Skip to content

[Fleet Monitor] petry-projects/TalkTerm β€” .github/workflows/pr-auto-review.ymlΒ #375

Closed

Description

@don-petry

🟑 WARNING: petry-projects/TalkTerm / .github/workflows/pr-auto-review.yml

Metric Value
Failure Rate 14.3% (1 / 7 runs)
Status WARNING
p50 Duration 996s
p95 Duration 1652s
Successful Runs 6
Cancelled Runs 0

Workflow: .github/workflows/pr-auto-review.yml

Last updated by Fleet Monitor run on 2026-07-18.
Threshold: failure rate > 10% over the monitored window.

Activity

  1. added
    dev-leadFor dev-lead agent pickup
    fleet-trackerTracked by the Actions Fleet Monitor
    health-checkAutomated health check result
    on Jul 18, 2026
  2. don-petry commented on Jul 18, 2026

    @don-petry
    ContributorAuthor

    Dev-Lead Implementation Plan

    Issue: #375 β€” [Fleet Monitor] petry-projects/TalkTerm β€” .github/workflows/pr-auto-review.yml

    Root-cause investigation

    The single counted "failure" over the monitored window is run 29587132848 β€” conclusion action_required, event pull_request_review, triggering actor Copilot (a bot). action_required means GitHub gated the run pending manual approval because it was initiated by a bot/App and consumes an org secret; the job never executed. This is inherent GitHub Actions approval-gating, not a defect in the workflow file, and it is not fixable from a thin caller stub (triggers/permissions/concurrency are centrally owned β€” see standards/ci-standards.md).

    While investigating I found a genuine, error-severity compliance drift on this exact file that the org compliance audit (check_stub_surface_drift) flags:

    Scope

    Re-sync the pr-auto-review.yml caller stub's permissions: surface to the canonical source of truth by removing the drifted, unused statuses: read grant.

    Implementation Checklist

    • .github/workflows/pr-auto-review.yml β€” remove the extra statuses: read permission line so the permissions: surface matches standards/workflows/pr-auto-review.yml

    Tests to Write

    • No unit-test surface exists for a thin YAML caller stub in this repo (no package.json). Per the Phase 2 exception this is a behavior-preserving configuration re-sync (the removed permission is unused at runtime), so TDD is skipped. Verification is done via the repo lint script, YAML-parse validation, and a diff of the permissions: surface against canonical.

    Test / Verification Command

    bash .dev-lead/scripts/dev-lead-lint.sh + python3 -c "import yaml,sys; yaml.safe_load(open('.github/workflows/pr-auto-review.yml'))"

  3. don-petry commented on Jul 18, 2026

    @don-petry
    ContributorAuthor

    Dev-Lead: Implementation Complete

    Plan Execution

    • Investigated the single counted failure (run 29587132848)
    • Identified genuine compliance drift on permissions: surface
    • Removed the drifted, unused statuses: read grant to re-sync to canonical
    • Verified YAML validity, permissions-surface match, and repo lint

    Root-cause of the 14.3% failure rate (honest assessment)

    The one "failure" in the window β€” run 29587132848 β€” concluded action_required, event pull_request_review, triggering actor Copilot (a bot). action_required is GitHub gating the run pending manual approval because it was initiated by a bot/App and consumes an org secret; the job never executed. This is inherent GitHub Actions approval-gating behavior, not a defect in the workflow file, and it cannot be fixed from a thin caller stub β€” on:/permissions:/concurrency: are centrally owned (see standards/ci-standards.md). The proper levers live outside this file (repo Actions "require approval" policy, or the central reusable), so no repo-level caller-stub edit changes that data point.

    What was actually fixed

    While investigating I found an error-severity compliance drift on this exact file that check_stub_surface_drift in the org compliance audit would flag: the permissions: block granted an extra statuses: read (commented "required by pr-auto-review-reusable.yml since #435"). This is not in the canonical standards/workflows/pr-auto-review.yml; the audit notes statuses: read is a dev-lead-specific remediation (#402) that was cargo-culted here. The reusable never uses GITHUB_TOKEN's statuses scope β€” every gh call uses the GH_PAT_WORKFLOWS PAT and reads checks via gh pr checks β€” so the grant was unused and safe to remove. Removing it re-syncs the stub to the source of truth.

    Verification Results

    === YAML validity ===
    YAML OK
    === permissions surface diff vs canonical ===
    LOCAL job-perms: ['permissions:', 'pull-requests: read', 'checks: read', 'actions: read']
    CANON job-perms: ['permissions:', 'pull-requests: read', 'checks: read', 'actions: read']
    MATCH: True
    === .dev-lead/scripts/dev-lead-lint.sh ===
      [lint] shellcheck on 3 file(s)...
      [lint] validating CODEOWNERS...
    OK: CODEOWNERS
      [lint] all checks passed
    

    No package.json exists in this repo (planning/automation phase), so there is no unit-test surface for a YAML caller stub. Per the Phase 2 exception this is a behavior-preserving configuration re-sync β€” the removed permission is unused at runtime β€” so TDD was skipped and verification was done via YAML parse, canonical-surface diff, and the repo lint script.

    Files Changed

    • .github/workflows/pr-auto-review.yml: removed the drifted, unused statuses: read line from the job permissions: block so the surface matches the canonical stub. on:/concurrency: surfaces and the @pr-auto-review/v1-ring1 ring-tier pin were already compliant and left unchanged.

    Notes

    • The recorded action_required failure is inherent GitHub bot-approval gating and is not resolvable in this caller stub; flagged here so the Fleet Monitor signal isn't misread as a code defect. If the spurious signal must be eliminated, that belongs in repo Actions approval settings or the central reusable, not this file.
    • The file lacks a trailing newline; this pre-existed and was left untouched to keep the change minimal and scoped.
  4. added a commit that references this issue on Jul 18, 2026
  5. added 4 commits that reference this issue on Aug 31, 2026
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

    dev-leadFor dev-lead agent pickupfleet-trackerTracked by the Actions Fleet Monitorhealth-checkAutomated health check result

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions