Forge-Rollback-Planning

Define observable rollback triggers first, then choose a revert or roll-forward path per change and record the reasoning. Rehearse the path before deployment and document it in the runbook rollback section.

Revert restores the previous version of the change, while roll-forward fixes the issue by deploying a new change. The Skill requires recording which path applies per change and why, decided before deployment rather than during an incident.

Forward-only migrations need a compatibility strategy rather than an assumed reversal. Record the reversibility decision explicitly and align it with backup and recovery objectives from Forge-Backup-and-Recovery.

Do not use it during a live incident; Forge-Incident-Management executes the plan under incident coordination. It also does not define backups or verify deployment success, which belong to other Skills.

No. Backup and restore capability is defined separately by Forge-Backup-and-Recovery, and its recovery objectives bound what the rollback plan can promise. Rollback planning consumes those objectives as an input.