Plug vulnerability intelligence into your AI
The patrowl-intel-mcp server exposes correlated CVE, CVSS, EPSS, CISA KEV, public-exploit and trending-attack data to any MCP client — so you can just ask, in plain language.
Works with your AI clients
Claude
Desktop · Code · API
ChatGPT
Developer mode
Codex
CLI & IDE
Cursor
Editor agent
Ollama
Local models
Connect in seconds
Read-only, no API key. Pick your client below.
claude mcp add --transport http patrowl-intel https://mcp.intel.patrowl.io/mcpclaude mcp remove patrowl-intel
claude mcp add --transport http patrowl-intel https://mcp.intel.patrowl.io/mcpClaude Desktop: Settings → Connectors → Add custom connector, then paste the endpoint URL.
claude mcp add patrowl-intel -- uvx patrowl-intel-mcpRequires uv. Runs the server on your machine over stdio.
What is PatrowlIntel MCP?
PatrowlIntel is an MCP (Model Context Protocol) server that enables AI agents to interact with Patrowl — a vulnerability intelligence security platform — through natural language. Built in Python, it provides a structured interface for vulnerability management operations and workflows.
It acts as a translation layer between AI assistants (like ChatGPT, Claude, or other LLMs) and the public PatrowlIntel API. It doesn't contain AI itself — instead, it provides AI assistants with the tools they need to understand and work with security data.
Just ask
Your AI picks the right tool and returns a decision, not raw JSON.
“Brief me on CVE-2025-32433 — is it being exploited?”
Triage“Which KEV-listed Fortinet CVEs have public exploits?”
Hunt“Rank these 20 CVEs by risk and tell me what to patch first.”
Prioritize“What are this week's critical trending attacks?”
Monitor“Assess our exposure to Apache vulnerabilities with EPSS above 0.3.”
Exposure“Summarize the SSVC decision for CVE-2024-3400.”
DecideGive your AI a security brain
Free, open source and read-only. Install it once, prompt forever.