Connects your Claude agent to the Nullcone threat intelligence network with 890K+ indicators of compromise. You get 30+ tools for checking IPs, domains, URLs, file hashes, and CVEs against live threat feeds, plus two AI-specific lookups: prompt injection payload detection and MCP tool definition scanning. The hosted endpoint at nullcone.ai/mcp requires no API key and supports incremental polling via poll_since for staying current without webhooks. Reach for this when you're building agents that handle untrusted input, load third-party skills, or need to validate artifacts against known malware hashes and malicious infrastructure. Submit your own IOCs back to the network or self-host the Python server for full control over stdio transport.
Real-time threat intelligence for AI agents, exposed as a Model Context Protocol server. Check IPs, domains, URLs, hashes, CVEs, prompt-injection payloads, and malicious AI-skill / MCP-tool definitions against the Nullcone network — 890K+ IOCs, free, no API key.
Backed by nullcone.ai.
The server is hosted at https://nullcone.ai/mcp over streamable HTTP. Add it to any MCP client:
Claude Code
claude mcp add --transport http nullcone https://nullcone.ai/mcp
Cursor / other MCP clients — add to your MCP config:
{
"mcpServers": {
"nullcone": {
"url": "https://nullcone.ai/mcp"
}
}
}
No signup or token required. Read tools and IOC submission are open; destructive tools (e.g. revoke_ioc) are disabled on the public endpoint.
30+ tools including:
lookup_ioc(value) — check any indicator against the feedrecent_threats(limit, min_severity) — current threat picturesubmit_ioc(...) / submit_batch(...) — contribute indicatorscheck_prompt(...) — sub-millisecond prompt-injection lookupvalidate_skill(...) / scan_skill_content(...) — vet MCP tools / AI skills before loadingpoll_since(last_id) — incremental sync, no persistent connectionget_stats(), list_families(), search_by_type(...), and moreResources: threat://stats, threat://recent, threat://families, threat://family/{name}, threat://ioc/{value}.
Prompts: analyze_ioc, triage_alert, threat_brief.
The server is built on the public nullcone SDK.
pip install -r requirements.txt
MCP_TRANSPORT=streamable-http MCP_PORT=8001 python server.py
Or with Docker:
docker build -t nullcone-mcp .
docker run -p 8001:8001 nullcone-mcp
Set MCP_PUBLIC=1 to run an anonymous public endpoint (disables destructive tools); omit it for full local control over stdio (MCP_TRANSPORT=stdio).
Rising Sun License v1.0 — see LICENSE. Free for individuals and small teams.
io.github.ericm1018/skillfm-llm-cost-optimizer-openai-anthropic-usage
io.github.mikerawsonnz/llm-orchestration-agent
io.github.mikerawsonnz/authenticated-llm-agent
labforgedev/copilot-memory-mcp
csoai-org/agent-prompt-injection-firewall-mcp
io.github.mikerawsonnz/authenticated-multi-llm-agent