voice-note-ingest

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill requires preserving the user's exact spoken words verbatim in the generated brain pages (the "User's Words" block and storage/source fields), so if a user speaks an API key, password, or token the LLM will output it verbatim, creating an exfiltration risk.