code-understanding
Fail
Audited by Snyk on Jun 7, 2026
Risk Level: HIGH
Full Analysis
HIGH W007: Insecure credential handling detected in skill instructions.
- Insecure credential handling detected (high risk: 1.00). The skill explicitly requires quoting exact source lines (file:line) as proof for assertions, which forces the model to output verbatim code lines — potentially including any hardcoded API keys, tokens, or passwords found in the codebase — creating a direct secret-exfiltration risk.
Issues (1)
W007
HIGHInsecure credential handling detected in skill instructions.
Audit Metadata