deep-research
Warn
Audited by Snyk on May 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). The SKILL.md explicitly requires ingesting and acting on public third-party sources (e.g., "public sources (web, public APIs, public docs)" and the "Source types to include: forums, Reddit, podcasts, court records, academic papers, GitHub") and instructs running a web-enabled agent and citing/clicking URLs, so untrusted user-generated/open-web content is read and can influence agent decisions.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata