What problem does it solve? AI agents running BigQuery and Google Cloud CLI commands often create or modify cloud resources without any attribution, making it impossible to track which environment or client initiated the operations. This Skill enforces consistent labeling so resource usage can be traced back to its source. ## Core Features & Use Cases - BigQuery Label Enforcement: Automatically adds the mandatory --label datacloud:<ide> flag to supported bq subcommands (query, load, extract, mk, update, cp) while omitting it from read-only commands like bq show or bq ls that reject the flag. - gcloud Metrics Attribution: Prepends CLOUDSDK_METRICS_ENVIRONMENT=datacloud.<ide> to every gcloud command, covering both read-only and mutation operations. - Environment Deduction: Maps the current client (VS Code, Cloud Shell, Cloud Workstations, Gemini CLI, etc.) to the correct label value, falling back to ai-agent when unknown. - Use Case: When an AI agent runs a BigQuery query from VS Code as part of a Data Agent Kit workflow, the command becomes bq query --use_legacy_sql=false --label datacloud:vscode "SELECT 1", ensuring the job is attributed correctly. ## Quick Start Ask the agent to run a BigQuery query or any gcloud command and it will automatically attach the correct attribution label or environment variable for your current client.
