# delego MCP server

> Provides intent-bound action authorization for AI agents using deterministic policies, human-in-the-loop approvals, and signed audit trails.

- Repository: https://github.com/Delego-Dev/delego
- Page: https://hostedmcp.org/servers/delego-dev/delego
- Transport: stdio
- Stars: 25
- License: Apache-2.0
- Latest version: 0.3.4

## Install

### Claude Code

```sh
claude mcp add delego -- uvx delego
```

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

```json
{
  "mcpServers": {
    "delego": {
      "command": "uvx",
      "args": [
        "delego"
      ]
    }
  }
}
```

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

```json
{
  "servers": {
    "delego": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "delego"
      ]
    }
  }
}
```

### VS Code CLI

```sh
code --add-mcp '{"name":"delego","type":"stdio","command":"uvx","args":["delego"]}'
```

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

```json
{
  "mcpServers": {
    "delego": {
      "command": "uvx",
      "args": [
        "delego"
      ]
    }
  }
}
```

### Codex

```sh
codex mcp add delego -- uvx delego
```

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

```toml
[mcp_servers.delego]
command = "uvx"
args = ["delego"]
```

### Gemini CLI

```sh
gemini mcp add -s user delego uvx -- delego
```

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

```json
{
  "mcpServers": {
    "delego": {
      "command": "uvx",
      "args": [
        "delego"
      ]
    }
  }
}
```

## Environment variables

| Name | Required | Secret | Description |
|---|---|---|---|
| `DELEGO_HOME` | no | no | Directory holding the policy, signing keys, and audit ledger (defaults to ~/.delego). Set this to keep state project-scoped, e.g. /abs/path/to/project/.claude/.delego. |

## FAQ

### Does delego hold my secrets or credentials?

No, delego is not a credential vault. It sits between an agent and a credential broker to authorize actions without holding the secrets itself.

### Can an AI agent approve its own actions?

No, approving and denying actions are deliberately excluded from the MCP interface to ensure a human must decide out-of-band.

### How does delego prevent prompt injection attacks from executing unauthorized actions?

It uses deterministic policy engines and intent binding, ensuring decisions are made outside the stochastic LLM loop and bound to a specific action fingerprint.

### What happens if the audit trail is tampered with?

The system uses an Ed25519-signed hash chain; any editing, reordering, or removal of receipts will break verification.
