Does it track metadata besides code changes?

Yes, it stores session context, including session numbers and token usage, to provide a complete picture of the AI's state at the time of the snapshot.

What does Checkpoint Manager do?

It creates snapshots of your code's state—including git data and feature progress—allowing you to revert to a previous state if something goes wrong during coding.

How does it handle git changes?

It captures the current commit hash and tracking of dirty files, utilizing git reset commands to restore the codebase during a rollback operation.

Can it trigger rollbacks automatically?

Yes, when integrated with error-recovery tools, it can be configured to automatically restore the last known good state if a risky operation fails.

What are the different checkpoint types available?

It supports Automatic (triggered before risky actions), Manual (user-defined), Feature (created after completion), and Session (created at start of work) checkpoints.

Checkpoint Manager

Name: Checkpoint Manager
Author: adaptationio

byadaptationio

0•

생산성 및 워크플로

Automates state snapshots and rollbacks to ensure safe experimentation and reliable recovery during autonomous coding sessions.

The Checkpoint Manager is a specialized utility for Claude Code that provides a robust safety net for autonomous development tasks. By capturing comprehensive snapshots including git states, feature progress, and session metadata, it allows developers and AI agents to experiment freely with refactors or complex logic without the risk of permanent failure. The tool supports multiple checkpoint types—from automatic snapshots before risky operations to permanent feature milestones—ensuring that the codebase can always be restored to a known good state with minimal effort.

주요 기능

01Comprehensive state snapshots including git commits, dirty files, and feature progress.

020 GitHub stars

03Automated rollbacks triggered by failure detection or manual request.

04Granular checkpoint categorization for sessions, features, and risky operations.

05Deep integration with autonomous loops for hands-free error recovery.

06Context-aware metadata tracking including session numbers and token usage.

사용 사례

01Protecting autonomous coding sessions from unrecoverable logic errors or bugs.

02Performing major code refactorings where reverting to a stable state may be necessary.

03Creating permanent progress snapshots after successfully implementing specific features.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add adaptationio/skrillz checkpoint-manager

For use in Claude.ai and ChatGPT

Download Skill