lev-social

Warn

Audited by Snyk on Mar 1, 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's Research Workflow and examples explicitly fetch and ingest public social-media content (bird search for Twitter/X and PostCrawl calls for Reddit/TikTok, plus Exa API background fetches) and instruct the agent to parse and act on that user-generated content for sentiment and report generation, creating a clear pathway for indirect prompt injection.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 1, 2026, 09:43 AM