Skills
skills/gemini-cli-extensions/data-agent-kit-starter-pack/gcp-spark/Gen Agent Trust Hub

gcp-spark

Pass

Audited by Gen Agent Trust Hub on Apr 28, 2026

Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands for managing environment configurations and file processing.
  • Evidence: Use of gcloud config, gcloud dataproc clusters list, jupyter nbconvert, and python3 -m py_compile to verify and manage resources (found in SKILL.md and resources/gcloud_dataproc.md).
  • [REMOTE_CODE_EXECUTION]: The skill facilitates the generation and remote execution of Spark scripts on GCP infrastructure.
  • Evidence: Submitting workloads to Dataproc via gcloud dataproc batches submit pyspark <SCRIPT_PATH.py> (found in resources/gcloud_dataproc.md).
  • [EXTERNAL_DOWNLOADS]: Fetches official connectors and configuration from trusted technology providers.
  • Evidence: Downloads the Spanner connector from Google's gs://spark-lib/ and references Microsoft's synapseml library for machine learning tasks (found in resources/read_write_data.md and resources/ml_tasks.md).
  • [DATA_EXFILTRATION]: The skill implements data ingestion workflows that read from cloud-hosted storage and databases.
  • Evidence: Instructions for reading data from Google Cloud Storage, BigQuery, Spanner, and BigLake Iceberg catalogs using spark.read and gcloud storage cat (found in resources/read_write_data.md and resources/schema_direct_inspection.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 28, 2026, 10:20 PM