Retire the dead --text-2xl-compact type step

[BigSimmo]

Summary

• Delete --text-2xl-compact from the src/app/globals.css @theme block. It was declared but selected by zero production surfaces — the one step the declared-but-unconsumed gate found on the day it landed, carried since as the only entry in UNUSED_TYPE_STEP_EXEMPTIONS with its retirement tracked as outstanding-issues #297. Deleting a step nothing selects renders identically; it gets its own revertible PR only because it edits @theme.
• Drop 2xl-compact from CLINICAL_TWMERGE_THEME.text in src/lib/tailwind-merge.ts and from that file's docstring. The paired test "declares exactly the --text-* tokens globals.css defines" keeps the config and the @theme block honest in both directions, so the two edits have to move together.
• Drop text-2xl-compact from CUSTOM_TEXT_STEPS in tests/tailwind-merge-config.test.ts.
• Delete the exemption entry in scripts/check-design-system-contract.mjs. The Map is now empty, and its header records that as the intended resting state rather than a list awaiting entries — an addition needs a ledger row saying when it leaves again. The gate itself is unchanged and still fails, by name, on any future declared-but-unconsumed step.
• Repoint the one synthetic var(--text-2xl-compact) fixture in tests/design-system-contract-utils.test.ts at --text-2xl-minus, so no fixture names a token that no longer exists. That fixture exercises the extraction utility on arbitrary strings; it was never a real consumer, which is why the step measured as unused.
• Docs: the docs/design-system/TOKENS.md legacy-type-steps row no longer lists the step, and docs/design-system/GATES.md now reads eight non-standard steps rather than nine. The 705-consumer total there is unchanged, precisely because this step contributed 0.
• Close #297 in docs/outstanding-issues.md via npm run issues:done.

docs/site-map.md and .design-sync/design-system-adoption.json are regenerated by the pre-commit hook, not hand-edited.

Verification

• npm run verify:pr-local — completed check:runtime, check:installed-lock-parity, format:changed, sitemap:check, docs:check-index, docs:check-inventory, docs:check-scripts, docs:check-links, check:branch-review-ledger, check:outstanding-issues, lint, typecheck. It then stopped at test on a pre-existing environment failure unrelated to this diff, so build and check:rag:fixtures were not reached by the chain; both were then run directly, see below.

The failure is tests/pr-handoff-stop.test.ts → "emits handoff context only when the marker file exists". The test chmods a directory to 0555 to force a marker write to fail; this container runs as root, which ignores the mode bits, so the write succeeds and the assertion markerExists === false is false. It reproduces unchanged on a clean tree at d812c76 (verified by stashing this diff and re-running the file), so it is a root-user environment artefact, not a regression here. Full suite otherwise: Test Files 1 failed | 552 passed (553), Tests 1 failed | 5993 passed | 4 skipped (5998).

Gates run directly for this change:

• npm run check:design-system-contract — passed: "Design-system contract passed (705 production files; raw colors 2; literal shadows 0; legacy tap classes 0; edge conflicts 27; 1px shadow spreads 2)", plus adoption (54 components, 59 roots) and design-sync (54 components, 7 guidelines). This is the gate that owns the retired exemption.
• npm run check:type-scale (--strict) — "✓ type-scale: no arbitrary text-[px|rem|em] font sizes in src."
• npm run lint — exit 0. npm run typecheck — exit 0.
• npm run build after rm -rf .next (no cache reuse) — exit 0, "✓ Compiled successfully in 63s", "Client bundle secret surface check passed." This is the check that matters for a @theme edit: the stylesheet still compiles with the step gone.
• npm run check:outstanding-issues — "Outstanding-issues guard passed: 299 rows (146 open, 153 archived), unique ids, next-id=302 above the highest".
• npm run format was run whole-tree and committed; Prettier reflowed the TOKENS.md table columns as a consequence of the shortened cell.

