Skip to content

[agent-perf] Premature-closure regression: Stale Lock Files, PR Description Updater, Label Closed PRs reopened within days #48894

Description

@github-actions

Caution

agentic threat detected
Threat detection flagged this output in warn mode. Manual review is REQUIRED before any follow-up automation.

Details

The threat detection engine failed to produce results.

Review the workflow run logs for details.

Premature-closure regression: 3 workflow-health issues reopened within days

Severity: High (systemic, recurring)

Multiple workflow-health/AR-cluster issues were closed after apparently seeing a green run, only to have action_required resume within days:

Workflow Latest AR sample (Jul 29) Closed issue(s) Closed date Days to reopen
Stale Lock Files 2/2 AR (100%) #48131 2026-07-27 2 days
PR Description Updater 2/2 AR (100%) #44510, #44954, #47925 2026-07-25 (latest) 4 days
Label Closed PRs 2/2 AR (100%) #44663, #44954, #47925 2026-07-25 (latest) 4 days

This matches a pattern already seen with AI Moderator on 2026-07-08 (issue #43925 closed → 4/4 AR the very next day, tracked as #aw_ai_mod_jul8). This is now the 4th distinct workflow to show the same closure/reopen churn within a 3-week window.

Likely root cause: PR Description Updater and Label Closed PRs were previously linked to the same "CLI hang-on-exit" cluster as Impeccable Skills Reviewer / Matt Pocock / Test Quality Sentinel, with a shared fix in PR #44254 (per shared-alerts.md, 2026-07-08). If #44254 merged but these two workflows are still at 100% AR three weeks later, either:

  1. The fix did not fully address these two workflows, or
  2. A regression reintroduced the hang, or
  3. Issues were closed based on a single transient success rather than a sustained-success window.

Recommendation:

  1. Verify PR fix: reduce post-completion idle watchdog and add cleanup timeouts to prevent Copilot CLI hang on exit #44254's merge status and confirm it actually covers Label Closed PRs / PR Description Updater / Stale Lock Files' code paths.
  2. Establish a 3+ consecutive successful runs bar before closing any workflow-health AR-cluster issue, rather than closing on the first green run.
  3. Re-open or re-triage the three workflows above with fresh root-cause analysis if fix: reduce post-completion idle watchdog and add cleanup timeouts to prevent Copilot CLI hang on exit #44254 is confirmed merged and insufficient.

Do not re-file if this exact regression-pattern issue already exists — check for an existing [agent-perf] or [workflow-health] issue referencing this same closure/reopen cycle before creating a duplicate.

Source: Agent Performance Analyzer — Meta-Orchestrator, run §30454899649, 2026-07-29.

Generated by ⚡ Agent Performance Analyzer - Meta-Orchestrator · aut00 · 98.7 AIC · ⊞ 16K ·

  • expires on Jul 31, 2026, 5:28 AM UTC-08:00

Activity

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

Metadata

Metadata

Labels

cookieIssue Monster Loves Cookies!

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions