Coding Tools MCP
A local MCP coding runtime gives AI clients file, command and Git access, with workspace boundaries, permission controls and Docker support.
Coding Tools MCP gives AI chat apps and coding agents access to a repository on your machine. It's for developers who want to use an existing MCP client to inspect code, make edits and check a running project.
- Code access and edits: Agents can search files, view images and apply changes across multiple files, with conflict checks and rollback support.
- Commands and Git: Clients can run commands, interact with ongoing processes and inspect Git history, diffs and working changes.
- Workspace controls: Path boundaries reject access outside the chosen workspace, while permission modes govern network access, scripts and destructive commands.
- Client compatibility: It connects to ChatGPT, Claude Desktop, Claude Code, Codex, Cursor, VS Code, Windsurf, Cline and Gemini CLI without tying the runtime to one model provider.
- Remote workstation access: Bearer tokens and OAuth support authenticated connections through an HTTPS tunnel so a remote client can work against your machine.
The Python server is open source under Apache 2.0 and can run locally or in Docker. It supplies tools to a connected client; it doesn't run that client's AI model locally. Tool output is summarized and capped to limit context use.
Secret filtering is included, and Linux can add filesystem confinement through Landlock. These controls aren't a complete operating-system sandbox; Docker or a VM provides additional isolation for untrusted work. Anonymous usage telemetry is on by default and can be disabled. It excludes file contents, paths, commands and arguments.
