Chops

Chops is a native macOS application built by Sabotage Media LLC (Josh Pigford) that centralizes the management of AI coding agent skills and agents across multiple tools. It is free and source-available under the FSL-1.1-MIT license, with the source code hosted on GitHub. The app requires macOS 15 (Sequoia) or later and is distributed as a downloadable .dmg.

What It Is

Chops solves a specific pain point for developers who use multiple AI coding assistants: skills and agent configuration files are scattered across different dotfile directories for each tool. Instead of manually digging through ~/.claude/, ~/.cursor/, ~/.codex/, and similar paths, Chops provides a single three-column native UI to discover, browse, edit, and organize all of them in one place. It supports Claude Code, Cursor, Codex, Windsurf, Amp, Copilot, and Aider.

How It Works

Chops is built with SwiftUI and SwiftData — no web views, no Electron. The app lifecycle works as follows:

• SkillScanner probes each supported tool's known filesystem directories on launch
• FileWatcher uses FSEvents to detect disk changes and re-scans automatically in real time
• SkillParser dispatches to either a YAML frontmatter parser (for .md files) or an MDC parser (for Cursor .mdc files)
• Skills are deduplicated via resolved symlink paths, so a file symlinked into multiple tool directories appears once with multiple tool badges
• The app intentionally disables the macOS sandbox to allow unrestricted filesystem access to dotfiles across ~/

The three-column layout mirrors macOS conventions: a sidebar for tool filters and collections, a list of filtered/searched skills, and a detail pane with a monospaced editor backed by NSTextView.

Key Features

• Multi-tool support — scans Claude Code, Cursor, Codex, Windsurf, Amp, Copilot, and Aider
• Skills and agents — discovers both skill files and agent definitions from each tool's directories
• Built-in editor — monospaced editor with Cmd+S save and frontmatter parsing
• Collections — organize skills and agents into groups without modifying source files
• Real-time file watching — FSEvents-based, instant updates on disk changes
• Full-text search — searches across name, description, and content
• Create new skills and agents — generates correct boilerplate per tool
• Remote servers — connects to servers running OpenClaw, Hermes, or compatible layouts to discover, browse, and install skills

Architecture and Tech Stack

The project is written in Swift and uses SwiftUI for the interface and SwiftData for persistence of Skill and SkillCollection models. Sparkle is the only external dependency, used for auto-updates and pulled automatically via Swift Package Manager. The Xcode project is generated from a project.yml file using xcodegen. The marketing website is a separate Astro 6 site in the site/ subdirectory.

Update: v1.15.0

The latest release is v1.15.0, published on April 29, 2026. The repository was created in March 2026 and has accumulated over 1,500 GitHub stars and 100 forks, signaling rapid early adoption within the AI developer tooling community. The project is actively maintained with the last push recorded in late April 2026. The FSL-1.1-MIT license grants a future MIT license on the second anniversary of the software's release date.