Help us improve
Share bugs, ideas, or general feedback.
From bigquery-agent-analytics-tracing
Checks BQAA tracing environment variables and Python runtime dependencies, reporting any missing configuration. Advisory only; no mutations.
npx claudepluginhub googlecloudplatform/bigquery-agent-analytics-sdk --plugin bigquery-agent-analytics-tracingHow this command is triggered — by the user, by Claude, or both
Slash command
/bigquery-agent-analytics-tracing:bqaa-setupThe summary Claude sees in its command listing — used to decide when to auto-load this command
# /bqaa-setup Run the BQAA tracing setup check and report what (if anything) needs fixing before agent events can land in BigQuery. ## What to do 1. Invoke the wrapper script: The wrapper sets `PYTHONPATH` to the vendored package root and runs `python -m bigquery_agent_analytics_tracing.setup_check`. Exit code is `0` when the runtime is ready, `1` when something required is missing or unconfigured. 2. Surface the script's output to the user verbatim. The check is designed to be copy-pasteable — it already prints the exact `pip install ...` command for any missing...
/initInitializes Self-Care: configures LangSmith or LangFuse trace source, validates credentials and projects, creates .self-care/config.json.
/agent-brain-verifyVerifies Agent Brain installation, Python version, provider config, project setup, and server status via CLI checks, producing a health report with fixes.
/preflightRuns two-phase environment detection and compatibility analysis on a project, producing a unified report with verdicts (✅/⚠️/❌), fixes for breaks, risks, and warnings on runtimes, GPU stack, env vars, deps, services, and build tools.
/auditAudits tech stack and sets up one dev tooling bundle per run (formatting, linting, types, tests, observability, git hooks, database), installing/validating/remediating monorepo-aware.
/analyze-usageAnalyzes BigQuery usage patterns, costs, and query performance for a project over a timeframe, producing a report with user breakdowns, top queries, and optimization recommendations.
Share bugs, ideas, or general feedback.
Run the BQAA tracing setup check and report what (if anything) needs fixing before agent events can land in BigQuery.
Invoke the wrapper script:
bash "${CLAUDE_PLUGIN_ROOT}/scripts/run_setup_check.sh"
The wrapper sets PYTHONPATH to the vendored package root and runs
python -m bigquery_agent_analytics_tracing.setup_check. Exit code is
0 when the runtime is ready, 1 when something required is missing
or unconfigured.
Surface the script's output to the user verbatim. The check is designed
to be copy-pasteable — it already prints the exact pip install ...
command for any missing dep and the exact export BQAA_PROJECT_ID=...
line for any missing env var.
Do not modify the user's shell rc files, install packages on their behalf, or run any GCP API calls in this command. This is advisory only — if the user wants a guided install, they can run the suggested commands themselves.
BQAA_PROJECT_ID — required. Falls back to GCP_PROJECT_ID /
GOOGLE_CLOUD_PROJECT if unset.BQAA_DATASET — required. Falls back to BQ_DATASET if unset.BQAA_PYTHON — the interpreter the hooks will exec; defaults to the
current python3. The check verifies it can import the tracing
package and required runtime deps.google-cloud-bigquery — always required (both writer paths use it).google-cloud-bigquery-storage + pyarrow — required only for the
lower-latency Storage Write path; absence is reported as advisory
(the drainer falls back to insert_rows_json).~/.claude/settings.json.For the bootstrap that actually does those things, see
plugins/claude_code/MARKETPLACE.md
for the IAM model and the manual gcloud / bq commands.