neo4j-vector-index-skill

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

• Third-party content exposure detected (high risk: 0.80). The skill's SKILL.md explicitly shows ingesting remote data via "LOAD CSV WITH HEADERS FROM 'https://example.com/data.csv' ..." (Cypher-Based Embedding Ingestion section), which instructs fetching arbitrary public URLs whose untrusted contents would be embedded and used in vector indexes and searches, so third‑party content can influence subsequent queries/actions.