engineering-ml-engineer

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

• Third-party content exposure detected (high risk: 0.90). The skill's workflow and examples explicitly instruct retrieval and indexing of external documents and public datasets/models (see SKILL.md RAG guidance and references/rag-patterns.md and references/transformers-patterns.md which use indexing/retrieval and load_dataset()/from_pretrained()), so the agent would ingest untrusted third‑party content (public datasets/documents on the web/HuggingFace) that can directly influence retrieval, generation, and subsequent actions.