[aw] PR Description Updater — 95% failure rate since Jul 1 (19/20 AR)

[github-actions]

Workflow Health Issue

Workflow: PR Description Updater (pr-description-caveman)
Trigger: pull_request: [closed] (when PR is merged)
Priority: P1
Run: https://github.com/github/gh-aw/actions/runs/28997022503

Failure Pattern

• Failure rate: 19/20 runs AR (95%) since July 1, 2026
• Status: Consistently action_required — only 1 success in 20 runs
• Engine: Copilot (uses sub-agents)
• Trigger pattern: Fires on every merged PR

Impact

This workflow rewrites merged PR descriptions with structured summaries optimised for downstream agentic analysis. With 95% failure rate, PR descriptions are not being updated, degrading downstream agentic context quality.

Observed Failure

All recent runs (Jul 4–9) result in action_required. The workflow uses strict: true with sub-agents for chunked diff processing. Possible causes:

• Safe-output validation failure (produces no safe outputs)
• Sub-agent model resolution issue (similar to other Copilot sub-agent failures)
• strict: true causing false-positive failures when PR has no description to update
• Token/timeout issues with large diffs

Example Runs

• https://github.com/github/gh-aw/actions/runs/28996995676 — AR (Jul 9)
• https://github.com/github/gh-aw/actions/runs/28995947747 — AR (Jul 9)
• https://github.com/github/gh-aw/actions/runs/28995329833 — AR (Jul 9)

Suggested Investigation

1. Check if safe-output update-pull-request is being called before workflow exits
2. Review sub-agent model resolution (may be hitting same Codex 404 pattern)
3. Consider: is strict: true appropriate here if some PRs legitimately need no update?
4. Check if the PR diff chunking logic is causing issues with small/auto-generated PRs

Related Issues

• [aw_pr_cq_tqs] PR Code Quality Reviewer — CLI hang-on-exit
• [aw] Daily Sub-Agent Model Resolution Audit failed #43335 Sub-agent model resolution
• [aw] PR Sous Chef is missing required tool #43143 PR Sous Chef — missing pr-processor

Generated by 🏥 Workflow Health Manager - Meta-Orchestrator · 86.5 AIC · ⌖ 11.8 AIC · ⊞ 7.7K · ◷

• expires on Jul 9, 2026, 9:59 PM UTC-08:00

[github-actions]

Status Update — 2026-07-10

Still failing: 8/10 most recent runs are action_required. Pattern persists ~10 days. No fix landed yet.

Latest runs (Jul 10):

• 05:32 UTC: success ✅
• 05:32 UTC: action_required ❌
• 05:31 UTC: action_required ❌
• 05:20 UTC: skipped
• 04:49–04:55 UTC: 6× action_required ❌

Priority: P1 — unresolved. Recommend investigating safe-output configuration or engine timeout.

Generated by 🏥 Workflow Health Manager - Meta-Orchestrator · 62.6 AIC · ⌖ 6.94 AIC · ⊞ 7.7K · ◷

[github-actions]

🍪 Issue Monster selected this for Copilot

I've identified this issue as a good candidate for automated resolution and requested assignment to the Copilot coding agent.

If assignment succeeds, the Copilot coding agent will analyze the issue and create a pull request with the fix.

Om nom nom! 🍪

🍪 Om nom nom by Issue Monster · 2.54 AIC · ⌖ 6.2 AIC · ⊞ 8.3K · ◷

[github-actions]

This issue was automatically closed because it expired on 2026-07-10T05:59:44.228Z.

Closed by Workflow