UI verification not run: no rendered output can change. --text-2xl-compact had zero class-utility and zero var(--text-*) consumers across the contract check's own walk of production sources — that measured absence is the entire premise of the change, and it is re-asserted mechanically by the gate above. Spending a Chromium journey run to confirm that deleting an unreferenced custom property changed nothing would not test a plausible changed failure path.

Risk and rollout

• Risk: Low. The only way this bites is if a consumer existed that the gate's walk cannot see — a dynamically composed class name, or a surface outside the walked production sources. The walk covers both class utilities and direct var(--text-*) references, and check:type-scale independently forbids the arbitrary-value escape hatch (text-[1.45rem]) that would otherwise reproduce the step's 1.45rem by hand. Mockups are excluded from the walk but were also grepped: no reference.
• Rollback: git revert this commit. It is deliberately a single self-contained commit touching one token, so the revert restores the declaration, the twMerge entry, both test lists and the exemption together; nothing else rides along.
• Provider or production effects: None. No provider-backed gate, eval, or hosted check was run, and none is required — the classifier returns clinicalRisk: false, operationalRisk: false, ragRanking: false for this file set.

Clinical Governance Preflight

Not applicable. This change touches none of ingestion, answer generation, search/ranking, source rendering, document access, privacy, production env, or clinical output — it deletes one unused typography custom property and its config/test/doc references. classifyPullRequestFiles over the changed paths returns clinicalRisk: false and ragRanking: false, so neither this section nor a RAG impact: line is required by scripts/pr-policy.mjs.

Notes

• This completes the recorded next action on ledger row #297 verbatim: delete the declaration, drop the twMerge entry and the test list entry, update the TOKENS.md row, then delete the exemption. The row also anticipated the outcome — "deleting a step with zero consumers renders identically, but it edits @theme, so give it its own revertible PR rather than bundling it with a gate change" — which is why nothing else is bundled here despite the diff being small.
• With the exemption list empty, the declared-but-unconsumed gate now holds the line with no standing carve-out: the next unconsumed step fails the build by name rather than being absorbed.

---

Generated by Claude Code

[coderabbitai]

Warning

Review limit reached

You’ve reached a temporary PR review limit under our Fair Usage Limits Policy.

Your recent review volume is higher than typical usage, so adaptive limits are currently applied.

Next review available in: 21 minutes

Your organization has reached its usage spending cap. Adjust your spending cap in the billing tab.

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 1f4a991f-602b-4dad-8cdc-ae0621544ce1

📥 Commits

Reviewing files that changed from the base of the PR and between ee4e478 and 1896aff.

📒 Files selected for processing (9)

• docs/branch-review-ledger.md
• docs/design-system/GATES.md
• docs/design-system/TOKENS.md
• docs/outstanding-issues.md
• scripts/check-design-system-contract.mjs
• src/app/globals.css
• src/lib/tailwind-merge.ts
• tests/design-system-contract-utils.test.ts
• tests/tailwind-merge-config.test.ts

---

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

[supabase]

This pull request has been ignored for the connected project sjrfecxgysukkwxsowpy 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 ↗︎.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 6bf3c7b2a0

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[github-actions]

CI triage

CI failed on this PR. Automated classification of the 3 failed job(s):

• Production UI critical — needs investigation: inspect the failing step and uploaded diagnostics; rerun only after classifying the cause.
• Lighthouse budget (advisory) — needs investigation: inspect the failing step and uploaded diagnostics; rerun only after classifying the cause.
• PR required — needs investigation: inspect the failing step and uploaded diagnostics; rerun only after classifying the cause.

Compared with main CI run #9571 (cancelled).

Classification is evidence routing, not permission to ignore a failure. Exact quarantined Playwright identities remain governed by the flake ledger.

[cursor]

Babysit / thread disposition (#1797)

Unresolved Codex P2 on outstanding-issues ledger wording: claimed full npm run test vs focused proof. Disposition: non-blocking documentation nuance — verification section already notes the focused path; no code change required for mergeability. Review thread resolved via GraphQL after inline reply API returned 403 for this token.

Proceeding to squash-merge once branch policy clears.