google-gemini-file-search

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • External Downloads (LOW): The skill references the installation of @google/genai. This package resides within the @google scope, which is a trusted organization. Per the [TRUST-SCOPE-RULE], this finding is downgraded to LOW.
  • Indirect Prompt Injection (LOW): The skill is designed to facilitate document ingestion for RAG (Retrieval-Augmented Generation), which creates an inherent surface for indirect prompt injection.
  • Ingestion points: Document upload operations via ai.fileSearchStores.uploadToFileSearchStore in references/setup-guide.md and references/error-catalog.md.
  • Boundary markers: Not present; the templates do not show specific delimiters or instructions to the agent to ignore embedded commands within the searched files.
  • Capability inventory: The skill uses the Google Gemini SDK to manage file stores and perform queries; it does not contain arbitrary shell execution or file-write capabilities.
  • Sanitization: No sanitization or validation logic for ingested file content is included in the provided script or documentation.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:05 PM