ai-output-evaluation
Fail
Audited by Snyk on May 10, 2026
Risk Level: HIGH
Full Analysis
HIGH W007: Insecure credential handling detected in skill instructions.
- Insecure credential handling detected (high risk: 0.80). The skill's mandatory "quote specific evidence" rule requires the LLM to reproduce exact snippets from the evaluated output, which would cause it to echo any embedded secrets (API keys, tokens, passwords) verbatim if present and offers no redaction guidance, creating an exfiltration risk.
Issues (1)
W007
HIGHInsecure credential handling detected in skill instructions.
Audit Metadata