docs(metadata-protocol): re-state row 6's trim justification on its live basis

[claude]

Fixes #15853

What was wrong

packages/metadata-protocol/src/protocol.destructive-409-face-inventory.test.ts argues, in the section titled Why row 6 is the one that forbids the trim, that duplicatePackage has no response schema in packages/spec at all. That clause has expired. Nothing is red — it is prose in a test docstring, not an assertion — but it is the argued justification for a restriction that is still in force, so the next person to re-derive the argument re-derives it from a false premise.

The measurement (re-located by TEXT, on a4816a79d)

⛔ Not inherited from triage — re-run here, with the control.

claim	where it actually is now
the false sentence	protocol.destructive-409-face-inventory.test.ts:64
DuplicatePackageResponseSchema	packages/spec/src/api/package-lifecycle.zod.ts:199
the route bound to it	packages/runtime/src/route-ledger.ts:369 (responseSchema: 'DuplicatePackageResponseSchema')
its conformance pin	packages/spec/src/api/package-lifecycle.test.ts:225

Control: git grep -l DuplicatePackageResponseSchema hits 12 tracked files — the search is live, not a silent zero. (Triage read the same 12. An earlier attempt of mine with an --include filter read 11 because the filter list omitted .mdx; the unfiltered git grep is the reading.)

Why row 6 forbids the trim TODAY

This is the part the card asked for first, because it is what the docstring was standing in for. Three legs, each independently measured:

1. Transport — duplicatePackage reports a per-item failure as response DATA on a 200 (POST /packages/:id/duplicate). No HTTP error boundary runs, so details.issues never exists.
2. Producer — its declared return still types failed[] inline with exactly type, name, error (protocol.ts:18403), and the push at protocol.ts:18714 emits exactly those three keys. Section 3 of this very file already pins that at runtime: expect('issues' in entry).toBe(false).
3. Spec — the leg that changed. A response schema now DOES exist and the route IS bound to it. But it is a describe-only transcription of that inline return: DuplicatePackageResponseSchema.failed[] declares type, name, error and no issues, no code. The comparator that separates row 6 from publishPackageDrafts therefore still holds, on a sharper axis — PublishPackageDraftsResponseSchema.failed[] (packages/spec/src/api/protocol.zod.ts:1097) does declare issues: z.array(RuntimeAuthoringIssueSchema).optional(), and that DECLARED channel is what the message was trimmed against there.

⇒ #10524's order — declare a structured channel, and only then trim — is still unsatisfied on row 6, so the verdict is unchanged. What the correction changes is the reason and its strength: the schema's existence does not supply the missing channel, it freezes its absence into a published, route-bound surface. That raises the bar rather than lowering it — declaring issues here now moves a spec schema carrying its own conformance pin as well as the inline producer type.

⛔ No assertion changed. Prose only, one file, +32/-6.

Wider staleness sweep (reported, ⛔ not fixed here)

Triage's this row is the only one reading searched only for that wording. I swept the other forms while in the file:

• Every referenced file path resolves: protocol.destructive-gate-reachable-types.test.ts, protocol.invalid-metadata-422-face-inventory.test.ts, meta-compound-save-force-parity.test.ts (the last in packages/rest, as stated).
• Every at-link symbol resolves in protocol.ts: ObjectStackProtocolImplementation, saveMetaItem, duplicatePackage, destructiveChangeRemedy.
• The table's the 409 is raised in ONE place claim still holds — the only assignment of code = 'DESTRUCTIVE_CHANGE' is protocol.ts:14998. ⚠️ My first pass at this used pathspec packages/*/src/**/*.ts, which silently excludes top-level package files and returned a near-false zero; the reading above is from packages/**/*.ts plus a firing positive control.
• One real second instance, in a different file: the identical expired sentence stands at the raise site itself, packages/metadata-protocol/src/protocol.ts:14967. Filed as [finding] The same expired justification stands at the 409 raise site: protocol.ts argues duplicatePackage has no response schema in packages/spec at all, which is false #16125 (no assignee, finding). ⛔ Deliberately untouched here: protocol.ts is held by open PR docs(metadata-protocol): the getMetaItems gate enumerates six rest-server.ts call sites, not five, and names the transitive diagnostics door #16117, and a bounded in-place fix requires the file to carry no other claim.

