rollback-manager

Automate task-level, full, and partial rollbacks to last-good-commit checkpoints.

Updated Jan 16, 2026
One-click install
npx skills add https://github.com/adiomas/claude-code-adiomas-plugins --skill rollback-manager
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: rollback-manager
Source: https://github.com/adiomas/claude-code-adiomas-plugins/tree/main/autonomous-dev/skills/rollback-manager
Command: npx skills add https://github.com/adiomas/claude-code-adiomas-plugins --skill rollback-manager

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill automates safe rollback of changes when verification fails during autonomous execution.

Core Features & Use Cases

  • Task-level rollback: revert changes from a single failed task.
  • Full rollback: revert all changes to the last checkpoint.
  • Partial rollback: selectively revert specific files while preserving others.
  • Checkpoint and state management: record and reference last-good-commit and tasks/state for deterministic recovery.

Quick Start

To execute immediately, tell the AI: rollback to the last good state after a verification failure.

  • Example: Rollback to the last good commit after verification failure.

Frequently Asked Questions about rollback-manager

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate git rollback when verification fails during an autonomous task?

To automate git rollback when verification fails, you can trigger a task-level or full rollback to the last-good-commit. This skill handles the automatic recovery by checking version-controlled history and reverting changes safely.

Can I do a partial rollback of specific files while preserving other changes?

Yes, partial rollback allows you to selectively revert specific files while preserving other changes. It uses a structured state/config file to drive conditional rollbacks and update task statuses during recovery.

What is checkpoint-based recovery and how does it manage state?

Checkpoint-based recovery records and references a last-good-commit alongside task states. It requires a structured state/config file to drive deterministic recovery and status updates when verification failure occurs.

Do I need a specific git history setup to enable automatic recovery?

Yes, automatic recovery requires a version-controlled history with established last-good-commit checkpoints. You must maintain this structured state alongside your codebase to drive conditional rollbacks effectively.

What is the best way to revert all changes to the last checkpoint after a failed task?

The best way to revert all changes to the last checkpoint is executing a full rollback. This approach uses your version-controlled history to safely restore the entire project to its last-good-commit state.

When should I use a task-level rollback instead of a full rollback?

Use a task-level rollback to revert changes from a single failed task, rather than a full rollback which reverts all changes to the last checkpoint. Choose based on whether the failure is isolated or system-wide.