# HOL Guard

A local-first security layer for AI agents. It intercepts terminal commands, file reads, and MCP tool calls to block risky actions before execution.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · ★ 977
- Author: hashgraph-online
- License: Apache-2.0 · policy=link-only · description only; the file stays with the author
- Source: https://github.com/hashgraph-online/hol-guard
- Updated: 2026-10-24
- HTML: https://affpapa.org/en/nevibecode/mcp/hol-guard

## What it does

HOL Guard intercepts AI agent operations before they execute on your operating system. It monitors tool calls, terminal commands, dependency installations, and access to private files. If it spots credential exposure, destructive commands, or prompt injections, it can halt the action or route it to a local approval queue.

Everything runs locally on your machine without requiring an account. You can inspect pending approvals, track action receipts, and audit installed tools directly through the CLI. Running the tool requires Python 3.10 or newer, so `uvx` fails immediately if the host lacks a supported Python environment.

## Requirements
- Needs: Python and uv (commands run via uvx)

## Install

### Claude Code (native platform)

**Platform CLI**
```bash
claude mcp add hol-guard -- uvx hol-guard
```
uvx runs the command named after the package; if the README names another one, use uvx --from hol-guard <command>

### Codex

**Platform CLI**
```bash
codex mcp add hol-guard -- uvx hol-guard
```
uvx runs the command named after the package; if the README names another one, use uvx --from hol-guard <command>

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "hol-guard": {
            "command": "uvx",
            "args": [
                "hol-guard"
            ]
        }
    }
}
```
uvx runs the command named after the package; if the README names another one, use uvx --from hol-guard <command>

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "hol-guard": {
            "type": "stdio",
            "command": "uvx",
            "args": [
                "hol-guard"
            ]
        }
    }
}
```
uvx runs the command named after the package; if the README names another one, use uvx --from hol-guard <command>

### Gemini CLI

**Platform CLI**
```bash
gemini mcp add hol-guard uvx hol-guard
```
uvx runs the command named after the package; if the README names another one, use uvx --from hol-guard <command>

## When to use it
- intercept destructive shell commands before execution
- stop sensitive credentials from leaking into prompts
- check package installations for supply-chain risks
- require manual approval for risky MCP tool calls

## Example prompts
- Check the list of pending approvals in HOL Guard.
- Scan workspace dependencies for supply-chain security risks.
- Audit the current MCP tool configurations for exposed secrets.

## Similar
- [REA](https://affpapa.org/en/nevibecode/mcp/rea.md) — MCP server · ★ 386989
- [World Monitor](https://affpapa.org/en/nevibecode/mcp/world-monitor.md) — MCP server · ★ 89408
- [mcp](https://affpapa.org/en/nevibecode/mcp/kubeshark-mcp.md) — MCP server · ★ 12109
- [Robin: AI-Powered Dark Web OSINT](https://affpapa.org/en/nevibecode/mcp/robin-ai-powered-dark-web-osint.md) — MCP server · ★ 7658
- [jshookmcp](https://affpapa.org/en/nevibecode/mcp/jshookmcp.md) — MCP server · ★ 2061
- [bernstein](https://affpapa.org/en/nevibecode/mcp/bernstein.md) — MCP server · ★ 1331

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md