Patch
Enables AI assistants to securely apply unified diff patches to files, offering comprehensive security validation and error recovery workflows.
About
This tool operates as a Model Context Protocol (MCP) server, designed to empower AI assistants with the capability to manage file changes safely and reliably. It facilitates the application, validation, reversal, and generation of unified diff patches, supporting complex multi-hunk and multi-file modifications atomically. With robust security features such as symlink and binary file protection, disk space validation, and automatic rollback mechanisms, it ensures that all file operations are secure, reversible, and maintain system integrity, significantly enhancing AI's capacity for automated code modification and refactoring tasks.
Key Features
- Create and restore automatic, timestamped backups
- Handle multi-file and multi-hunk changes atomically
- Apply code changes using standard unified diff format
- Revert changes safely if something goes wrong
- 0 GitHub stars
- Validate patches before applying them with dry-run mode
Use Cases
- AI assistant performing targeted code modifications in files
- Automated error recovery and rollback for AI-driven file modifications
- AI assistant refactoring code across multiple files with atomic changes
