websocket-patterns

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

• Third-party content exposure detected (high risk: 0.70). The skill is a WebSocket server/client implementation that ingests only runtime network messages from connected peers (outsider-authored free text/JSON via ws.on('message') / ws.onmessage and JSON.parse(event.data)), which can contain attacker-controlled strings that become LLM-readable context if the agent uses them.