rag-architect

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

• Third-party content exposure detected (low risk: 0.10). The skill’s required runtime workflow is a RAG retrieval pipeline that only ingests the caller-provided query string and then uses it for embedding/hybrid search, reranking, and evaluation, but it does not specify reading any outsider-authored free text sources (e.g., emails, tickets, GitHub/issues) into the LLM context by default.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The LateChunker calls AutoModel.from_pretrained with trust_remote_code=True using the external model "jinaai/jina-embeddings-v2-base-en", which will fetch and execute remote repository code at runtime (risk: remote code execution from a model repo).