nx-ci-monitor

Run the ci-monitor command in an Nx workspace connected to Nx Cloud. It spawns a ci-watcher subagent that polls CIPE status, reports state changes, and triggers default actions like applying self-healing fixes until the pipeline succeeds or a timeout is reached.

When a CIPE fails, Nx Cloud self-healing generates a suggested fix. The monitor can apply it via the MCP update_self_healing_fix tool, apply it locally with nx apply-locally for enhancement, or reject it and fix from scratch, then push and wait for a new CIPE.

No. The Skill checks nx.json for nxCloudId or nxCloudAccessToken before starting and exits immediately if neither exists. It is designed specifically for Nx Cloud CIPEs, not local task execution or generic CI debugging.

The subagent returns a no_new_cipe status after the new-CIPE timeout (default 10 minutes). If --auto-fix-workflow is enabled, the monitor attempts a lockfile update and pushes it; otherwise it exits with guidance to check the CI provider for pre-Nx failures.

Yes. User instructions passed as arguments override defaults, such as never auto-applying fixes, prompting before each git push, auto-rejecting e2e task fixes, or requiring a minimum confidence threshold before applying.

Do not use it for running local Nx tasks (use nx-run-tasks instead) or for debugging CI systems outside Nx Cloud. It only orchestrates monitoring and self-healing for Nx Cloud-connected workspaces.