researching-topics

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: 0.90). This skill explicitly invokes WebSearch and WebFetch and instructs the agent to consult public/community sources (e.g., "community discussions (Reddit, HN, etc.)" in the Research Strategy and "Invokes
  • WebSearch, WebFetch" in Integration Points), so it fetches and synthesizes untrusted third‑party content that can influence recommendations.

Issues (1)

W011
MEDIUM

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

Audit Metadata
Risk Level
MEDIUM
Analyzed
May 14, 2026, 07:43 AM
Issues
1