Skip to content

source-control: observe a ready run that skips the local verify #8994

Closed

Description

@kyle-sexton

Problem

#8753 makes /source-control:pull-request ready skip the local full-suite verify when a passing CI test-lane run exists at the head. A lane can prove the skill text and the venue script with tests, but not a live ready run, which marks a real PR ready.

Criterion moved from #8753

  • A ready run on a PR whose CI lanes are healthy reports the verify step as skipped (CI covers it) and flips the PR in minutes.

Blocked by: #8753

Activity

  1. added
    needs-humanHuman-in-the-loop required; autonomous sessions must not resolve items carrying this.
    status: confirmedA maintainer confirmed the item is real and in scope; ready to be worked.
    on Oct 22, 2026
  2. github-actions commented on Oct 22, 2026

    @github-actions
    Contributor

    Proposed work class: none

    Thanks for the report. Automated triage handed this issue to a maintainer.

    Automated triage. A maintainer reviews it.

  3. removed
    needs-humanHuman-in-the-loop required; autonomous sessions must not resolve items carrying this.
    on Oct 25, 2026
  4. kyle-sexton commented on Oct 25, 2026

    @kyle-sexton
    MemberAuthor

    Released for re-triage under the owner's 2026-10-24 rule: an issue stops for a person only for an intent or vision question, a security call, a host or attended run, or spend outside prepaid credit. The hold label came off and intake triage runs again now; it re-applies a hold, with a numbered question, when one of those applies.

    To hold this issue again, add needs-human (or status: needs-decision) and say why in a comment.

  5. added
    agent-readyFully specified and briefed; eligible for autonomous pickup from the frontier.
    on Oct 25, 2026
  6. added
    needs-humanHuman-in-the-loop required; autonomous sessions must not resolve items carrying this.
    and removed
    agent-readyFully specified and briefed; eligible for autonomous pickup from the frontier.
    on Oct 25, 2026
  7. melodic-automation-lanes commented on Oct 25, 2026

    @melodic-automation-lanes
    Contributor

    The work-resolve-issue lane did not start this issue: it needs a person to make it clearer first.

    Why, in the run's own words:

    This issue can't be worked yet, and this CI lane can't do the work it asks for. (1) The blocker #8753 is still open, with no linked branch or pull request (read https: //github.com/melodic-software/claude-code-plugins/issues /8753). The checkout doesn't contain its behavior either: searching plugins/source-control/skills/pull-request for a skipped (CI covers it) verify outcome or a CI test-lane skip of ready's local verify finds nothing. That leaves no feature to observe. (2) The only criterion is to watch a live /source-control:pull-request ready run mark a real PR ready, which needs an attended or host run against a live PR. This lane can't do that: it has no gh access and leaves its changes uncommitted. Question for a person: once #8753 lands, who runs the live ready on a PR with healthy CI lanes and records the result? Should this issue carry needs-human (an attended run) rather than agent-ready?

    What the run says it did not do:

    • The live observation of a ready run reporting verify as skipped (CI covers it) and flipping the PR: blocked on #8753 landing, and it needs an attended run against a real PR, which this lane can't perform

    Run: https://github.com/melodic-software/claude-code-plugins/actions/runs/40233438204

    To retry, answer the question here, remove needs-human and add agent-ready again.

  8. removed
    needs-humanHuman-in-the-loop required; autonomous sessions must not resolve items carrying this.
    on Oct 26, 2026
  9. kyle-sexton commented on Oct 26, 2026

    @kyle-sexton
    MemberAuthor

    Hold changed: this issue waits on #8753 (now a native blocked-by edge), not on a person. Under the owner's 2026-10-26 rule a person is involved only for an intent or vision change, an implementation finding that alters intent or the primary design, or a HIGH/CRITICAL security issue; none applies. Admission defers it until #8753 closes; the lane's open question is answered from the facts then.

  10. added
    host-runNeeds an unattended run on a host; the host runner takes it, not a person or the issue lane.
    on Nov 1, 2026
  11. kyle-sexton commented on Nov 1, 2026

    @kyle-sexton
    MemberAuthor

    Decision recorded.

    The blocker #8753 closed as completed on 2026-10-29, and its behavior is on origin/main b9ae6bbaed: plugins/source-control/skills/pull-request/SKILL.md:69 says the verify gate's suites are "skipped when CI's test lanes already passed on that head, reported as skipped (CI covers it)", and reference/ready-for-review.md:225 and :345 give the report line. What is left is the one criterion moved from #8753: a live ready run on a real PR. The issue lane cannot run it (2026-10-25 lane run: it "needs an attended or host run against a live PR"). Agents decided this on evidence under standing direction §1.

    Acceptance criteria:

    • A /source-control:pull-request ready run on a PR whose CI lanes are healthy reports the verify step as skipped (CI covers it) and flips the PR in minutes, recorded here in a Probe result: comment that links the PR.

    Route: host-run.

  12. kyle-sexton commented on Nov 3, 2026

    @kyle-sexton
    MemberAuthor

    Lane 3 claims this: the live ready run happens on its next PR once CI's test lanes pass.

  13. kyle-sexton commented on Nov 3, 2026

    @kyle-sexton
    MemberAuthor

    Probe note, criterion not met. Lane 3 ran ready on #12626. The venue read on its head printed tests ci healthy: 3 test-lane check runs succeeded on ebb7aef10a, so the input the skip depends on works live. The skipped (CI covers it) line cannot appear here, though: docs/conventions/source-control.yaml:36-38 turns the verify step off in pre_ready, and a disabled step is reported as disabled, never as skipped by CI (ready-for-review.md, 2.5.6). Observing the line takes a repository with verify on, or a PR on a non-default base, where the pre_ready list is not honored. Lane 3 releases this claim and leaves the issue open.

  14. kyle-sexton commented on Nov 3, 2026

    @kyle-sexton
    MemberAuthor

    Probe result: criterion met on ci-workflows#972, a repository with verify on (read-pre-ready.sh printed source default, verify enabled).

    • Head e50d559, 0 commits behind main. The PR went back to draft at 04:02:06Z and ready at 04:02:39Z, 33 seconds later. The ready run 41575752483 started at 04:02:43Z.
    • Venue line on the draft's head: tests ci healthy: 1 test-lane check runs succeeded on e50d5598ce82095071b46019300eb631826e7394 (test-contracts). A draft that had its lanes run before going back to draft therefore keeps the CI result.
    • Verify step: suites skipped (CI covers it), naming test-contracts. The non-test part ran (actionlint on the 3 changed workflows, exit 0) and verify-gate.sh record stored pass; verify-gate.sh check then allowed the flip.

    Limit: this session ran ready's steps by hand from reference/ready-for-review.md, using its scripts, not through a single command. The reference's line "a draft's head always reads local" holds only for a draft that has never gone through a ready run; this probe's draft had. The first flip of the same PR ran the suites locally in 23 seconds (03:51:16Z to 03:51:39Z).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    host-runNeeds an unattended run on a host; the host runner takes it, not a person or the issue lane.status: confirmedA maintainer confirmed the item is real and in scope; ready to be worked.

    Type

    Fields

    No fields configured for Task.

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions