agentjj

Automate agent-oriented version control with structured JSON outputs.

6|Updated Feb 1, 2026

Search Tags:#checkpoints #undo #version-control #AI-agents #bulk-operations #JSON-output #agentjj

Ready to run

npx skills add https://github.com/2389-research/agentjj --skill agentjj

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: agentjj
Source: https://github.com/2389-research/agentjj/tree/main/docs
Command: npx skills add https://github.com/2389-research/agentjj --skill agentjj

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Automates agent-oriented version control for AI agents.

Core Features & Use Cases

JSON-first outputs: all interactions are machine-friendly and parseable.
Orient, status, and changes: complete repo briefing, current state, and recommended actions.
Safety & recovery: checkpoints and undo for safe experiments.
Bulk operations: read and analyze multiple files efficiently.

Quick Start

In any Git repository, run agentjj orient to get a complete orientation. Optional: run agentjj init to create an agent manifest and start using features like checkpoints and bulk operations.

Frequently Asked Questions about agentjj

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

FAQPage Schema

How do I set up version control for AI agents to safely manage and undo code changes?▼

Automated version control for AI agents is set up by running the orientation and initialization commands in your repository. This creates an agent manifest, enabling reliable checkpoints, bulk operations, and undo functionality for safe experiments.

Can AI agents get a structured JSON briefing of repository status and recommended actions?▼

Yes, AI agents can receive a structured JSON briefing of the repository status. The tool provides machine-friendly JSON outputs for orientation, current state, and recommended actions, ensuring workflows can parse the data directly.

What is the best way to bulk read and analyze multiple files in a repository using AI workflows?▼

The best way to bulk read and analyze files in AI workflows is using an agent-oriented version control tool. It performs bulk operations efficiently to read and analyze multiple files, outputting structured results for the agent.

Do I need external dependencies or libraries to use agent-oriented version control for AI agents?▼

No external dependencies or libraries are required. The agent-oriented version control system operates entirely using the standalone agentjj tool, requiring no additional components to manage checkpoints and output JSON.

How does an AI agent create checkpoints before applying code changes in a Git repository?▼

AI agents create checkpoints by running the initialization command to generate an agent manifest. Once initialized, the version control tool provides checkpoint and undo features, allowing agents to safely experiment and revert changes.

Why use a specialized VCS for AI agents instead of standard Git commands for automated workflows?▼

A specialized VCS for AI agents provides JSON-first outputs for machine parsing, complete repo orientation, and bulk file operations. It optimizes automated workflows by offering reliable orientation and structured state data standard Git lacks.