What problem does it solve?
Developers often need to try risky changes or refactorings without fear of breaking their work or losing progress. This skill provides a robust safety net, allowing for experimentation with automatic rollback to a stable state.
Core Features & Use Cases
- Atomic Restore Points: Create full snapshots of your working tree and staging area before making any changes.
- Safe Experimentation: Try out complex refactorings or new features, knowing you can revert to a stable state with a single command.
- Verified Destructive Operations: Clear your working tree with confidence, as the system verifies the state before proceeding, preventing accidental data loss.
- Use Case: Before an AI agent attempts a major refactoring, it creates a checkpoint. If the refactoring introduces critical bugs, the agent can automatically restore the previous stable state, saving hours of manual debugging and ensuring "AI works, you rest."
Quick Start
Create a checkpoint before making risky changes:
CHECKPOINT_HASH=$(scripts/create.sh "my-feature-experiment")
If changes fail, restore it:
scripts/restore.sh "$CHECKPOINT_HASH"