# workit MCP server

> Provides workflow rails for agentic coding by managing git operations, recording test evidence, and enforcing verification policies.

- Repository: https://github.com/BrainerVirus/workit
- Page: https://hostedmcp.org/servers/brainervirus/workit
- Transport: stdio
- Stars: 1
- License: MIT

## Install

### Claude Code

```sh
claude mcp add workit -- npx -y @workit-dev/core
```

### Cursor (`~/.cursor/mcp.json`)

```json
{
  "mcpServers": {
    "workit": {
      "command": "npx",
      "args": [
        "-y",
        "@workit-dev/core"
      ]
    }
  }
}
```

### VS Code (`.vscode/mcp.json`)

```json
{
  "servers": {
    "workit": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "@workit-dev/core"
      ]
    }
  }
}
```

### VS Code CLI

```sh
code --add-mcp '{"name":"workit","type":"stdio","command":"npx","args":["-y","@workit-dev/core"]}'
```

### Claude Desktop (`claude_desktop_config.json`)

```json
{
  "mcpServers": {
    "workit": {
      "command": "npx",
      "args": [
        "-y",
        "@workit-dev/core"
      ]
    }
  }
}
```

### Codex

```sh
codex mcp add workit -- npx -y @workit-dev/core
```

### Codex config.toml (`~/.codex/config.toml`)

```toml
[mcp_servers.workit]
command = "npx"
args = ["-y", "@workit-dev/core"]
```

### Gemini CLI

```sh
gemini mcp add -s user workit npx -- -y @workit-dev/core
```

### Gemini settings.json (`~/.gemini/settings.json`)

```json
{
  "mcpServers": {
    "workit": {
      "command": "npx",
      "args": [
        "-y",
        "@workit-dev/core"
      ]
    }
  }
}
```

## FAQ

### What does workit do for agentic coding?

The agent decides what to do, while workit handles the execution details like branching, committing, pushing, PRs, CI, and merging, while recording proof of observations.

### How does workit ensure tests actually passed?

It uses 'workit check' to run real test commands and records the exit code, HEAD, and tree, providing observed evidence rather than just claims.

### Can an agent verify its own code changes?

No, the system enforces a rule where the session that wrote the code can never be the one to verify it.

### How can I control what an agent is allowed to do?

You can use per-workspace grants to decide how far an agent may go, such as whether it can push, create PRs, or merge without asking.
