biorxiv-api

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly instructs the agent to fetch and ingest public preprint metadata and abstracts from the open API endpoints (e.g., https://api.biorxiv.org/details/... and /pubs/...), and shows workflows that filter and act on that returned content (monitoring, alerting), so untrusted third‑party content is read and can influence agent actions.