search-specialist

Warn

Audited by Snyk on Feb 15, 2026

Risk Level: MEDIUM
Full Analysis

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

  • Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs the agent to search, scrape, and ingest content from open public sources and user-generated platforms—e.g., Reddit, G2/Capterra reviews, LinkedIn, GitHub, and arbitrary websites via web scraping and Google search—which the agent would read and interpret as part of its workflow.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 15, 2026, 08:48 PM