semanticscholar-skill

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and ingests public third‑party content from the Semantic Scholar API (see SKILL.md Phase 2 and s2.py functions like search_bulk, get_paper, get_citations, search_snippets), and those returned abstracts/TLDRs and citation contexts are read and used by the agent (format_citations, format_results, recommendation flows) to decide follow-up actions, so untrusted external content can materially influence the agent's behavior.