# Tripwire MCP server

> Blocks prompt-injection-driven tool calls by verifying if actions are grounded in evidence and consistent with user intent.

- Repository: https://github.com/bonesdefi/Tripwire
- Page: https://hostedmcp.org/servers/bonesdefi/tripwire
- Transport: stdio
- Stars: 1
- License: MIT
- Latest version: 0.3.1

## Install

### Claude Code

```sh
claude mcp add tripwire -- npx -y tripwire-mcp
```

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

```json
{
  "mcpServers": {
    "tripwire": {
      "command": "npx",
      "args": [
        "-y",
        "tripwire-mcp"
      ]
    }
  }
}
```

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

```json
{
  "servers": {
    "tripwire": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "tripwire-mcp"
      ]
    }
  }
}
```

### VS Code CLI

```sh
code --add-mcp '{"name":"tripwire","type":"stdio","command":"npx","args":["-y","tripwire-mcp"]}'
```

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

```json
{
  "mcpServers": {
    "tripwire": {
      "command": "npx",
      "args": [
        "-y",
        "tripwire-mcp"
      ]
    }
  }
}
```

### Codex

```sh
codex mcp add tripwire -- npx -y tripwire-mcp
```

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

```toml
[mcp_servers.tripwire]
command = "npx"
args = ["-y", "tripwire-mcp"]
```

### Gemini CLI

```sh
gemini mcp add -s user tripwire npx -- -y tripwire-mcp
```

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

```json
{
  "mcpServers": {
    "tripwire": {
      "command": "npx",
      "args": [
        "-y",
        "tripwire-mcp"
      ]
    }
  }
}
```

## FAQ

### How does Tripwire prevent prompt injection attacks?

It uses a three-tier verification pipeline including deterministic receipts, provenance indexing, and multi-model consensus to ensure actions are grounded in evidence.

### How can I verify that the audit logs have not been tampered with?

You can use the `tripwire verify-log` command to cryptographically re-validate the hash-chained, append-only audit log.
