databricks-unity-catalog

Pass

Audited by Gen Agent Trust Hub on Jun 28, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill exposes the mcp__databricks__execute_sql tool, which allows for the execution of arbitrary SQL queries. While intended for administrative use within the Databricks environment, this represents a significant capability that could be misused if the agent is compromised.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface (Category 8).
  • Ingestion points: Untrusted data enters the agent context via system tables documented in 5-system-tables.md, specifically logs such as system.access.audit and system.access.table_lineage which record activities of external actors.
  • Boundary markers: The provided instructions and reference files lack explicit boundary markers or system instructions to ignore potential commands embedded within the queried data.
  • Capability inventory: The skill provides extensive capabilities for data manipulation and code execution, including SQL execution via mcp__databricks__execute_sql (mentioned in SKILL.md) and comprehensive file operations (upload, delete) in 6-volumes.md.
  • Sanitization: There is no evidence of input validation or sanitization routines to process the data retrieved from system tables before it influences subsequent agent actions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 28, 2026, 05:51 PM
Security Audit — agent-trust-hub — databricks-unity-catalog