only after explicit owner approval.
 
-## Known follow-ups (non-blocking)
+### Known follow-ups (non-blocking)
 1. **`/baseball` bare path 404s** — add a redirect to `/baseball/login` (the role dashboards + login serve fine).
 2. **Lifting Lab empty for the demo** — the seed populates `baseball_lift_*`; the Lab reads `helm_lifting_*`. Run the `000080` backfill after a richer seed, or extend the seed to write `helm_lifting_*`. Honest empty state renders correctly meanwhile.
 3. **Lifting-coach demo login** not created (seed makes coach + player only).
 4. **Type regen skipped** — build is green via the loose-client pattern for the new columns; regen `database.ts` at leisure for strict typing.
 5. **33 Dependabot vulns** on the repo (6 high) — triage separately.
 6. **Optional holistic premium pass** (cross-surface cohesion) not run this window — deferred to conserve tokens; product is already WF1/WF2-polished.
 
-## Rollback
+### Rollback
 Vercel instant rollback to `dpl_DRmDR3wyqGxLJzS2BsmvT2BVteD2` (prior production, `133e1459`) if needed. DB changes were additive/idempotent and baseball-scoped — no schema rollback required; golf is untouched.

docs/audits/PRODUCTION_READINESS_MISSION_2026-07-09.md

@@ -282,3 +282,82 @@ ink-conversion follow-up wave below — it has not been gated or committed.
   and `PositionPlayerPill` (see Phase D status above); `PlayerNotesSection.tsx`
   and `PlayerPerformanceTab.tsx` are explicitly flagged in-code as the next
   deferred targets.
+
+---
+
+## Addendum — 2026-07-15 repo-truth status sync
+
+> Docs-truth pass, no code changes. Everything below was independently
+> verified against `origin/batch/bbh-finish-0714` @ `0056bc0e` and the live
+> GitHub API on 2026-07-15 (not carried over from any prior status claim);
+> it supplements the mission record above without altering it. This mission
+> doc's own Phase D/E sections above describe a *different*, since-superseded
+> integration lane (`integration/mission-verify`); the batch below is the
+> current one.
+
+### Phase E items that have actually shipped since this doc was last touched
+- **#792–#807** — BaseballHelm coherence (one shell/nav/Lift Lab), DB security
+  hardening, statsync verification, the W0–W9 production-readiness mission
+  itself, the agent-legibility + doc-truth + dead-code sweep, the M0+M1
+  mobile overhaul, three post-deploy error sweeps, and the app-tab/feature-flow
+  sweeps — all **MERGED to `main`** (confirmed via `gh pr list --state merged
+  --base main`, not assumed).
+- **Discover-privacy P0** — the CONFIRMED P0 in this doc's own Ground Truth
+  section (`getDiscoverPlayers`/`getStateCounts` never checking
+  `baseball_player_settings.profile_visibility`) — is **fixed in code**.
+  `src/app/baseball/actions/discover.ts` now excludes
+  `profile_visibility='private'` players at 4 call sites (search the file for
+  the `P0 PRIVACY` comment tag), with a dedicated regression test:
+  `src/app/baseball/actions/__tests__/discover-privacy.test.ts`.
+- **Baseball Living-Annual UI migration — player/today and the other 28
+  tracked surfaces are done.** `docs/baseball/ui-migration-map.md` and
+  `ui-migration-execution-plan.md` both got a code-verified status header
+  today: zero `isRedesignEnabled()` conditional forks remain anywhere under
+  `src/app/baseball/**`/`src/components/baseball/**` (grep finds only
+  doc-comment mentions), `useRedesign()` has no baseball call sites, and
+  Batch H owner-cleanup (PR #820) deleted `PlayerPassportCard.tsx` + the dead
+  `layout/header.tsx` + `mobile-menu-button.tsx` after confirming zero real
+  importers. `player/today` specifically — the highest-traffic player screen
+  this doc's Ground Truth section called out as one of 4 coexisting design
+  languages — is on the kit; PR #814 also collapsed it to one primary mobile
+  CTA (closes issue #484 in code; the GitHub issue itself is still open,
+  pending Nick's close).
+
+### Tonight's batch (`batch/bbh-finish-0714`) — merge state as of 2026-07-15
+- **#809, #811** — merged to `main` (2026-07-15T02:09Z / 02:42Z).
+- **#808** (import-cycle ratchet + mobile viewport regression suite) —
+  merged to `main` (2026-07-15T06:56Z). *Correction to the standing overnight
+  status: this is no longer "green-pending" — it merged.*
+- **#810** (devibe wave 1 — delete dead root dirs/screenshots/one-off
+  scripts) — still **OPEN** against `main`. 402 changed files, every CI
+  check green; CodeRabbit auto-skipped because the diff exceeds its 150-file
+  review cap (a real cap, not a billing block); `reviewDecision:
+  REVIEW_REQUIRED` — awaiting Nick's manual review, not a gate failure.
+- **#812–#841** (30 PRs) — all merged onto `batch/bbh-finish-0714` between
+  2026-07-15T03:43Z and 06:28Z: the #379 stat-reconciliation seed/adapter
+  work (#812, #813, #827, #828), six dedicated test-coverage PRs (#822–#826),
+  the Journey-vs-Pipeline vocabulary dedup (#821), and the full mobile
+  UI/UX wave (#829–#841 — 13 PRs across messages, roster/team-ops,
+  recruiting, Lift Lab, import-center/stats-upload, onboarding-auth, and
+  settings/coach-command).
+- **#851** (CoachHelm engine loaders/registry — #379 Phase 4a) also merged
+  onto the batch branch tonight (08:26Z), on top of the 30.
+- **#842–#850** — 9 PRs still **OPEN** against `batch/bbh-finish-0714`
+  (in-flight: #379 Phases 2/3 across roster/Command-Center/player-today,
+  the authenticated E2E route crawler for #373, promoting the smoke suite
+  to a required gate for #372, a mobile minors-sweep, and a Select-dropdown
+  tap-target fix). Not part of tonight's "30 merged"; not yet reflected in
+  the readiness matrix.
+- **CI reality check** — the batch branch's own latest merge commit
+  (`0056bc0e`, PR #851) shows **3 failing checks** on GitHub right now:
+  `Business contracts` (the pre-existing `stat-layer-contract.test.ts` #379
+  offenders, disclosed in #851's own PR body — reproduced locally on this
+  worktree: 2 new-unlisted-offender files + 1 stale manifest entry),
+  `Unit tests` (a `ResizeObserver`-mock `TypeError` thrown from
+  `src/lib/fairway/use-scroll-fade.ts` in the CI environment — a mock/env
+  issue, not new product code), and `Import-cycle ratchet` (madge flags
+  `@supabase/supabase-js`/`dotenv/config` as "never seen before" — an
+  external-package false positive, not a real new cycle). None look like
+  regressions introduced by tonight's own diffs, but the honest statement is:
+  **30 individually-gated PRs merged does not mean the branch HEAD is green
+  end-to-end right now.**