// agents
AI Coding Agents
An AI coding agent is a coding tool extended with rules, skills, and MCP servers so it can run one job on its own. Browse working agents below, or learn how the parts fit together in the build guide.
Browse AI Coding Agents
Every category has its own shelf.
What Is an AI Coding Agent?
An AI coding agent is a coding assistant set up to run one job on its own. You build one from four parts. A tool like Claude Code or Cursor does the work. Rules set your standards. Skills teach it specialized tasks. MCP servers connect it to outside services like GitHub or your database.
Every agent in this directory lists the exact parts it uses, so you can install the same setup in your own environment. If a term is new, the glossary covers each one in plain language.
Build Your Own Agent
Assemble one from parts in the workshop. Browse skills, MCP servers, and rules, then follow the build guide from first setup to a working agent.
Agent Questions
What is the difference between an agent and a skill?
A skill is one packaged procedure, like generating tests or reviewing a diff. An agent is a full setup that combines a tool with rules, skills, and MCP servers to run a complete job. Skills are ingredients. Agents are the assembled result.
Which coding tool should I build my first agent on?
Start with the tool you already work in. Rules, skills, and MCP servers exist for Claude Code, Cursor, Copilot, Codex, Cline, and Windsurf, and the build steps are the same on each. Switching later rarely means starting over.
Can coding agents handle work beyond writing code?
Yes. Agents built on coding tools also handle jobs like bug triage, standup summaries, documentation upkeep, and research. If the job touches your repo, your issues, or your docs, an agent can usually run it.
Are the agents in this directory free?
Every agent recipe here is free to copy. You pay only for what the underlying tool costs, such as a Claude or Cursor subscription, plus any services an MCP server connects to.