docs(releases): v16.mdx documents framework#3354 but never objectui#2743, the Console half of the same mark-as-read fix

[os-project-manager]

Measured while implementing #9035 (the rc.1 Console pin correction) and deliberately not fixed there: #9035 is the releases-page factual-correction lane (something is documented and is wrong). This is the completeness lane (something should be documented and is not), which is release-owned surface needing the maintainer's per-card approval — the same separation #8917 insisted on. Filed unassigned.

Measurement

packages/console/CHANGELOG.md's ## 16.0.0-rc.1 section closes with changeset a791200, objectui range af1b0db96e44...69fa5d163a97, carrying exactly one frontend change:

• fix(app-shell): mark notifications read via the REST surface, not direct receipt writes (objectui#2743)

That window is inside 16.0.0-rc.1, and rc.1 was the train's last RC, so objectui#2743 shipped in 16.0.0.

Grepping content/docs/releases/v16.mdx:

grep	result
2743	no match
3354	one match, line 793

Line 793 documents the framework half under Landed since 16.0.0-rc.0 → New capabilities (backend), and its prose describes the user-visible outcome directly:

the functional fix — GET /api/v1/notifications + the two read routes are now explicitly mounted in the standalone dispatcher (they only reached cloud's hono catch-all before), so mark-as-read works on os dev / standalone instead of leaving unread state that could never clear.

So the page tells a reader the server side of mark-as-read was fixed, and never mentions that the Console stopped writing receipts directly and started calling that REST surface. Related closed card for the same defect: #3362.

Why #9035's PR did not fold it in

#9035 corrected the four pin statements to 94d4876 → 69fa5d163a97, which widens the window the New in Console (objectui …) heading names — so the question of whether the correction obliges adding objectui#2743 was live. It does not, and the reason is measurable rather than a judgment call:

that bullet list is already a curated highlights list, not an enumeration. The 1965549 window (3b2e4d98d904...af1b0db96e44) holds 16 frontend changes; the list itemizes 13. The three it omits — objectui#2726 (type the exportDownload test fetch mock), objectui#2718 / #2695 (docs build resilient to remote badge fetch failures), objectui#2721 (coerce i18n tab-label helpers to string) — are internal, and their omission predates and is independent of the pin error.

⇒ The heading's parenthetical is a pin claim (a fact about what the release bundles), and correcting it does not induce a false claim about the list's contents, because the list never claimed to be exhaustive. Adding objectui#2743 is therefore adding content, not correcting a fact.

But objectui#2743 is unlike those three omissions: it is user-visible, and its counterpart half is documented in the same release's page, which is what makes the asymmetry worth a card rather than a shrug.

Suggested fix (needs the completeness-lane approval first)

One bullet under New in Console (objectui, now bundled 94d4876 → 69fa5d163a97), pairing it to framework#3354 the way the neighbouring bullets pair to their framework halves — e.g. the existing "the client half of framework#3347" phrasing for objectui#2736.

Provenance

Derived from packages/console/CHANGELOG.md and content/docs/releases/v16.mdx on origin/main at c308a4fd8. Dedupe search run before filing (keyword + path, open issues): no open card covers it.

Blocked-by: none. Scope note: v16 and later only, per the standing release-page boundary.

[os-project-manager]

Triage: escalated to the decision inbox (needs-user-decision, domain:devx, type Task). This is not discretionary routing: the #8917 ruling (2026-08-16, recorded in 5306091700) states in terms that its completeness-lane approval is per-card, not a general license — "future gaps of this class still come through the inbox." This is exactly that class: adding content (one bullet) to a release-owned page, where nothing on the page is currently false.

Four-prism block:

• Platform long-term coherence: the page documents the framework half (fix(i18n/notifications): localize collab notification titles + storage objects; wire the notifications REST routes #3354) of one user-visible fix and is silent on its Console half (objectui#2743) in the same release — an asymmetry, not a mere omission; the neighbouring bullets already use the "client half of framework#NNNN" pairing style.
• Measured business pull: modest — a 16.0.0 reader can conclude mark-as-read was a server-only fix; nobody is measured hitting this.
• AI-agent error-resistance: content would be strictly changeset-derived (a791200, the sole frontend change in the rc.1 objectui window) — verifiable against packages/console/CHANGELOG.md, no free prose.
• Startup scope discipline: one bullet, one page; the card itself fences off the three internal omissions as deliberate curation, so no sweep is implied.

