YggShpigford/chops
↗ GitHubYour AI agent skills, finally organized. A macOS app to browse, edit, and manage skills across Claude Code, Cursor, Codex, Windsurf, and Amp.
1,126
Stars
60
Forks
3
Watchers
8
Open Issues
Safety Rating A
The repository appears to be a legitimate open source macOS utility. The only notable security-adjacent design choice is the deliberate disabling of the macOS app sandbox to allow broad filesystem reads across ~/. This is clearly documented in the README and entitlements file as an intentional, required architectural decision rather than a red flag. No hardcoded secrets, malicious patterns, suspicious dependencies, or prompt injection attempts were found.
ℹAI-assisted review, not a professional security audit.
AI Analysis
Chops is a native macOS desktop application (SwiftUI + SwiftData) that provides a unified interface for discovering, browsing, editing, and organizing AI coding agent skills and agent definitions across multiple tools including Claude Code, Cursor, Codex, Windsurf, Copilot, Aider, and Amp. It scans tool-specific dotfile directories for skill and agent markdown files, provides a built-in editor with frontmatter parsing, supports collections for organization, real-time file watching via FSEvents, full-text search, and remote skill server connectivity.
Use Cases
- Centralized management of AI coding agent skills and prompts stored across multiple tools and dotfile directories
- Browsing, editing, and creating new skills or agent definitions without manually navigating dotfiles
- Organizing skills into collections and searching across all installed tools from a single UI
- Watching for real-time filesystem changes to skill files and syncing updates instantly
- Discovering and installing skills from remote servers such as OpenClaw
Tags
Security Findings (2)
No hardcoded API keys, tokens, or secrets were found in the repository content provided.
No obfuscated code, backdoors, or data exfiltration patterns detected. The intentional disabling of the macOS sandbox (noted explicitly in Chops.entitlements and the README) is by design to enable broad filesystem access to dotfile directories, which is the core use case. This is disclosed transparently.
Project Connections
Claude-Code-Game-Studios
→Claude Code Game Studios generates large numbers of specialized skills and agent definitions for Claude Code; Chops provides the native macOS UI to browse, organize, and edit exactly those skill files across tools.
code-review-graph
→code-review-graph exposes codebase intelligence via MCP to AI coding agents (Claude Code, Cursor, Windsurf), while Chops manages the skills and agent configurations that govern how those same tools behave — the two tools serve the same user persona from complementary angles.
GitNexus
→GitNexus provides MCP-based code intelligence to AI coding agents like Cursor and Claude Code; Chops manages the skill and agent configuration files that control those same tools, making them natural companions in an AI-assisted development workflow.
clawvault
→ClawVault is a CLI-based local-first memory and skills management system for AI agents; Chops addresses a similar problem of organizing AI agent context/skills but does so through a native macOS GUI rather than a CLI, and focuses specifically on coding tools.
kavach
→Kavach monitors and constrains AI agent behavior at the OS level for security; Chops manages the skill definitions that define agent behavior. Together they represent governance (Kavach) and configuration management (Chops) layers for AI coding agents running locally.