prod-readiness

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The judge prompts force the model to include verbatim excerpts from repository files (file:line plus literal text) as evidence for pass/warn, which will cause any secrets present in those files (API keys, tokens, passwords, cookies, etc.) to be output by the LLM — a direct secret-exfiltration risk.