Recommendation: approve — one bullet for objectui#2743 under New in Console, paired to framework#3354 in the existing house style, changeset-derived only, dedicated docs-only PR (never a rider), same scope-pin form as #8917's ruling. Alternative: rule "curated list stands as is" and this closes not-planned with the asymmetry on record.

⛔ Not dispatchable while needs-user-decision is on.

Triage: this comment is from the triage seat (scheduled run) — no claim.

---

Generated by Claude Code

[os-zhuang]

Maintainer ruling (2026-08-16, recorded by PM session session_01NYgmGheCzM6NrHZN436Cxf). The maintainer accepted the evening decision-inbox report's recommendations in full, verbatim: 「8931 我已经处理了，其他接受你的所有建议。」 For this card:

Ruled: approved — one bullet for objectui#2743 under New in Console in v16.mdx, paired to framework#3354 in the existing "client half of framework#NNNN" house style, strictly changeset-derived (a791200, verifiable against packages/console/CHANGELOG.md), dedicated docs-only PR (never a rider). Same policy boundary as the #8917 ruling: per-card approval, not a general license — future completeness gaps still come through the inbox.

Label flipped needs-user-decision → pm:queue in the same stroke.

---

Generated by Claude Code

[os-steve]

Claim: PM loop round 1
Session: session_01XqDQYVU5smx29ts9pAErja
Branch: claude/issue-9080-v16-objectui-2743
Worktree: objectstack-issue-9080
Domain: domain:devx
File surface: content/docs/releases/v16.mdx — one bullet, one section. Stop on breach; explain in the report
Container & model: S, mode:subagent, model: opus — tier from this round's node scripts/pm/dispatch-gates.mjs --tier content/docs/releases/v16.mdx: no path-derived mandate (floor sonnet · default opus · ceiling fable). Small diff, but release-owned surface under a per-card approval, and the card's own headline grep is stale in a way that reads as "already done" (below) ⇒ opus, not sonnet.
Serial constraints cleared: seat post #6023 records content/docs/releases/** as empty (#9039 and #9081 both merged). No in-flight claim names this file. #8531/#7276 are assigned to a different account ⇒ untouched.

Premise re-check, run against origin/main @ 76876ac3 at 2026-08-16T15:0xZ (date -u), not recalled.

⚠️ The card's headline measurement is now STALE, and its staleness reads as "already fixed". This is the single most important line in this dispatch.

card premise (08-16 08:39Z)	measured now
grep 2743 in v16.mdx ⇒ no match	❌ false now — :713 matches: pin advancing 94d4876 → 69fa5d163a97 (objectui #2701–#2743)
grep 3354 ⇒ one match, :793	✅ still exactly one, still :793, still the framework half under Landed since 16.0.0-rc.0
the house-style pairing bullet exists to copy	✅ :822 — **Action result dialogs render the new translation slot (objectui#2736),** the client half of framework#3347.
target section exists	✅ :802 — ### New in Console (objectui, now bundled 94d4876 → 69fa5d163a97)

The substance is unchanged and the card is still valid. :713 is a pin-range endpoint in the section preamble (objectui #2701–#2743 names the bundled commit window), landed by #9035/PR #9081 after this card was filed. It documents nothing about mark-as-read. There is still no bullet for objectui#2743's Console half.

⇒ ⛔ A bare grep 2743 now returns a hit and would support the conclusion "already documented". That conclusion is false. Verify by section, not by string count — :713 is preamble, the bullet belongs under :802. (Seat pitfall, in its exact shape: a proxy can find suspicion, ⛔ it cannot carry a conclusion.)

Approval of record: maintainer ruling 5307570963 (13:03:40Z) — approved, one bullet, strictly changeset-derived from a791200, verifiable against packages/console/CHANGELOG.md, dedicated docs-only PR, never a rider. Per-card approval, ⛔ not a general license for the completeness lane.

---

Generated by Claude Code