Unified Search
Warn
Audited by Snyk on Apr 8, 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). This skill performs live searches against public providers (Tavily, Brave, Linkup, SearXNG) and ingests their search results — explicitly including user-generated sources like Reddit and Hacker News via SearXNG as shown in the SKILL.md uber_search usage and categories — which the agent is expected to read and act on, allowing third-party content to influence behavior.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata