docs(standards): (scope) is the component, never the ticket number

[LukasWodka]

Four PRs failed closing-ref today for the same reason — .github#349, #350, #352, #353 — and all four were fixed by retitling. That's a missing rule, not four mistakes.

The collision

The canon says commit subjects are type(scope): summary, "referencing the ticket (backend#1234)". Read quickly, (scope) looks like the ticket slot — so sessions write sec(2157): ….

closing-ref then reads a bare number in the title as a promise to close that ticket, and fails unless the body carries Closes <owner>/<repo>#N. For partial work that keyword would be actively wrong. All four PRs were partial work, all four correctly said Part of tracebloc/backend#N, and all four went red for saying so.

The canon's own example is already the qualified form (backend#1234), not a bare number. What it never says is that the number doesn't belong in the subject, and that a bare one carries a specific machine meaning.

Why not loosen the gate

The gate is right, and its failure message is one of the better ones in the fleet — a title reference is inert, and it can't tell which repo owns a bare number, so it would happily accept a link to any repo at that number. The defect is the convention pointing one way while the check reads the other. Fixed on the convention side.

Verification

• classify reports IN_SYNC — canon and CLAUDE.md agree
• standards-sync selftest 31 checks, 0 failed

Follows the same route as .github#350: the canon changes here, and the 19 CLAUDE.md copies pick it up on the next standards-sync dispatch with create-prs: true.

---

Note

Low Risk
Prose-only updates to engineering standards with no runtime, auth, or data-path changes.

Overview
Updates the org-wide Branches & PRs bullet on naming and commits in org-standards.md and the synced block in CLAUDE.md.

The change spells out that (scope) in type(scope): summary is the component (e.g. mint-scope, kanban), not the issue number, and that a bare number in a PR title (e.g. sec(2157): …) is treated by the closing-ref check as a promise to close that ticket unless the body has Closes <owner>/<repo>#N. For partial work, the canon now says to use Part of tracebloc/backend#N in the body and keep the ticket number out of the subject.

This is documentation-only alignment with existing gate behavior (not a change to closing-ref itself); downstream CLAUDE.md copies pick it up via standards-sync.

^{Reviewed by Cursor Bugbot for commit 1a279df. Bugbot is set up for automated code reviews on this repo. Configure here.}

[aptracebloc]

Reviewed — accurate and useful. Clarifies that (scope) in type(scope): summary is the component (mint-scope, kanban), never the ticket number, because a bare number in the title (sec(2157): …) is read by closing-ref as a promise to close that ticket and fails the check unless the body carries Closes <owner>/<repo>#N. That matches the actual closing-ref behavior, and the guidance (number in the subject only when the PR truly closes it; otherwise Part of tracebloc/backend#N in the body) is correct. Same sentence added identically to CLAUDE.md and org-standards.md. Docs-only, green.

Approving.

— drafted with Claude Code

[saadqbal]

Good rule, and the right two lines to spend on it — sec(2157): as the worked example is exactly the shape that bites, because a bare number resolves against this repo and the ticket lives in backend.

One refinement for whenever this text is next touched, non-blocking: "the check fails unless the body carries Closes <owner>/<repo>#N" describes the fix correctly but the mechanism slightly loosely. The gate reads GitHub's computed closingIssuesReferences, not body text — so a body that carries the keyword and the ref but not adjacently (Closes the interim half of tracebloc/backend#2327) satisfies the sentence as a reader might parse it and still computes zero links. Following your line literally produces the right result; it's the loose reading that fails. Worth "immediately followed by" if you're ever back in here. 👍