stevens-network-protocols

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

• Third-party content exposure detected (high risk: 0.80). The skill's SKILL.md includes a capture_and_analyze routine that opens a raw socket (socket.AF_PACKET) to capture live network packets and PacketAnalyzer decodes and displays application payloads (e.g., payload_preview = tcp.payload[:100].decode(...)), meaning the agent will ingest untrusted, third-party network traffic as part of its required workflow.