You will apply a repeatable seven-step triage flow whenever a run produces unexpected output.
- Open the live log in Actions and scan for
[error]lines first. - Check
[plan]density. More than four consecutive plan lines without a tool call usually means your brief is underspecified. - Inspect
[tool]and[result]lines to confirm expected data is returned. - Look for
safe-output: BLOCKEDand decide whether to increasemaxor tighten "post once" guidance. - Read the run summary and compare it to your expected write behavior.
- Open the safe-output record in the job details and treat it as source of truth for writes.
- If behavior is still unclear, ask the
agentic-workflowsskill to diagnose your workflow with a pasted snippet.
- I can run this checklist in order without skipping steps
- I know where to find both live logs and safe-output records
- I can decide whether the root cause is prompt, data, permissions, or output limits
- I can gather a minimal log snippet to share for deeper diagnosis