open-agent-hub

Why we included this project

People who work day to day with Claude Code, Cursor, Trae, or similar assistants tend to collect a messy pile of custom skills and slash commands that never quite match up between projects. open-agent-hub is built around that problem. A single zero-dependency CLI command ties modular prompting skills, expert agent roles, and slash commands into the config directories of eight coding assistants, either per workspace or globally. The bundled library already holds more than eighty ready-to-use skills plus agent prompt templates, and because every component is plain Markdown with YAML frontmatter, you can read it, edit it, and commit it just like any other source file. Teams that want consistent assistant behavior across repos and machines can switch capabilities on or off and check their status with straightforward commands, instead of hand-copying config around.