evaluation

Warn

Audited by Snyk on Mar 16, 2026

Risk Level: MEDIUM
Full Analysis

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

  • Third-party content exposure detected (high risk: 0.80). The skill's required Step 4 "Research Each Option" explicitly directs the evaluator to consult public third-party sources (official documentation, GitHub repos, Stack Overflow/forums, benchmarks, CVE databases), meaning the agent/workflow will ingest and act on untrusted user-generated web content that could influence decisions.

Issues (1)

W011
MEDIUM

Third-party content exposure detected (indirect prompt injection risk).

Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 16, 2026, 06:15 PM
Issues
1
Security Audit — snyk — evaluation