titan-forge

Automating phase-driven code refactor workflows involves processing targets in defined phases with per-target progress tracking, applying gauntlet-guided changes, and validating commits with gate checks before finalizing state.

Managing git staging and rollbacks in a multi-phase pipeline requires tracking the current phase, completed targets, and rollback readiness, while handling per-target diffs and commits automatically to reduce manual toil.

Gate validation works by running checks before finalizing state and progress, ensuring that applied per-target changes pass safety gauntlets and commit validation before the phase state is updated.

Tracking per-target execution progress independently is supported through robust state management, which monitors the current phase, completed targets, and rollback readiness to prevent drift across the codebase.

Titan-state management is needed to track the current phase, completed targets, and rollback readiness, preventing coordination errors and drift when applying plan-driven changes and safety checks across multiple targets.

Automated phase-driven workflows should not be used for simple, single-file edits that do not require gauntlet guidance, gate validation, or multi-target state tracking, as the overhead outweighs the coordination benefits.