Operational Excellence

Self-hosted automation for reliable CI recovery.

CI Autopilot provides a hardened runner, disciplined workflows, and a clear operational playbook so incidents are resolved quickly and safely.

Runner setup Operations runbook
// Quick verification
gh auth status

$serviceName = Get-Content .\.service
Get-Service -Name $serviceName
// Runner health
gh api repos/Coding-Autopilot-System/ci-autopilot/actions/runners \
  -q '.runners[] | {name, status, online}'
Architecture

Designed for deterministic recovery

Runner host

  • Windows service for consistent execution.
  • Separation between orchestration and execution.
  • Local logs for host-level diagnostics.

Workflow layer

  • Manual dispatch, schedules, and issue triggers.
  • Explicit approvals and audit trails.
  • Safe rollbacks and idempotent steps.

Agent runtime

  • Python execution with reproducible environments.
  • Structured logging and artifacts.
  • Designed for reliable remediation.
Operations

Runbooks built for busy teams

cd C:\actions-runner
$serviceName = Get-Content .\.service
Restart-Service -Name $serviceName

See docs/operations.md for upgrade, rotation, and health checks.

Security

Least-privilege by default

Token scopes

  • Required: repo, workflow, read:org.
  • Optional: admin:org based on policy.
  • Always fetch short-lived tokens.

Auditability

  • GitHub Actions logs are authoritative.
  • Host logs retained for forensics.
  • Explicit approvals for sensitive tasks.
Reference

Quick links

Architecture Runner setup Operations Security Troubleshooting