Knowledge Router

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). The skill explicitly instructs querying Context7 for documentation using mcp__context7__query-docs against public library IDs (e.g., /websites/fastapi_tiangolo, /fastapi/fastapi, /vercel/next.js) in its Execution Protocol, meaning it fetches and ingests open third-party docs that the agent will read and synthesize into actions—creating a clear avenue for indirect prompt injection.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The skill makes runtime calls to the Context7 MCP (e.g., mcp__context7__query-docs with libraryId "/websites/fastapi_tiangolo" and "/fastapi/fastapi") to fetch documentation that is merged into model responses, so external content fetched at runtime directly controls prompts.