knowledge-graph-construction

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md workflow (Step 1: "Catalog the input data" listing "web pages" and other public documents) and the methodology (resources/methodology.md) explicitly instruct ingesting and LLM-processing content from web pages, literature, and public knowledge bases (e.g., DrugBank, Wikidata), meaning the agent will read and act on untrusted third‑party content that can influence extraction and downstream actions.