debug-loop
Fail
Audited by Snyk on Mar 6, 2026
Risk Level: HIGH
Full Analysis
HIGH W007: Insecure credential handling detected in skill instructions.
- Insecure credential handling detected (high risk: 0.90). This skill requires running real commands and "capture and display the REAL output" (and grepping the codebase), so it can force the agent to print verbatim command output that may contain API keys, tokens, or passwords, creating an exfiltration risk.
Audit Metadata