fix-my-look

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt instructs the agent to copy and reuse long JWT-style task_id values verbatim and to pass asset/video/audio URLs directly in calls and run notes, which forces the LLM to output token-like or signed-URL secrets exactly and thus creates an exfiltration risk.