docs(skills): correct six wrong shipped-API facts in the published objectui guides

[claude]

Fixes #7676

Six statements of a shipped-API fact in the published skills/objectui/guides/ bundle contradicted the built dist/*.d.ts. Each is corrected here to what the types declare, one line per fact. ⛔ No fence is rewritten and no os:check marker is added — the fences stay unmarked, and their remaining diagnostics are the prose placeholders they reference by design.

The sites were found by running check-skill-examples.mjs --measure over the whole corpus (that sweep's other result — that the guides' fences are fragments by construction and none is markable — is recorded on the card it came from, which is closed not planned).

改前 → 改后

#	guide · site	改前	改后	the shipped fact
1	mobile.md gesture fence	const listRef = usePullToRefresh(...)	const { ref: listRef } = usePullToRefresh(...)	returns an object — ref, isRefreshing, pullDistance — not a bare ref (packages/mobile/dist/usePullToRefresh.d.ts)
2	auth-permissions.md provider fence (and the nesting-order fence, same fact)	authClient={authClient}	authUrl="/api/v1/auth" client={authClient}	the prop is client, and authUrl is required (AuthProviderOptions, packages/auth/dist/types.d.ts:493)
3	auth-permissions.md fetch fence	createAuthenticatedFetch(authClient)	createAuthenticatedFetch()	it takes CreateAuthenticatedFetchOptions (only sameOriginOnly?), never the client; the bearer token comes from storage
4	auth-permissions.md guard fence	fallback={ the AccessDenied element }	fallback="custom" fallbackContent={ the AccessDenied element }	fallback is the closed union 'hide' | 'disable' | 'custom'; custom content goes in fallbackContent
5	i18n.md formatting fence	const { formatCurrency, formatDate, formatNumber, formatRelativeTime } = useObjectTranslation(); and formatDate(date, 'YYYY-MM-DD')	import { formatCurrency, formatDate, formatNumber, formatRelativeTime } from '@object-ui/i18n'; and formatDate(date, { style: 'short' })	all four are module-level exports (packages/i18n/dist/index.d.ts:49), not members of the hook's return, and each takes an options object, not a positional format string
6	project-setup.md provider fence	value={metadata}	adapter={myAdapter}	app-shell's MetadataProvider takes adapter and optional ttlMs (packages/app-shell/dist/providers/MetadataProvider.d.ts)

Row 5 is two lines because the fact is stated twice in the same fence: where the helpers come from, and how they are called. Correcting only the import would have introduced a fresh error on the positional format string, which is the same wrong fact one line down.

Differential evidence, per row

Method: the named diagnostic is present in a --measure log taken before the edit and absent from one taken after, with the fence's remaining diagnostics being the placeholder class. Nothing else in the corpus moved.

#	before	after
1	mobile.md:48:35 TS2322 — the pull-to-refresh result object is not assignable to the element ref prop	absent; fence keeps only TS2304 CardContent
2	auth-permissions.md:29:19 TS2322 — property authClient does not exist on the provider props	absent; the interim TS2741 authUrl is missing this exposed is also absent after the required prop was supplied; fence keeps only TS2304 AppContent
3	auth-permissions.md:110:53 TS2559 — AuthClient has no properties in common with the options type	absent; the fence now has no diagnostics at all
4	auth-permissions.md:242:56 TS2322 — an element is not assignable to the closed union	absent; fence keeps only TS2304 AccessDenied, BulkDeleteButton
5	i18n.md:98:11 / :27 / :39 / :53 TS2339 x4 — none of the four helpers exists on the hook's return	all four absent, and no new diagnostic replaced them; fence keeps only TS7031 implicit-any on its own untyped example props
6	project-setup.md:280:21 TS2322 — property value does not exist on the provider props	absent; fence keeps only TS2304 myAdapter

Corpus-level: semantic failures 81 → 80 over 95 judged fences, and exactly one verdict flipped — auth-permissions.md:105 (row 3) went FAIL → pass. It is now the first unmarked fence in skills/objectui/guides/ that compiles, so it is the one candidate a follow-up could opt in (floor 13 → 14). ⛔ Deliberately not marked here: this PR corrects facts and moves no marker.

Gate verdicts — union re-run on the final commit a7dfbcc

check:skill-examples        exit 0   Marked: 13 ts fence(s) (floor 13), 70 json fence(s) (floor 70)
                                     Semantic phase: 13 of 13 ts fence(s) judged, 0 failed.
                                     JSON phase:     70 fence(s) parsed, 0 failed.
                                     Bare `any`: 1 finding(s), 1 declared, 0 NOT declared, 0 stale.
                                     Every marked skill example holds up against the built types.
check:skills-paths          exit 0   OK (88/89 stated path(s) resolve across 20 guide file(s); 1 baselined)
check:skill-eval-tokens     exit 0   Every must_contain token is taught by its own skill bundle.
check:doc-fences            exit 0   every TypeScript block in 227 document(s) is fenced ts/tsx/typescript
check:control-bytes         exit 0   OK (scanned 6258 tracked text file(s); skipped 85 binary)
check:governed-queue-guard  exit 0   132 cases pass

The marked population and both floors are unchanged at 13 ts / 70 json — no marker moved, so MARKED_FLOOR is not re-pinned.

Changeset — objectui has no skip-changeset label, so the script decides:

node scripts/check-changeset-presence.mjs   exit 0
Compared the working tree with a472b0716 (merge-base with origin/main): 4 file(s) changed,
0 of them published source of a package the release covers, 0 of them a manifest whose
published contract moved, 0 under a package changesets ignores, 0 changeset(s) added.
✅  No source or published contract of a released package changed in this range, so no changeset is owed.

Ratchets (whole-file lines, before → after): auth-permissions.md 348 → 348; i18n.md 187 → 185; project-setup.md 345 → 345; mobile.md 168 → 168. Published SKILL.md bundle sum 137 lines, untouched. Net −2 lines — no content was purchased.

Review notes

Governed surface skills-catalog. Stays a draft; two approvers; human merge. The packages are untouched — every change is documentation converging on the API that already ships.

---

🤖 Generated with Claude Code

https://claude.ai/code/session_019RfFHiRCSs3JXLK4cwcfox

---

Generated by Claude Code