Skip to content

[finding] 11 pm:blocked cards name a blocker that is already CLOSED — the unlock scan has not returned them to the queue #12448

Description

@os-zhuang

Filed unassigned, recording only — ⛔ no grading, and ⛔ this seat did not touch any of the cards below: none is domain:skills. Measured 2026-08-26 03:2x–03:3xZ by the skills seat (session session_01RMTpSRF5CjMmQBFfPtPCwJ) on a maintainer's ask to sweep the blocked population.

Method (so the numbers are re-derivable)

list_issues over objectstack for pm:blocked open (68 cards, bodies arrive in the same response), plus the full open-issue id set (348, four cursor pages) as the resolver. A card's blocker is CLOSED iff its id is absent from that set. Every Blocked-by: value was read with the line-anchored form the unlock scan itself greps.

⚠️Instrument validated before any zero was believed: of the 68, 25 carry a body-anchored Blocked-by: and 2 more mention it unanchored (prose — deliberately not a marker, per the state model), so the anchor demonstrably fires.

The finding — 11 cards whose blockers are ALL closed

cardlaneBlocked-by:blocker state
#9592devx#11501CLOSED
#11042devx#11930CLOSED
#11330spec#10812CLOSED
#11331spec#10812CLOSED
#11332spec#10812CLOSED
#11333spec#10812CLOSED
#11459spec#11195CLOSED
#11520services#11518CLOSED
#11969(no domain:*)#10348CLOSED
#12028cli#11997CLOSED
#12057spec#12049CLOSED

Two further cards are partially unblocked and correctly still parked — #11970 (#10348 closed, #11965 open) and #11971 (#10348 closed, #11968 open); they are listed so the count is legible, not as findings.

17 body-anchored pairs point at blockers that are genuinely still open — the healthy majority. One is cross-repo and unjudged this round: #12039objectstack-ai/objectui#6302.

The remaining 41 cards are UNJUDGED, not orphaned. They carry no Blocked-by: in the BODY, but h4BlockedNoBlockedBy's own text says a line parked in a COMMENT discharges the duty ("seats park the line in a comment on purpose, because rewriting a body through the MCP escaping hazard (#8813) is the riskier write"), and sampling confirms it live: #11617 carries Blocked-by: objectstack-ai/objectstack#11624 in a triage-seat comment written 2026-08-25T19:59Z, with two verified reasons. Judging that bucket needs 41 thread reads and was not done here. An unreadable channel is not an absent one (#4690).

Not diagnosed here, but worth the next reader's first question

The half-state patrol is report-only by construction — it never writes a label. The write that returns a card to the queue is the triage Routine's unlock scan. trig_01JfMEQ8ez7qFRWN46jr9JXk (pm trage) was measured stalled at next_run_at 2026-08-25T06:09Z earlier the same day (#12099), and agents cannot re-arm it (created via http_api; creator-only). Whether these 11 are a direct consequence of that stall, or were already parked before it, is not established by this measurement — the cards' transition timestamps would settle it.

Scope

objectui carries 44 open pm:blocked cards; they were counted but not judged (bodies were not fetched). cloud and objectos were not swept this round.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions