agent-platform-rag-engine-management

Pass

Audited by Gen Agent Trust Hub on May 28, 2026

Risk Level: SAFE
Full Analysis
  • Authentication and Environment Management: The skill guides users through authenticating with Google Cloud using the gcloud CLI and setting up a Python virtual environment. These are standard practices for secure local development and cloud interaction.- Official SDK Integration: Operations for listing, inspecting, and querying RAG corpora are handled through the vertexai and google-genai libraries. Using these official SDKs ensures compatibility and adherence to vendor-supported communication patterns.- Retrieval-Augmented Generation (RAG) Surface: The skill involves fetching and processing data from external RAG corpora to provide grounded answers.
  • Ingestion points: Content is ingested through rag.retrieval_query and model generation tools as described in SKILL.md.
  • Boundary markers: There are no explicit instructions provided to the agent to treat retrieved RAG context as data rather than instructions.
  • Capability inventory: The skill uses shell commands for environment configuration and package management in SKILL.md.
  • Sanitization: The provided code does not include explicit sanitization of the content retrieved from the RAG corpus before it is processed by the model.
Audit Metadata
Risk Level
SAFE
Analyzed
May 28, 2026, 06:21 PM
Security Audit — agent-trust-hub — agent-platform-rag-engine-management