verification-loop

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill's security scan phase runs grep commands that will print matching lines (e.g., patterns like "sk-" or "api_key") and so, if the agent reports findings verbatim, it would require handling/outputting secret values directly, creating an exfiltration risk.