skills/agentient/vibekit/rag-patterns/Gen Agent Trust Hub

rag-patterns

Pass

Audited by Gen Agent Trust Hub on Mar 31, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements standard integration patterns for Vertex AI RAG Engine using official SDKs (google-cloud-aiplatform, google-cloud-storage, google-genai).
  • [SAFE]: Data ingestion patterns for local files and Google Cloud Storage follow standard cloud service practices and are consistent with the skill's primary purpose.
  • [SAFE]: The integration instructions include specific prompt engineering constraints (e.g., 'ONLY answer using information from the retrieved documents') which serve as boundary markers to mitigate hallucination and indirect injection risks.
  • [EXTERNAL_DOWNLOADS]: The skill references authoritative external documentation from Google Cloud and well-known technology providers like Weaviate for developer education.
  • [DATA_EXPOSURE]: The code correctly utilizes environment variables for project and location configuration and follows standard GCS bucket access patterns.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 31, 2026, 09:08 AM
Security Audit — agent-trust-hub — rag-patterns