semantic-scholar-recs-guide

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

• Third-party content exposure detected (high risk: 0.70). The SKILL.md explicitly fetches and processes JSON from public third‑party APIs (https://api.openalex.org and https://api.crossref.org), using returned titles/abstracts/referenced_work IDs to drive selection, ranking, and follow‑on requests, so untrusted third‑party content can materially influence the agent's actions.