gemini-interactions-api

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill contains instructions that explicitly direct the agent to disregard its internal training data regarding model lifecycle and versioning in favor of the provided IDs.
  • [EXTERNAL_DOWNLOADS]: The skill provides references to official documentation, API references, and OpenAPI specifications hosted on Google's developer domains.
  • [PROMPT_INJECTION]: The skill describes a surface for indirect prompt injection by documenting the use of tools that ingest untrusted data. 1. Ingestion points: Google Search, URL context, and file search result types in the data model. 2. Boundary markers: None described in the prompt construction logic. 3. Capability inventory: The skill enables model interaction and background tool execution via the described SDKs. 4. Sanitization: No sanitization or filtering is specified for external content.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 12:53 PM
Security Audit — agent-trust-hub — gemini-interactions-api