agent-platform-rag-engine-management

Pass

Audited by Gen Agent Trust Hub on Sep 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the user to execute shell commands for environment setup and authentication.
  • Includes gcloud auth login and gcloud auth application-default login for Cloud authentication.
  • Recommends the creation of a virtual environment using python3 -m venv and its activation.
  • [EXTERNAL_DOWNLOADS]: The instructions include the installation of external Python libraries using the package manager.
  • Fetches official libraries from the Google Cloud ecosystem: google-cloud-aiplatform and google-genai via pip install.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses an indirect prompt injection attack surface as it is designed to ingest and process data from external RAG corpora.
  • Ingestion points: Data is retrieved from RAG corpora using rag.retrieval_query in SKILL.md.
  • Boundary markers: The provided Python snippets for grounded generation in Section 4 do not include specific delimiters or "ignore instructions" warnings for the retrieved content.
  • Capability inventory: The skill uses the google-genai SDK to make network calls to Vertex AI for content generation and retrieval.
  • Sanitization: There is no explicit sanitization or filtering logic shown for the context text retrieved from the corpus before it is used for grounded generation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 18, 2026, 05:40 AM
Security Audit — agent-trust-hub — agent-platform-rag-engine-management