verification-loop

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The security scan explicitly runs grep for patterns like "sk-" and "api_key" and returns matching lines (head -10), which would expose any secret strings verbatim in the agent's output.