Repository navigation
[workflow-health] PR Triage Agent: push_repo_memory job failing (post-cleanup error) #23109
Description
Problem
PR Triage Agent started failing on 2026-03-26 with 2 consecutive failures in the push_repo_memory job.
- Run #289 — failed 2026-03-26T06:21
- Run #288 — failed 2026-03-26T00:12
- Run #287 — ✅ success 2026-03-25T18:16
Root Cause
The push_repo_memory job switches the workspace branch to memory/pr-triage to commit and push memory files. After the push succeeds, the GitHub Actions post-job cleanup for the ./actions/setup local action fails because the workspace is now on the memory/pr-triage branch (which doesn't have the actions/ directory).
Error in post-cleanup:
##[error]Can't find 'action.yml', 'action.yaml' or 'Dockerfile' under
'/home/runner/work/gh-aw/gh-aw/actions/setup'. Did you forget to run
actions/checkout before running your local action?
Important: The actual memory push succeeds — data is correctly written to memory/pr-triage branch. The failure only affects the post-job cleanup phase.
Evidence
From run #289 logs (push_repo_memory job):
- ✅ Checked out branch
memory/pr-triage - ✅ Committed and pushed 3 files (
pr-triage-latest.json,pr-triage-summary.md,README.md) - ❌ Post-job cleanup fails: can't find
actions/setup
Impact
- PR Triage Agent reports are still being generated (the
agentjob succeeds) - Repo memory is being saved successfully
- But the overall workflow run is marked as
failuredue to the cleanup error
Suggested Fix
Options:
- Add a checkout step at the end of
push_repo_memoryto restore the workspace to the main branch before cleanup - Use
persist-credentials: false+ sparse checkout on main as the last step to restoreactions/setup - Set
continue-on-error: trueon steps that might leave the workspace on the memory branch
The fix should be implemented in the workflow compiler's push_repo_memory job generation, so all workflows that use repo-memory benefit from the fix.
Priority
P1 — New regression, 2 consecutive failures starting 2026-03-26T00:12
Detected by §23593331592 — Workflow Health Manager
Note
🔒 Integrity filter blocked 3 items
The following items were blocked because they don't meet the GitHub integrity level.
- #21703
search_issues: has lower integrity than agent requires. The agent cannot read data with integrity below "approved". - #18407
search_issues: has lower integrity than agent requires. The agent cannot read data with integrity below "approved". - #21784
search_issues: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".
To allow these resources, lower min-integrity in your GitHub frontmatter:
tools:
github:
min-integrity: approved # merged | approved | unapproved | noneGenerated by Workflow Health Manager - Meta-Orchestrator · ◷
- expires on Mar 27, 2026, 12:14 PM UTC
Activity
github-actions commentedon Mar 26, 2026
🍪 Issue Monster has assigned this to Copilot!
I've identified this issue as a good candidate for automated resolution and assigned it to the Copilot coding agent.
The Copilot coding agent will analyze the issue and create a pull request with the fix.
Om nom nom! 🍪
Copilot commentedon Mar 26, 2026
The agent encountered an error and was unable to start working on this issue: This may be caused by a repository ruleset violation. See granting bypass permissions for the agent, or please contact support if the issue persists.
github-actions commentedon Mar 27, 2026
Status Update — 2026-03-27T12:03Z | §23645263363
Escalation: 6 consecutive failures (was 2 when issue was filed)
Pattern Confirmed Systemic
The
push_repo_memory → Post Setup Scriptsfailure is confirmed to affect multiple workflows:push_repo_memorypush_repo_memory → Post Setup ScriptsfailureRoot cause remains: after pushing to the
memory/*branch, the workspace is left on that branch. When GitHub Actions runs post-job cleanup for./actions/setup, it can't findaction.ymlbecause the memory branch doesn't have theactions/directory.Auto-Fix Status
Issue Monster assigned this to Copilot on 2026-03-26T12:34, but Copilot's agent reported:
The fix requires adding a checkout-restore step to the
push_repo_memoryjob generator inpkg/workflow/compiler_yaml.go. Manual intervention needed.Recommended Fix
In the compiled
push_repo_memoryjob, add a final step to restore the workspace branch tomainbefore cleanup runs:This should be implemented in the workflow compiler so all workflows with repo-memory benefit automatically.
github-actions commentedon Mar 27, 2026
This issue was automatically closed because it expired on 2026-03-27T12:14:04.829Z.