Wire the repo guards that executed never, and the ratchet nothing ran by njrini99-code · Pull Request #1694 · njrini99-code/helmv3

njrini99-code · 2026-08-31T03:37:39Z

An audit of one auditable question: does every check that exists actually run,
and does every ratchet actually ratchet? Full method and per-guard verdicts in
docs/reports/REPO_WIRING_AUDIT_2026-08-30.md.

Two things were not running.

19 guard tests executed never

vitest.config.ts already documented the trap and ships a self-check — 51 files
in scripts/__tests__, 31 listed. The unlisted ones all import node:test, and
nothing in this repo runs node --test. Promoting one is a port, not a
config line.

Run under node --test, all 19 fail. What matters is why, and it splits
cleanly on whether the files each guard names still exist. Every missing path was
traced to a dated deletion commit — mostly ffd0fd8ab (W1 Fairway
consolidation) and a259fa296 (dead player-CoachHelm cluster).


promoted	`admin-tables-mobile`, `no-arbitrary-text-px-fairway-pages` — stale targets dropped, both green. `guards`: 11 files/32 tests → 13/37
deleted	`chart-tooltip-consolidated`, `genome-fluid`, `microcopy-banned-phrases` — subjects deleted or rewritten
repaired, still red	`badge-consolidation` — failure is now honest: two live CRM badges genuinely stopped delegating to `<Badge>`
recorded, not fixed	10 guards with real design drift, counts in the report

Deleted rather than repaired, deliberately: making a retired rule pass is
implementing retired policy. genome-fluid is the clearest case — its subject
was rewritten onto recharts, so it asserts a fixed viewBox and
aspect-square on a hand-rolled SVG that no longer exists.

One path was a move, not a deletion (GenomeRadar → fairway/charts/).
Worth separating: dropping a moved file silently loses coverage.

Every one of the ten drift guards was checked against
.claude/rules/design-system.md and encodes still-live policy, so none was
deleted. Those are design debt a working guard would have caught — product work,
not audit work.

`lint:duplicate-exports` never ran

It has a committed baseline (27 grandfathered) and exits 1 on anything new, and
no job invoked it. A ratchet nothing compares against is a green that means
nothing. Now a fifth step in the Lint ratchet job, following the existing
continue-on-error + aggregate pattern.

Six other "unwired" scripts were false positives — CI invokes them as
node scripts/… inside multi-line run: blocks. Recorded as method, because
the naive check is the obvious one to repeat.

Bottleneck — measured, not where it was feared

The config warns these sweeps walk ~4,066 files each and carries a 120s timeout.
The whole guards project runs in 1.48s; each promoted guard costs 0.1s.
No argument for promoting in waves.

One finding that is not about wiring

Two ChartTooltip components, one exported name.
fairway/charts/ChartTooltip.tsx (220 lines) has five real importers.
ui/chart-tooltip.tsx (108 lines) has zero — but is re-exported as
ChartTooltip from @/components/ui, so importing from that barrel silently
gets the dead one. This is exactly what the deleted
chart-tooltip-consolidated guard existed to prevent, and it happened while
that guard was not running. Not touched here: product code, and Knip's weekly
job is the right owner.

Verified

typecheck 0 · lint 0 · npm test 1275 files / 12,209 passed · guards 37/37 ·
duplicate-exports ratchet 0 · docs:schema-drift 0 · docs:path-drift 0 ·
knowledge:check 0 · inventory-check 0 · markdown ratchet at baseline ·
ci.yml parses as valid YAML

🤖 Generated with Claude Code

https://claude.ai/code/session_01NbDxyygyXRUEERuGocpZZH

…t nothing ran An audit of whether every check in this repo actually runs. Full method and per-guard verdicts in docs/reports/REPO_WIRING_AUDIT_2026-08-30.md. Two things were not running. **19 guard tests executed never.** vitest.config.ts already documented the trap and ships a self-check: 51 files in scripts/__tests__, 31 listed. The unlisted ones all import `node:test`, and nothing in this repo runs `node --test` — not a script, not a workflow. Promoting one is a port, not a config line. Run under node --test, all 19 fail. What matters is why, and it splits cleanly on whether the files each guard names still exist. Six fail because every path they reference was DELETED — traced to dated commits, mostly ffd0fd8 (the W1 Fairway consolidation) and a259fa2 (the dead player-CoachHelm cluster). Ten fail on real violations in files that exist. promoted admin-tables-mobile, no-arbitrary-text-px-fairway-pages — stale targets dropped, both now green. guards: 11 files/32 tests -> 13/37. deleted chart-tooltip-consolidated (all four subjects gone — a guard with an empty subject list is not a guard), genome-fluid (its subject was REWRITTEN onto recharts; it asserts a fixed viewBox and aspect-square on a hand-rolled SVG that no longer exists), and microcopy-banned-phrases (five of seven subjects rebuilt in #1058 and #984). Deleted rather than repaired: making a retired rule pass is implementing retired policy. repaired, badge-consolidation — six dead entries and one stale allowlist still red entry removed so its failure is now HONEST: two live CRM badges genuinely stopped delegating to <Badge>. Left unlisted. That tree is NOT dead — next.config.mjs:183 redirects /golf/admin on the exact path only, and 144 files live under it — so this is a colour-fidelity refactor, which is product work, not audit work. One path was a MOVE, not a deletion (GenomeRadar -> fairway/charts/). Worth separating: dropping a moved file silently loses coverage. **lint:duplicate-exports never ran.** It has a committed baseline (.duplicate-exports-baseline.json, 27 grandfathered) and exits 1 on anything new, and no job invoked it. A ratchet nothing compares against is a green that means nothing. Wired as a fifth step in the Lint ratchet job, following the existing continue-on-error + aggregate pattern so it reports its own outcome. Six other "unwired" scripts were false positives: CI invokes them as `node scripts/…` inside multi-line run: blocks, not `npm run <name>`. Recorded in the report as method, because the naive check is the obvious one to repeat. Bottleneck, measured rather than feared: the config warns these sweeps walk ~4,066 files each and carries a 120s timeout. The whole guards project runs in 1.48s and each promoted guard costs 0.1s. No argument for promoting in waves. Verified: typecheck 0, lint 0, npm test 1275 files / 12,209 passed, guards 37/37, duplicate-exports ratchet 0, docs:schema-drift 0, docs:path-drift 0, knowledge:check 0, inventory-check 0, markdown ratchet at baseline, ci.yml parses as valid YAML. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NbDxyygyXRUEERuGocpZZH

