audit-full

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly loads entire codebases into the model context and instructs the agent to "scan for hardcoded credentials, API keys, tokens" and produce findings/reports without any redaction rules, which will require the LLM to read and likely output secret values verbatim (exfiltration risk).