Incident Response Runbook

This runbook covers high-priority operational incidents for joy release consumers.

Severity Guidelines

• SEV-1: active supply-chain risk (malicious dependency, compromised release artifact)
• SEV-2: broken release affecting installs/builds broadly
• SEV-3: localized issue with workaround

Scenario A: Compromised Dependency

1. Freeze releases immediately.
2. Identify affected versions and package IDs.
3. Remove/disable affected dependency source in registry metadata.
4. Publish a patched release and changelog/security advisory.
5. Notify users with explicit upgrade/rollback instructions.

Scenario B: Bad Release Artifact

1. Mark the release as yanked in release notes.
2. Repoint package-manager channels (Homebrew/Scoop metadata) to the prior safe release.
3. Publish fixed artifacts under a new version tag.
4. Verify checksums/signatures and smoke test install paths.

Rollback Drill (Per Release)

Perform a rollback drill once per release cycle:

1. Simulate a bad release in a staging repository.
2. Execute package-manager metadata rollback.
3. Confirm install commands resolve to the prior known-good version.
4. Record elapsed time and gaps in the release notes archive.

Communication Checklist

• Open internal incident tracker entry.
• Publish public status update within 24 hours for SEV-1/SEV-2.
• Update SECURITY.md if policy/process changes were required.