databricks-cli
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches the official Databricks CLI installation script and binary releases from the Databricks organization's GitHub repository.
- [COMMAND_EXECUTION]: The helper scripts in the
scripts/directory utilizesubprocess.runto execute thedatabrickscommand-line utility for authentication, metadata retrieval, and command execution. - [INDIRECT_PROMPT_INJECTION]: The skill ingests and displays data from external Databricks workspaces (such as table schemas, comments, and query results). Malicious instructions embedded in this data could potentially influence agent behavior if not treated as untrusted content.
- Ingestion points:
scripts/run.py(fetching command output viaprint_result) andscripts/metadata.py(fetching table descriptions). - Boundary markers: None identified in the provided prompt structures.
- Capability inventory:
subprocess.runcalls to the Databricks CLI for file imports, job runs, and API interactions. - Sanitization: Minimal string escaping is performed when wrapping SQL queries in Python for execution.
Audit Metadata