[aw] Auto-Triage Issues failed

[github-actions]

Workflow Failure

Workflow: Auto-Triage Issues
Branch: main
Run: https://github.com/github/gh-aw/actions/runs/25878139409

⚠️ No Safe Outputs Generated: The agent job succeeded but did not produce any safe outputs.

This typically indicates:

• The safe output server failed to run
• The prompt failed to generate any meaningful result
• The agent should have called noop to explicitly indicate no action was taken

Action Required

Assign this issue to Copilot using the agentic-workflows sub-agent to automatically debug and fix the workflow failure.

Debug with any coding agent

Use this prompt with any coding agent (GitHub Copilot, Claude, Gemini, etc.):

Debug the agentic workflow failure using https://raw.githubusercontent.com/github/gh-aw/main/debug.md

The failed workflow run is at https://github.com/github/gh-aw/actions/runs/25878139409

Manually invoke the agent

Debug this workflow failure using your favorite Agent CLI and the agentic-workflows prompt.

• Start your agent
• Load the agentic-workflows prompt from .github/agents/agentic-workflows.agent.md or https://github.com/github/gh-aw/blob/main/.github/agents/agentic-workflows.agent.md
• Type debug the agentic workflow auto-triage-issues failure in https://github.com/github/gh-aw/actions/runs/25878139409

Tip

Stop reporting this workflow as a failure

To stop a workflow from creating failure issues, set report-failure-as-issue: false in its frontmatter:

safe-outputs:
  report-failure-as-issue: false

Generated from Auto-Triage Issues · ◷

• expires on May 15, 2026, 6:36 AM UTC

---

Auto-closed by [aw] Failure Investigator (6h)

This tracking issue is now stale: the underlying failure was caused by the wazero compilation cache permissions bug already fixed by #32193 (commit 1797426, merged 2026-05-14 at 19:13:34Z), which bumped mcpg to v0.3.9 and relocates the wazero cache out of /tmp/gh-aw/mcp-logs/.

Evidence

Failure pattern from run [§25878139409](https://github.com/github/gh-aw/actions/runs/25878139409)

agent: ##[error]EACCES: permission denied, scandir '/tmp/gh-aw/mcp-logs/wazero-cache'
agent/Upload agent artifacts: ##[error]EACCES: permission denied, scandir '/tmp/gh-aw/mcp-logs/wazero-cache'
conclusion/Download agent output artifact: ##[error]Unable to download artifact(s): Artifact not found for name: agent
detection/Setup threat detection: ##[error]❌ Prompt file not found at: /tmp/gh-aw/threat-detection/aw-prompts/prompt.txt

This matches the symptom described in #32192 (the issue PR #32193 resolves).

Cross-cluster confirmation

The same root cause produced 9 agentic workflow failures across 7 distinct workflows in the 6h window (all engines affected — Claude, Copilot, Gemini), all pre-fix. Closing this issue rather than spawning a new tracker because the platform fix is already shipped.

If Auto-Triage Issues continues to fail on runs created after 2026-05-14 19:13:34Z, a new tracking issue should be opened — those failures would indicate a separate regression.

Generated by [aw] Failure Investigator (6h) · ● 13.4M · ◷

[github-actions]

Agent job 25879784811 failed.

⚠️ No Safe Outputs Generated: The agent job succeeded but did not produce any safe outputs.

This typically indicates:

• The safe output server failed to run
• The prompt failed to generate any meaningful result
• The agent should have called noop to explicitly indicate no action was taken

Generated from Auto-Triage Issues · ◷