What it does
The server provides security assessment tools based on the OWASP Top 10 for AI Agents. It scans agent configurations for prompt injection vulnerabilities, tool poisoning risks, excessive agency and data leakage. A full scan covers the whole OWASP Agentic Top 10, and separate tools check prompt injection, tool poisoning and excessive agency, or assess against selected categories.
It suits teams reviewing an agent's system prompt, tool definitions, permissions and allowed tools before deployment. Each check takes the relevant configuration inputs, such as agent_config, tools, permissions, system_prompt, tool_definitions or allowed_tools, depending on the tool used.
The server runs as an MCP server, and the publisher's instructions cover installation in Claude Desktop, Cursor, VS Code and Windsurf.
Limitations and requirements
The free tier is limited to 10 scans per day. Higher tiers with more scans and reports are offered as paid plans.
Safety
Our team has not reviewed this resource. Only automated rules checked it; resources with high-risk signals are not published. These are signals, not a guarantee that installing it is safe. Try it in a sandbox project first.
- Runs code or changes an assistant's behaviour
Sources
- Official MCP Registry first seen 13 Nov 2026 · last seen 13 Nov 2026 · its page on the source
io.github.CSOAI-ORG/owasp-agentic-mcp
Owasp Agentic MCP Server by MEOK AI Labs
v1.0.9As shown on registry.modelcontextprotocol.io when we last read it