topical-authority-mapper

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs the agent to fetch and ingest public third‑party content — e.g., running site-content-catalog against competitor URLs in Phase 1B and running reddit-post-finder/web_search for question mining in Phase 2B — and to read and act on that content to build topic maps and prioritization, so untrusted user-generated pages could inject instructions that change tool use or decisions.