academic-paper

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

• Third-party content exposure detected (high risk: 1.00). The skill's required Phase 1 "Literature search — systematic search strategy, source screening" (SKILL.md Orchestration Workflow) and the citation_compliance_agent's DOI/URL and Retraction Watch verification steps explicitly fetch and verify public web sources via DOI or "WebSearch"/RetractionWatch (agents/citation_compliance_agent.md and SKILL.md), so the agents ingest untrusted third‑party web content that is read and used to drive drafting, citation fixes, and review decisions.