What problem does it solve?
This Skill prevents chaotic incident-driven recovery by helping release owners pre-plan how to rollback or forward-recover every changed surface—code, configuration, database, feature flags, caches, jobs, and external integrations—using measurable triggers and a validated runbook.
Core Features & Use Cases
- Per-surface rollback planning: Defines independent rollback methods for code, schema, config, flags, caches, background jobs, and external integrations instead of assuming a single “redeploy previous version” is sufficient.
- Database rollback strategy: Specifies reversible migrations with tested DOWN scripts or, when not reversible, requires an explicit forward-fix plan reviewed before production deployment.
- Mixed-version compatibility sequencing: Ensures expand-contract ordering so the deployment avoids the unsafe state (new code + old schema) during rolling and canary/blue-green phases.
- Measurable rollback triggers: Converts rollback decisions into thresholds tied to monitoring metrics, including trigger duration and measurement tool expectations.
Quick Start
Use the release-rollback skill to generate a complete rollback runbook for your next production release, including changed-surface actions, rollback triggers, decision owner, mixed-version sequencing, and post-rollback validation.
