search-cluster

Warn

Audited by Snyk on Feb 23, 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). scripts/search-cluster.py explicitly fetches and parses open/public third‑party content — e.g., reddit via https://www.reddit.com/search.json (reddit_search), arbitrary RSS/Atom feeds via rss_fetch, Google CSE (google_search), Wikipedia API (wiki_search), and NewsAPI (newsapi_search) — and returns their text snippets as aggregated results that an agent would read and could use to drive subsequent actions, exposing it to untrusted user-generated content that could carry indirect prompt injection.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 23, 2026, 06:54 AM