maps-geo MCP Server
Intent classifier for MCP server packages. Static analysis + behavior extraction + OWASP MCP Top 10 mapping. Five verdicts (benign/shallow/unknown/suspicious/malicious) so coverage gaps aren't hidden behind false greens.
Discovered via github-topic:model-context-protocol and last synced 3mo ago.
Install instructions not detected yet
Check the source repository for the latest setup steps.
Sigstore-style trust + runtime proxy
Runtime security plane: capability tokens, IFC, output sealing
Workspace config-file injection scanner (`.cursorrules`, `mcp.json`)
An intent rule fired with low/medium confidence
Layer
Meaning
Count
An intent rule fired with high confidence
No tool registration detected (SDK pattern unsupported) — **NOT a safe verdict**
Tools detected, zero behavior extracted from any of them — **NOT a safe verdict**
Tools detected, behavior extracted from at least one, no rule fired (bounded by [L4](./LIMITATIONS.md))