| name | ci-status-monitor |
| description | Monitor CI conservatively with snapshot-first status checks and actionable failure triage. Use when tracking pipeline runs, investigating build failures, or validating PR readiness. |
| argument-hint | Branch/commit and monitoring strictness |
| user-invocable | true |
| disable-model-invocation | false |
CI Status Monitor
Inherits from task-core-loop.
Use this skill to monitor CI with clear, conservative status handling.
Policy
- Default to conservative polling.
- Prefer snapshot checks over continuous watch.
- Escalate polling frequency only for active failures or user-requested live monitoring.
Procedure Additions
- Apply
task-core-loop evidence and output discipline.
- Capture latest CI snapshot via
gh run list --branch <branch> or gh run view for a specific run.
- Classify state: queued, in_progress, success, failure, stale.
- For queued/in_progress: report once and continue local work.
- For failure: pull failing job details and isolate first actionable failure.
Staleness Heuristic
- Treat long-running in_progress runs as stale only after a conservative threshold.
- Do not spam repeated status calls unless there is a decision to make.
Output
- Current CI state table.
- Recommended next action.
- Confidence and unresolved risks.
