llm-council

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The prompt instructs the agent to scan and read arbitrary workspace files and to "include key context from workspace files" in the framed question and saved reports/transcripts, which can cause the model to ingest and reproduce secrets verbatim even though it doesn't explicitly ask for API keys — creating a realistic exfiltration risk.