code-review
Fail
Audited by Snyk on Mar 9, 2026
Risk Level: HIGH
Full Analysis
HIGH W007: Insecure credential handling detected in skill instructions.
- Insecure credential handling detected (high risk: 0.90). The prompt requires reading files and outputting changed lines and code suggestions verbatim (via sub-agents and consolidated review) without any instruction to redact secrets, so repository-stored API keys or passwords could be included in the LLM's output.
Audit Metadata