qodo-code-review · 2026-08-31T03:37:44Z

ⓘ Qodo reviews are paused because your trial has ended. Ask your workspace admin to add credits to resume reviews. Manage billing

vercel · 2026-08-31T03:37:44Z

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated (UTC)
helmv3	Ignored	Preview	Aug 31, 2026 3:38am

supabase · 2026-08-31T03:37:45Z

This pull request has been ignored for the connected project qmnssrrolpinvwjjnufo because there are no changes detected in supabase directory. You can change this behaviour in Project Integrations Settings ↗︎.

Preview Branches by Supabase.
Learn more about Supabase Branching ↗︎.

coderabbitai · 2026-08-31T03:37:48Z

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: 8d44735f-27b1-4625-bab9-e60209ffe890

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

🔍 Trigger review

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

_{Comment @coderabbitai help to get the list of available commands.}

…#1694 #1693 was transitionally closed at its own merge; the grace ended when this branch added a commit past it. Cleared in the next ordinary PR, which registers itself at the same time. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NbDxyygyXRUEERuGocpZZH

…#1695 Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NbDxyygyXRUEERuGocpZZH

* fix(lifecycle): stop inventing "no PR" for the canonical checkout's own branch The PR lookup was skipped for the canonical checkout and substituted with a literal: const pr = w.branch && !isCanonical ? prFor(w.branch) : { lookup: 'OK', state: 'NONE' }; `{ lookup: 'OK', state: 'NONE' }` does not mean "not checked". It means "checked, and there is no PR" — and classifyBranch read it exactly as written, returning UNKNOWN_PR ("no PR found — cannot prove the work landed") for canonical's branch even with an OPEN PR against it. Reproduced against #1694 on 2026-08-30: `gh` answered `1694 OPEN` from the shell while the tool's own report printed PR `NONE`. Nothing was lost — UNKNOWN_PR fails safe to KEEP — but the tool was reaching a right answer from an invented fact, in the evidence layer, which is the exact failure this repo keeps removing everywhere else. The shortcut also bought nothing. classifyWorktree returns ACTIVE for canonical on its FIRST line, before any PR fact is read. Only the BRANCH verdict ever consumed these — and that is the separation AGENTS.md insists on: may this CHECKOUT go, versus may this BRANCH be deleted. The short-circuit conflated the two inside the tool that defines the distinction. The test took two attempts, and the first one is worth recording because it looked fine. It passed with the bug reverted — because canonicalRoot() resolves .claude/hooks/lib/workspace-identity.mjs against the repo under test, fixtures do not have one, so isCanonical is ALWAYS false in the harness and the bug is unreachable there. That null is deliberate and correct (a fixture must never be able to claim it is the real canonical checkout), but it means every canonical-only branch of this tool was untestable. The fixture now writes its own identity stub, and the test asserts ACTIVE first — proving it really is being treated as canonical — before asserting the branch verdict. Verified by injection: reverting the one-line fix turns the new test red and everything else green; restoring it returns the file byte-identical by sha256. typecheck 0, lint 0, npm test 1275 files / 12,210 passed, guards 37/37. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NbDxyygyXRUEERuGocpZZH * chore(control-plane): clear #1694's expired transitional row, register #1695 Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NbDxyygyXRUEERuGocpZZH --------- Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>

njrini99-code merged commit 66a0ae0 into main Aug 31, 2026
48 checks passed

njrini99-code deleted the fix/wire-dead-repo-guards branch August 31, 2026 03:48

njrini99-code mentioned this pull request Aug 31, 2026

Stop inventing "no PR" for the canonical checkout's own branch #1695

Merged

njrini99-code mentioned this pull request Aug 31, 2026

Audit addendum: the branch-durability blind spot, and a zsh trap that reads as a git bug #1696

Merged

Navigation Menu

Wire the repo guards that executed never, and the ratchet nothing ran - #1694

Wire the repo guards that executed never, and the ratchet nothing ran#1694
njrini99-code merged 2 commits into
mainfrom
fix/wire-dead-repo-guards

njrini99-code commented Aug 31, 2026

Uh oh!

qodo-code-review Bot commented Aug 31, 2026

Uh oh!

vercel Bot commented Aug 31, 2026 •

edited

Loading

Uh oh!

supabase Bot commented Aug 31, 2026

Uh oh!

coderabbitai Bot commented Aug 31, 2026 •

edited

Loading

Review skipped

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

njrini99-code commented Aug 31, 2026

19 guard tests executed never

lint:duplicate-exports never ran

Bottleneck — measured, not where it was feared

One finding that is not about wiring

Verified

Uh oh!

qodo-code-review Bot commented Aug 31, 2026

Uh oh!

vercel Bot commented Aug 31, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

supabase Bot commented Aug 31, 2026

Uh oh!

coderabbitai Bot commented Aug 31, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Review skipped

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

`lint:duplicate-exports` never ran

vercel Bot commented Aug 31, 2026 •

edited

Loading

coderabbitai Bot commented Aug 31, 2026 •

edited

Loading