verification-loop

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill's security scan explicitly greps for patterns like "sk-" and "api_key" and uses head to show matches, which will surface secret strings and could cause the agent to include those secret values verbatim in its output since no redaction or safe-handling is required.