arxiv-search
Warn
Audited by Snyk on Feb 16, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 0.70). The skill directly fetches and parses user-submitted papers from the public arXiv API (http://export.arxiv.org/api/query?... in arxiv_search.ts) and ingests paper titles and abstracts (untrusted, third‑party content) as part of its output, which could carry embedded instructions.
Audit Metadata