verification-loop

Fail

Audited by Snyk on Mar 4, 2026

Risk Level: HIGH
Full Analysis

HIGH W007: Insecure credential handling detected in skill instructions.

  • Insecure credential handling detected (high risk: 0.90). The skill explicitly runs pattern searches (e.g., grep "sk-" and "api_key") and would surface matching lines in the report, meaning the LLM could read and output secret values verbatim.
Audit Metadata
Risk Level
HIGH
Analyzed
Mar 4, 2026, 03:44 PM