# ClawGuard Shield

ClawGuard Shield scans text and MCP server configs for prompt injections, jailbreaks, and tool poisoning across 204+ attack patterns in 15 languages.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · 395 installs
- Author: joergmichno
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://smithery.ai/server/joergmichno/clawguard
- Updated: 2026-03-22
- Security check: automated, 2026-11-16 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/mcp/clawguard-shield

## What it does

ClawGuard Shield functions as a hosted agent security scanner accessible at `https://clawguard--joergmichno.run.tools`. The `scan_text` tool inspects raw text inputs for jailbreak attempts and adversarial prompts. Every scan reports severity levels, confidence scores, and remediation steps.

Its companion tool, `scan_mcp_config`, audits settings of other MCP integrations. It flags malicious descriptions, tool poisoning, privilege escalation paths, hidden directives, and exfiltration risks. The scanner is designed with EU AI Act compliance in mind.

Because this server runs remotely over HTTP, your environment must allow outbound traffic to its endpoint.

## Requirements
- Remote server: `https://clawguard--joergmichno.run.tools`

## Install

### Claude Code (native platform)

**Platform CLI: remote server**
```bash
claude mcp add --transport http clawguard https://clawguard--joergmichno.run.tools
```

### Codex

**Platform CLI: remote server**
```bash
codex mcp add clawguard-shield --url https://clawguard--joergmichno.run.tools
```

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "clawguard-shield": {
            "url": "https://clawguard--joergmichno.run.tools"
        }
    }
}
```

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "clawguard-shield": {
            "type": "http",
            "url": "https://clawguard--joergmichno.run.tools"
        }
    }
}
```

### Gemini CLI

**Platform CLI: remote server**
```bash
gemini mcp add -t http clawguard-shield https://clawguard--joergmichno.run.tools
```

## When to use it
- detect prompt injection attacks in user input
- audit MCP server configs for poisoned tool definitions
- identify hidden instructions and data exfiltration vectors
- verify compliance with EU AI Act security requirements

## Example prompts
- Scan this user input for prompt injection with `scan_text`
- Check our MCP config JSON for tool poisoning using `scan_mcp_config`
- Analyze tool descriptions for hidden commands or exfiltration vectors

## Similar
- [Koma Gate MCP](https://affpapa.org/en/nevibecode/mcp/koma-gate-mcp.md) — MCP server · ★ 13 repo stars
- [ai-red-teaming](https://affpapa.org/en/nevibecode/skills/ai-red-teaming.md) — skill · 401 installs
- [llm-redteam-specialist](https://affpapa.org/en/nevibecode/agents/llm-redteam-specialist.md) — subagent · 139 installs
- [safety-scan](https://affpapa.org/en/nevibecode/skills/safety-scan.md) — skill · 744 installs
- [godmode](https://affpapa.org/en/nevibecode/skills/godmode.md) — skill · 214 installs
- [Aperion Shield](https://affpapa.org/en/nevibecode/mcp/aperion-shield.md) — MCP server · ★ 13 repo stars

---
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
