explain-system-tradeoffs

Fail

Audited by Snyk on Feb 28, 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 instructs subagents to scan and report specific artifacts and "Detail" lines from files (including exact file paths and line ranges) without any guidance to redact secrets, which can cause the LLM to read and reproduce API keys, tokens, or passwords verbatim in its output.
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 28, 2026, 06:44 PM