NYC

huashu-info-search

Warn

Audited by Snyk on Feb 17, 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 performs automated multi-channel searches and ingests/saves public web and social content (e.g., TechCrunch, The Verge, Medium, Reddit, Hacker News, X/Twitter) into its knowledge base, so the agent reads untrusted, user-generated third-party content and could be exposed to indirect prompt injection.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 11:00 AM