long-running-job-monitor

Monitors long-running test suites, builds, installations, and deployments, including a run the agent starts itself with no user request, using timestamped logs, heartbeats, terminal markers, out-of-band probes, and WORKING/STALLED/DONE/FAILED classification. Survives control-channel drops; a restart starts a new readiness epoch. USE FOR: run the test suite, full suite, Pester run, Invoke-Pester, build.ps1, test.ps1, verification run, regression run, live test, live proof, proof run, integration test, deployment, installation, long-running job, is it stuck, still running, watch deployment, remote job, SSH, WinRM, keep me posted, answering while a job runs. DO NOT USE FOR: diagnosing why a build or test failed (use sampler-build-debug), writing test code (use pester-patterns), AutomatedLab Proxmox provisioning (use automatedlab-proxmox), WinRM connectivity diagnosis (use winrm-troubleshooting), short commands, skill authoring (use skill-creator), agent evals (use agent-evals).