ha-core

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill includes code that accepts a token parameter and directly inserts it into Authorization headers and WebSocket auth messages, which would require the model to place a secret value verbatim into generated requests or example code if a real token is provided.

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

• Third-party content exposure detected (medium risk: 0.30). Home Assistant WebSocket real-time event listening (SKILL.md lines 180-193) ingests free-text JSON event data into the runtime via async for message in self.ws → yield json.loads(message) and then reads/prints fields from event["event"]["data"].