reference-checker

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime verification workflow requires fetching/consulting external sources (e.g., DOI resolution via Crossref/publisher pages, PubMed, and especially Chinese title searches on CNKI/Wanfang/VIP/official journal sites), which can ingest outsider-authored free text from those web/database pages into the agent’s LLM context.