Adjacency, run rather than assumed

PR #16117 changes exactly one file, packages/metadata-protocol/src/protocol.ts (read from its files endpoint). This PR changes exactly one file, the inventory test. Disjoint. check:single-claim-paths self-test passes (54 cases); the wired run needs a PR number and is reported below.

Verification

Built dep closure first (pnpm --filter '@objectstack/metadata-protocol^...' build), so nothing below reads stale dist/. Heavy steps went through scripts/pm/os-verify-lock.sh; verdicts quoted from its own VERDICT line, exit codes captured by redirect-then-read, ⛔ never through a pipe.

• vitest run src/protocol.destructive-409-face-inventory.test.ts src/protocol.destructive-gate-reachable-types.test.ts — 2 files, 30 tests passed.
• pnpm --filter @objectstack/metadata-protocol typecheck — clean. ⚠️ Verified this is a real reading and not a vacuous one: tsc --listFiles shows the edited test file is in the program (the package's tsconfig does not exclude tests).
• Ablation for the published-surface question — subject file swapped back to a4816a79d, rebuilt, then restored. Mutation proved on disk before measuring (new text 0 hits, old text 1 hit); all 24 files under dist/ byte-identical across the ablation, sourcemaps included; restore proved after (hash-object equals the HEAD blob, git diff HEAD empty), then a final rebuild so dist/ matches HEAD.
• Gate family derived, not recalled: node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack (264 families across 31 workflows). All 50 derived pnpm check:* plus 37 derived node-script invocations were run — ⛔ the family was not narrowed on the grounds that the diff is "only a docstring".
• Line-number anchors, checked explicitly — this diff is +26 lines (665 to 691), so everything below the docstring moved. check:system-context-census green (105 elevation read sites in 19 packages across 44 files, all anchored; 140 anchors resolve, 27 declared non-read), check:doc-anchors, check:adr-anchors, check:adr-symbol-anchors, check:scripts-symbol-anchors all green. Direct search: nothing anchors a line number in this file (positive control: content/docs/permissions/system-context.mdx carries 82 such anchors, so the pattern search fires).

Declared non-readings — ⛔ these are NOT green

• check:dual-build-cjs-loads and check:published-readme-exports exit 3, PREREQUISITE NOT MET in their own words: both read built output and the whole monorepo is not built. Their self-tests pass. Structurally unreachable by this diff — the ablation above proves every published artefact of the only package touched is byte-identical.
• packages/spec check:react-declaration-parity exit 1: it needs a browser-dumped objectui sdui.manifest.json, and packages/console/dist/sdui.manifest.json does not exist in this repo. A missing cross-repo prerequisite, not a finding, and it reads nothing in this diff.
• check-single-claim-paths.mjs and check-partof-closing-keyword.mjs bare runs exit 2, NOT WIRED in their own words — they need PR_NUMBER. Run against this PR once it exists.

Release impact

None — skip-changeset. Measured, not assumed: @objectstack/metadata-protocol publishes files: ['dist', 'README.md', 'CHANGELOG.md']; the only changed file is a src/*.test.ts that no source imports (index.ts imports no test module) and that appears in no sourcemap's sources[] — it is named only inside the embedded sourcesContent of protocol.ts, which mentions it in three comments. The ablation confirms dist/ is byte-identical either side. This PR releases nothing.

Report follows on #15853, first line os-dev-report.

Generated by Claude Code, session session_01ARYe3yQTQCUFm5qPYNgKaJ.

---

Generated by Claude Code

[github-actions]

📓 Docs Drift Check

Nothing in this diff resolved to a documentable surface (no symbol, route or SDK anchor derived from 0 changed package(s)), so this run has no opinion about the docs.

What this run could not see

• a page that states a rule by its inputs shares no identifier with the emitter that implements the rule, so an emitter-only diff cannot list it — not on this run and not on any run. Measured on fix(driver-sql): emit varchar(maxLength) for a text field a declared index keys on #11430: content/docs/protocol/objectql/types.mdx documents the text-family column mapping by the ObjectQL type names it maps FROM (text / textarea / html) while the diff changed createColumn; it went unlisted, and it was the page that diff falsified, in four places. No shared token exists to detect this on, so a rule your change carries has to be re-read by hand in the pages that restate it.

Coarse fallback — 0 page(s) merely mention a changed package (the pre-#9192 predicate, kept for the deliberately-wide backstop): node scripts/docs-audit/affected-docs.mjs --json 932acc3df41e016812cfc17ffe1527fd7b5df538 → packageMentionDocs.