Skip to main content
Glama
OWASP Agentic MCP

OWASP Agentic MCP

README.md
<!-- mcp-name: CSOAI-ORG/owasp-agentic-mcp -->
[![MCP Scorecard: 86/100](https://img.shields.io/badge/proofof.ai-86%2F100-5b21b6)](https://proofof.ai/scorecard/owasp-agentic-mcp.html)

# Owasp Agentic MCP

[![MEOK AI Labs](https://img.shields.io/badge/MEOK-AI%20Labs-667eea)](https://meok.ai)
[![GSPC](https://img.shields.io/badge/GSPC-UNMEASURED-9ca3af)](https://councilof.ai/api/gspc)
[![License](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE)
[![PyPI](https://img.shields.io/badge/PyPI-Install-3775a9)](https://pypi.org/project/owasp_agentic_mcp/)

> OWASP Top 10 for Agentic AI security MCP server โ€” prompt injection detection, tool poisoning, exc...

OWASP Top 10 for Agentic AI security MCP server โ€” prompt injection detection, tool poisoning, excessive agency, data leakage assessment
<div align="center">

# OWASP Agentic MCP

**OWASP Top 10 for AI Agents Security Assessment โ€” Prompt Injection, Tool Poisoning, Data Leakage**

[![MCP](https://img.shields.io/badge/MCP-Server-blue)](https://github.com/CSOAI-ORG)
[![License](https://img.shields.io/badge/License-MIT-green)](LICENSE)
</div>

---

## ๐Ÿš€ Quick Start

```bash
# Install via pip
pip install owasp_agentic_mcp

# Or install via Smithery
npx -y @smithery/cli@latest install owasp-agentic-mcp --client claude
```

## โœจ Features

- MCP protocol compliant
- Easy installation
- Well-documented API
- Production-ready
- Active maintenance

## ๐Ÿ“– Documentation

- [Full Documentation](https://docs.meok.ai/owasp-agentic-mcp)
- [API Reference](https://api.meok.ai)
- [EU AI Act Compliance Guide](https://councilof.ai/compliance)

## ๐Ÿ›ก๏ธ Compliance

This MCP server is built with **EU AI Act compliance** built-in:

- โœ… Article 9 โ€” Risk Management System
- โœ… Article 13 โ€” Transparency & Instructions for Use
- โœ… Article 15 โ€” Bias Detection & Testing
- โœ… Article 26 โ€” FRIA Support (where applicable)
- โœ… Article 50 โ€” AI Content Watermarking (where applicable)

Need help getting compliant? **[Book a free 15-min diagnostic โ†’](https://cal.com/csoai/august-audit)**

## How MEOK charges

MEOK has no subscriptions and no tiers. You pay per job, and only when the agreed artifact has been delivered to you. There is no public price list. See <https://meok.ai/pricing>.

This repository's source is MIT-licensed and free to self-host.

## ๐Ÿค Part of the MEOK Ecosystem

This server is part of the **[MEOK AI Labs](https://meok.ai)** ecosystem โ€” 300+ MCP servers for sovereign AI governance.

| Domain | Purpose |
|--------|---------|
| [councilof.ai](https://councilof.ai) | EU AI Act compliance marketplace |
| [safetyof.ai](https://safetyof.ai) | AI safety & monitoring |
| [meok.ai](https://meok.ai) | Sovereign AI platform |
| [cobolbridge.ai](https://cobolbridge.ai) | Legacy modernization |

## ๐Ÿ“œ License

MIT ยฉ [CSOAI-ORG](https://github.com/CSOAI-ORG)

---

<p align="center">
  <sub>Built with ๐Ÿ’œ by <a href="https://meok.ai">MEOK AI Labs</a> ยท UK Companies House 16939677</sub>
</p>
Security assessment tools based on the OWASP Top 10 for AI Agents. Scan agent configurations for prompt injection vulnerabilities, tool poisoning risks, excessive agency, data leakage, and more.

## Tools

| Tool | Description | Parameters |
|------|-------------|------------|
| `full_agent_scan` | Full OWASP Agentic Top 10 security scan | `agent_config`, `tools`, `permissions` |
| `assess_agentic_security` | Assess against specific OWASP Agentic categories | `categories`, `agent_config` |
| `check_prompt_injection` | Test for prompt injection vulnerabilities | `system_prompt`, `user_input_template` |
| `check_tool_poisoning` | Check for tool poisoning risks | `tool_definitions`, `input_validation` |
| `check_excessive_agency` | Assess agency level vs minimum required | `allowed_tools`, `required_tools`, `permissions` |

## Installation

```bash
pip install mcp
```

### Claude Desktop / Cursor / VS Code / Windsurf
```json
{
  "mcpServers": {
    "owasp-agentic": {
      "command": "python",
      "args": ["path/to/server.py"]
    }
  }
}
```

## Usage Examples

### Full agent security scan
```json
{
  "agent_config": {
    "system_prompt": "You are a helpful assistant that can access email, calendar, and file system.",
    "allowed_tools": ["send_email", "read_calendar", "write_file"]
  },
  "permissions": ["email:send", "calendar:read", "files:write"]
}
```

### Check prompt injection
```json
{
  "system_prompt": "You are a financial advisor bot",
  "user_input_template": "{user_query} Please provide investment advice."
}
```

TDQS

A4.4/5.0

Scored across 5 tools

Disambiguation5/5

Each tool targets a distinct security aspect (overall assessment, data leakage, excessive agency, prompt injection, tool poisoning) with clear, non-overlapping descriptions, ensuring an agent can easily distinguish them.

Naming Consistency5/5

All tool names follow a consistent verb_noun pattern (e.g., check_data_leakage, assess_agent_security) with no mixed conventions or deviations, making navigation predictable.

Tool Count5/5

Five tools is a well-scoped count for a security assessment server, covering key OWASP agentic AI vulnerabilities without being too sparse or overwhelming.

Completeness4/5

The set covers major OWASP Top 10 areas but lacks tools for supply chain security or model theft. However, the comprehensive assess_agent_security tool partially fills gaps, making the surface reasonably complete.

Maintenance

ActivityMaintained
ResponsivenessOutdated data