network-recon

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

• Third-party content exposure detected (high risk: 0.90). Phase 4: Web Enumeration instructs the agent to fetch and parse arbitrary web pages from target hosts (e.g., curl -s ... | head -50, curl -s .../robots.txt and nmap --script "http-*"), which are untrusted third-party contents the skill reads and uses to decide follow-up actions.