topic-agent

Warn

Audited by Snyk on Apr 5, 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). Yes — the SKILL.md explicitly instructs the agent to use WebSearch to fetch and ingest open/public third‑party content (e.g., Reddit posts, Product Hunt entries, blog articles, official blogs) and requires specific article/post URLs as inputs for generating and scoring topics, so untrusted user-generated content can directly influence downstream decisions.

Issues (1)

W011
MEDIUM

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

Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 5, 2026, 07:35 AM
Issues
1
Security Audit — snyk — topic-agent