databricks-agent-bricks

Pass

Audited by Gen Agent Trust Hub on Aug 11, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill facilitates the creation of Knowledge Assistants and Supervisor Agents that process untrusted data from Unity Catalog Volumes and external MCP servers, introducing a surface for indirect prompt injection.
  • Ingestion points: Document files in Unity Catalog Volumes and external MCP server responses.
  • Boundary markers: The documentation suggests using instructions to guide behavior and citing sources, but no strict boundary markers for untrusted content are enforced.
  • Capability inventory: The skill allows the agent to read document volumes, call model serving endpoints, execute Unity Catalog functions, and interact with external systems via HTTP connections.
  • Sanitization: No mention of sanitizing or filtering external content before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 11, 2026, 02:51 PM
Security Audit — agent-trust-hub — databricks-agent-bricks