Ancient Near East Research

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 1.00). The skill explicitly fetches and scrapes open/public third‑party sites and arbitrary URLs (e.g., Omnisearch/Exa discovery, Firecrawl scrape, Obscura/ane_stealth_fetch.sh, and fetch_jstor.py as shown in SKILL.md and scripts) and then parses/uses that content in its pipeline (scrape → PDF download → Mistral OCR → compile_quotations), so untrusted web content can directly influence tool actions and decision-making.