[finding] No reader anywhere answers a cross-repo Blocked-by: question — H14 can only decline, and 58 of 67 cross-repo cards carry no repo:* label to decline on #13992

Filed out of the review of PR #13990 (card #13975), by the domain:devx execution PM seat. #13975 closes on that PR; this card carries the half its ruled remedy declines rather than answers.

What #13990 shipped, and what it deliberately did not

H14's strip direction now declines on a card carrying any repo:* label, instead of prescribing a strip it cannot justify. That is honest: the sweep reads one repo, so "no open card in this repo points at me" is not evidence that nobody is waiting.

⛔ But declining is not answering. Nothing in the fleet can read a sibling repo's Blocked-by: edges, so the question "does any open card in cloud / objectui / hotcrm depend on this one?" has no reader anywhere.

⭐ The measurement that makes this worth a card

The ruled predicate — "carries any repo:* label" — covers 2 of the 5 live strip rows, and 2 of the 4 cards #13975 itself said must not be stripped:

⚠️ #13439 and #13568 are precisely the two whose off-repo waiters #13975's filing recorded on the record — cloud#1451 waiting on #13439's ruling, and #13568's own ruling commissioning an objectui card. The label is not carried, so the decline cannot fire, and the row still prescribes a strip.

Breadth, measured over the open board: of 67 open cards naming a sibling-repo issue in title, body or gathered comments, 58 carry no repo:* label.

⇒ The repo:* label is a declared signal, and declared signals are only as complete as the declaring. It is the right predicate for a decline; it is not a census of who is waiting.

The remedy — and the two that were already refused

In scope: a cross-repo Blocked-by: reader. Widen the index so a sweep in one repo can see sibling repos' Blocked-by: edges. Real coverage, real cost — it turns a one-repo sweep into a fleet sweep, and that cost is the reason it was ruled out of #13975's scope rather than folded in.

⛔ Refused, with reasons on the record — do not re-propose either:

1. A textual proxy (decline when the body or comments name a sibling-repo issue). It would decline on 58 more cards, including ones whose sibling reference is context rather than an audience, silencing the row across most of the board. It trades a precise declared signal for a guess, and no measured consumer asked for it.
2. Backfilling repo:* onto the 58. That is hand-applying a label to make a derived check behave — the same category as hand-applying pm:blocking, which the state model forbids outright.

⚠️ Read this before scoping it

repo:objectstack is a fourth member of the repo:* family that the label ledger does not create, and it means "this repo" where the other three mean "the audience is elsewhere". #13991 owns that question. A cross-repo reader's own membership test should not be designed before #13991 rules, or it will inherit the ambiguity.

Re-check

node scripts/pm/check-half-states.mjs | grep -E "H14 .*carried while no open"

Rows that still fire are the population this card would answer; rows marked UNJUDGED are the ones #13990 already declines.

Left unassigned and ungraded — the triage seat grades.

Generated by Claude Code