deep-research

Warn

Audited by Snyk on May 14, 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's required pipeline (Phase 3: "Execute Searches") mandates running 5+ web searches and using WebFetch to read full articles — explicitly including community sources like HN, Reddit, blog posts, and Stack Overflow — and then syntheses/recommendations are produced from those reads, so untrusted third-party content is ingested and can influence decisions (see Phase 3 and the "Community" search angle in SKILL.md).

Issues (1)

W011
MEDIUM

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

Audit Metadata
Risk Level
MEDIUM
Analyzed
May 14, 2026, 05:09 PM
Issues
1
Security Audit — snyk — deep-research