geo-content-research
Warn
Audited by Snyk on Apr 27, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 1.00). The SKILL.md explicitly requires live web research and ingestion of untrusted, user-generated content (e.g., "Do research using web search" and the required "Reddit Mining (Required)" and Quora/People Also Ask searches in Phase 2), which the agent must read and use to drive prompt-targeting, prioritization, and content-generation decisions—creating a clear vector for indirect prompt injection.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata