ci
Skill1 starsChanged 21 days agoMITView on GitHub
Triage the current branch's CI. Find the relevant current-head run, wait if pending, diagnose failures without flooding context, compare against main when appropriate, and report or fix. In fix mode, push and re-watch until the new current HEAD is green. Never merge.
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 22
Use sy:sweep for large failing logs.
- Very specific: line 24
Before calling something pre-existing, verify the corresponding latest main job when relevant.
- Very specific: line 40
Every subagent dispatch resolves its model and effort from config and passes them as the Agent invocation's actual model and effort …
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Get the current branch's CI green. $ARGUMENTS: fix applies fixes; explicit PR/run overrides auto-detect.
This skill inherits caller model/effort. Raw logs never enter the calling context.
Find and watch
- Detect branch/PR and current HEAD SHA.
- Find checks/runs that actually cover that SHA; do not mistake an older green run for current-head coverage.
- Pending ⇒ wait with the shared poller: launch
${CLAUDE_PLUGIN_ROOT}/scripts/ci_poll.sh poll <pr> --repo <the PR's base repository, never a fork's own origin>as a single token-freerun_in_backgroundBash call (it sleeps between checks and exits when nothing is pending); add--head <SHA>whenever triage already knows the expected head, so a stale or empty check set cannot read as green. Never hand-write a poller loop. Never poll once per reasoning turn and never self-resume a monitor at a turn-budget boundary; both bleed tokens on a large matrix. Its default 1800s timeout comes fromci.poll_timeout— raise it for repos/matrices whose CI routinely runs longer, so a single call spans the wait instead of timing out and bouncing to the parent.
Diagnose failure
Use sy:sweep for large failing logs; it returns failing jobs/tests, a few decisive root-cause lines, and implicated code pointers. If the failure surface is tiny, read the small error region directly. Never page full logs.
Before calling something pre-existing, verify the corresponding latest main job when relevant. Distinguish branch bug, inherited main breakage, flake, resource limit, dependency drift, and infra/auth/network failure.
Report or fix
- Report: failing job, real error, cause, concrete fix.
fix: apply, run local reproducer/checks, push, and re-watch the new HEAD until green.
Return status plus the SHA covered:
CI status: green | fixably-red | blocked
CI_GREEN_SHA: <sha or none>Never merge.
Every subagent dispatch resolves its model and effort from config and passes them as the Agent invocation's actual model and effort parameters, per ${CLAUDE_PLUGIN_ROOT}/skills/shared/references/model-dispatch.md; a nested dispatch inherits nothing and must resolve again.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

