serghei-qa
Fail
Audited by Snyk on Feb 23, 2026
Risk Level: HIGH
Full Analysis
HIGH W007: Insecure credential handling detected in skill instructions.
- Insecure credential handling detected (high risk: 1.00). The prompt's report template explicitly instructs the agent to include "offending code" verbatim (in "The Crime" blocks) so if the codebase contains hard-coded API keys, tokens, or passwords the agent would be required to reproduce those secret values in its output, creating an exfiltration risk.
Audit Metadata