skills/bmsuisse/skills/spark-connect/Gen Agent Trust Hub

spark-connect

Pass

Audited by Gen Agent Trust Hub on Oct 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The helper script scripts/check_spark_env.py uses subprocess.run to call the official Databricks CLI and the uv package manager for environment checks and configuration. \n- [EXTERNAL_DOWNLOADS]: The skill manages the installation of the databricks-connect package from a public registry, ensuring the local version matches the remote cluster's runtime version. \n- [INDIRECT_PROMPT_INJECTION]: The script scripts/check_spark_env.py parses JSON output from the Databricks CLI, representing a potential data ingestion surface. \n
  • Ingestion points: Output from the databricks clusters get command. \n
  • Boundary markers: None. \n
  • Capability inventory: The script executes subprocesses to install packages via uv. \n
  • Sanitization: The script uses json.loads for structured parsing and a strict regular expression (\d+\.\d+) to validate the version string before use. \n- [DYNAMIC_EXECUTION]: The skill dynamically constructs installation commands to pin the databricks-connect package to the specific version required by the target cluster.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 6, 2026, 08:47 AM