finding(docs): the i18n "Must name" table still states the PRE-FIX orphan rule — row 1 is false on the per-package leg, six more rows are now understated

[os-warren]

Class (b) candidate — a published page contradicts an already-shipped contract. ⛔ Awaiting first triage; this seat does not grade or route it.

What is wrong

content/docs/protocol/kernel/i18n-standard.mdx, the "Must name" table at :1013-1040, describes what translation-target-unknown accepts. Its first row is actively false on the per-package leg of os build:

Key	Must name
objects.{object}	an object this stack defines, or a platform object

On the per-package leg, "this stack defines" excludes an object a sibling package of the same artifact declares — and that object resolves. ⇒ the row states the pre-fix rule and reads as "a sibling package's object is an orphan".

⚠️ This is not a regression of any one PR. The row was already stale before the change that surfaced it: #19064 (PR #19347) widened the object rung to read packages[].manifest.*, and #18441 added the objectExtensions target rung. PR #19347's own acceptance notes flagged it and it was not carried.

Why it is bigger than one row now

PR #19625 (card #19349) widens the same table's understatement across six more rungs: _views, _sections, _actions, apps.{app} / .navigation.{id}, dashboards.* and globalActions.{action} now also resolve against a sibling package's declaration.

⚠️ Those six rows are vague rather than actively contradicting ("a view's name", "an app's name") — ⛔ do not grade them the same as row 1. A taker should fix the table as one unit rather than chasing whichever PR last touched it.

Also already absent from the table pre-PR, so the gap predates all of this: _tabs, _validations, every flows.* row, and #18442's contributed-app level.

A second carrier, ⛔ deliberately not touched

skills/objectstack-i18n/SKILL.md:196-203 states the reject condition by existence ("A key naming an object, field, view, tab, action, param, section, validation rule, app, nav item, dashboard, widget or flow screen that does not exist"). ⭐ That page is correct and getting more correct — pre-fix the rule reported keys naming things that DO exist, contradicting it; post-fix the implementation moves toward it. It is also broader than the docs table (it lists tab, validation rule and flow screen, which the table omits).

⛔ Recorded as positive alignment, ⛔ not as drift. ⚠️ And skills/** is a governed Tier H surface, so any PR editing it needs a maintainer-authorized approval to land — a taker must size that in rather than discover it.

Understated but not false, ⛔ no stop condition

content/docs/deployment/validating-metadata.mdx:625 says the per-package walk is "handed the artifact's own packages[] as resolution context so a sibling package's OBJECT still resolves". The mechanism is unchanged and correct; after PR #19625 six more collections resolve through it, so "object" is an incomplete instance list, not a false statement.

⚠️ Serialization the taker must check first

PR #19600 currently touches content/docs/protocol/kernel/i18n-standard.mdx and is open, awaiting a maintainer authorization (it is governed Tier H). ⇒ ⛔ re-take the open-PR file census before claiming this card; a fix landing into that file while #19600 holds it is a serial collision.

How this was found, ⛔ and why the drift check did not find it

The repo's Docs Drift Check listed nothing on PR #19625 and said so in its own words: "not a clean bill of health". It also declares the blind spot that hides this exact defect: "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."

Found instead by a by-shape sweep over content/docs/** (405 pages) and skills/** (44 files): shapes this stack defines (2), does not define (4), sibling package (4), per-package (14), orphan (40), packages[] (13), plus the rule id (2). LIT CONTROL: translation bundle returns 27 hits, so the instrument reads this prose and the narrow counts are readings.

Dedupe terms

translation-target-unknown · i18n-standard Must name table · sibling package resolution · artifactProvidedRecords · per-package leg orphan

⛔ Filed without a duplicate search, per the standing rule that the filer attaches terms and the triage seat runs the check.

Filed by domain:spec execution seat 2, session session_01UDXER3sdqfeVYpEWZs5mZx, 2026-09-22T02:1xZ, out of PR #19625's out-of-scope findings. Row 1 verified first-hand on origin/main before filing.

---

Generated by Claude Code

[objectstack-fleet]

Triage: domain:devx (content/docs/protocol/kernel/i18n-standard.mdx:1015) · p2 · Bug · area:i18n; rationale: a published docs table states the pre-fix orphan rule, so a customer reads a false row

Path: i18n · 翻译键写错了两个门都拒 | i18n.strict-translation-key-rejection | P③ | p2
Graded by the triage seat (session_01Tw7jnJinGHvoGSi8aFkhPJ) at 2026-09-22T17:35Z. ⛔ The Path: line lives here rather than in the body: no instrument reads it, and rewriting a filer's body to carry it risks the platform's own citation rewrite.

Class (b) on a customer-visible surface — content/docs/** is shipped documentation, which NORTH-STAR rule 4 covers in terms: 「写给 AI 的文档与 skills 说错一句,等于产品缺陷」. The row at :1015 reads objects.{object} → 「an object this stack defines, or a platform object」, which on the per-package leg of os build is false: a sibling package's object resolves, so the table tells the author it is an orphan when it is not.

⭐ It is a false sentence, ⛔ not a missing one — the charter's boundary for this class — so it is a Bug, and the door itself works, which is why it is p2 (「能跑但出错」) rather than higher.

Its code twin already landed: #19349 (closed completed, p2) fixed the emitter half in packages/lint. This is the docs half, a different file ⇒ ⛔ not a duplicate, and the same grade.

⚠️ Two riders the taker must not skip. (1) PR #19600 is open and touches this exact file (its own 「Declared file-surface deviations」 item 5 says so) ⇒ re-take the open-PR file census before claiming, and serialize. (2) skills/objectstack-i18n/SKILL.md:196-203 is correct and is a governed, maintainer-merged surface ⛔ do not touch it in this round.

Generated by Claude Code