Port the decision-box dependency-flags section (instruction ④) of the half-state patrol to objectui's `check-half-states.mjs` copy (objectstack PR #18015)

[claude]

What

objectstack's half-state patrol (scripts/pm/check-half-states.mjs) gained a report-only inventory section beside its H17 trigger-file index: Decision-box dependency flags (instruction ④) — one line per open needs-user-decision card that some open card's Blocked-by: line names, a N of M count pair, a declared NOT-MEASURED number (H4's own count of pm:blocked cards with no readable line), and a population-scope line. Landed as objectstack PR #18015 (card objectstack#17968). objectui runs its own copy of the patrol (.github/workflows/half-state-patrol.yml + scripts/pm/check-half-states.mjs) and its anchor body does not carry the section, so instruction ④'s reading is available on one board and not the other.

Work item

Port the section to objectui's scripts/pm/check-half-states.mjs the way this repo ports patrol rows: decisionDependentIndex + renderDecisionDependents, wired into both renderers inside the reserved block below H17, the ruling constant verbatim, the self-test battery (two dependents · quiet inbox pair · comment channel with its unread control · NOT-MEASURED equals H4 · cap with overflow). No new request, no second parser: the section is one Map.get per decision card over the Blocked-by: index the copy already builds for H14. Reference diff: objectstack PR #18015 (+372 −5, one file).

Dedupe keywords

check-half-states · instruction ④ · decision-box dependency flags · Blocked-by reverse index · patrol port

Source

Filed by the skills execution seat (session session_01DAcomhvR9kKizeYgg89Vo8) at objectstack PR #18015's landing, as its ACCEPT (objectstack#17968 comment 5653169259) said it would. Left bare for triage per ruling ③ (grading and routing are triage's; this filer did not pull the board).

---

Generated by Claude Code

[claude]

pm:retriage — the work item as written cannot land here; asking triage for the shape

domain:skills execution PM seat (objectstack #7623), session session_01DAcomhvR9kKizeYgg89Vo8, 2026-09-14T08:24Z. Raised at dispatch-time premise check, not from the dev; pm:queue and priority:p3 left in place (this seat does not grade or reshape). Readings below were taken by the seat within the ten minutes before this stamp.

What falsifies the premise

• scripts/pm/check-half-states.mjs in this repo is a byte-pinned verbatim port of objectstack's file: scripts/upstream-port-pin.json on origin/main (c5cd9c01) pins it at upstream ref bf10debd587f6ba891be9eadc2b76c91e15bd82b, digest 449a0aec…ac9, with 11 declared divergences, and check:upstream-port-parity (wired in lint.yml) reverses the divergences and compares SHA-256 — byte equality, no "close enough".
• The section this card asks for is absent from the copy — git grep -c -E 'renderDecisionDependents|decisionDependentIndex' origin/main -- scripts/pm/check-half-states.mjs → 0 here, control 13 on objectstack origin/main — and it lives upstream at a39e385eed (objectstack PR #18015), far ahead of the pinned ref.
• A hand edit of the copy reds the gate by construction, measured on this board already: PR fix(pm-gate): H26 reports no SCHEDULED releaser, not an impossible one #9391 (objectui#9317, the domain:devx seat) hand-edited six sites of this same file, self-test green, and pnpm check:upstream-port-parity answered exit 1 「reconstruction does not match the pinned upstream blob」 on the change and exit 0 on the unedited base. Its dev's count: the copy is ~11,000 lines behind (13,103 here vs 24,443 upstream).
• The gate's own header names the only sanctioned mover: node scripts/check-upstream-port-parity.mjs --resync with the fetched upstream file and --ref naming the upstream commit, which rewrites the ported file wholesale and re-applies the declared divergences forward. So "port the section the way this repo ports patrol rows" resolves to a re-sync of the pin to a ref at or after a39e385eed, which carries every upstream change since bf10debd5, not one section.

What is asked

One of the following, decided by triage (or the maintainer on the direct channel):

• A — reshape this card into 「re-sync the pinned patrol copy to an upstream ref ≥ a39e385eed」 (instruction ④ arrives with it, as does every other upstream row; H26's corrected wording arrives once objectstack fixes it at the producer, per PR fix(pm-gate): H26 reports no SCHEDULED releaser, not an impossible one #9391's route 3). Routing is triage's: the pin and the parity tool are domain:devx @ objectui's (PR hooks: the three remaining guards name the environment their hatch variable must be set in, never a command prefix #9300), the patrol's content is domain:skills's. Recommended.
• B — close this card not_planned as "arrives with the re-sync", with the re-sync card filed as its successor (no such card exists: 388 open issues scanned, control hits on Port the decision-box dependency-flags section (instruction ④) of the half-state patrol to objectui's check-half-states.mjs copy (objectstack PR #18015) #9395 and finding(pm-gate): H26's "can never CLOSE" premise is asserted, not counted — and check-half-states.mjs refutes it twice in its own text #9317 titles only).
• C — declare a divergence for a hand port. Not recommended: the gate's header rules that the divergences are 「a checklist, not a licence」 and 「the set is meant to SHRINK」, and PR fix(pm-gate): H26 reports no SCHEDULED releaser, not an impossible one #9391's dev declined the same route for the same reason.

Adjacent half-state, named not touched

#9317 carries pm:blocked with no Blocked-by: line in its body and no visible transition comment: its timeline shows nothing after the pm:retriage flip at 2026-09-13T12:11Z, which is the shape a suspended account's erased events leave behind (objectstack #18147). The same decision is pending there. Once a re-sync card exists, both cards can carry Blocked-by: pointing at it.

Dedupe keywords: upstream-port-parity · upstream-port-pin · re-sync · check-half-states · instruction ④ · decision-box dependency flags.

---

Generated by Claude Code