BigTesty plugin
BigTesty runs SQL tests on real BigQuery inside short-lived datasets, and four skills teach a coding agent to write them.
- 68 ★ on the repository
- Last commit
Extended access
Why this level
- It creates and deletes datasets and runs queries in your Google Cloud project.
- It consumes BigQuery quota on every test run.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/bigtestyDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- It creates and deletes datasets and runs queries in your Google Cloud project.
- It consumes BigQuery quota on every test run.
This is third-party code. Review the repository files before installing.
What it does
BigTesty checks production BigQuery queries on the real engine, with no emulator. A test is written in YAML or JSON in a given, when, then shape: load rows, run SQL, compare a query result or table contents. References to datasets are rewritten to isolated ones, so production SQL stays unchanged. Each scenario gets its own datasets, created and destroyed by the tool itself. Parallel runs from developers and CI do not collide. It ships a pytest plugin, the `bigtesty skills install` command and a Claude Code plugin with skills for setup, test writing and failure debugging.
Who it is for. Data engineers on Google Cloud who verify BigQuery SQL, MERGE statements and procedures and want an agent to write the tests.
Good fit when
- A MERGE or a procedure needs coverage that mocks cannot give.
- Several developers and a CI pipeline run checks at once.
- An agent has to set up `bigtesty.yaml` and write the scenarios itself.
- A pytest test fails and the troubleshooting skill should walk through it.
Not a fit when
- The project does not use BigQuery.
- There is no Google Cloud project access or default credentials.
- You need free local tests with no cloud calls.
Example request
Set up BigTesty in this project, write a scenario for the MERGE on the job_failure table and run it through pytest.Limitations
Python 3.11 and BigQuery access through Application Default Credentials are required, via `gcloud auth application-default login`. Tests run on the real engine and use quota, because BigQuery bills at least 10 MiB per query. A CI service account needs rights to create and delete datasets. Versions before 1.0 used Pulumi and another file format, with a migration section for them.
How to disable. Remove the package with `pip uninstall bigtesty` or `uv remove bigtesty`, uninstall the Claude Code plugin and delete the skill folders from `.claude/skills` and `.agents/skills`. Remove any leftover datasets in the BigQuery console.
FAQ
How do I install the agent skills?
Run `bigtesty skills install` and the skills land in the project's `.claude/skills` and `.agents/skills`. Claude Code also has a plugin: `/plugin install bigtesty --marketplace tosun-si/bigtesty`.
Is a BigQuery emulator needed?
No. Tests run on the real engine, and isolation comes from short-lived datasets per scenario.
Can I write tests in Python?
Yes. The package works as a pytest plugin that collects `*.bigtesty.yaml` files, and the shared infrastructure is available through the `bigtesty_infra` fixture.
What happens to test data after a run?
The tool deletes the scenario datasets itself. The `labels` option in the config helps attribute cost per team.
What do the bundled skills cover?
Four skills: setup, writing-tests, python-tests and troubleshooting. They lead an agent from configuration to debugging a failed scenario.
Discussion
No questions or reviews yet.
Related
PluginsQA agents, skills and an MCP server: test generation, coverage gaps, flaky tests, security scans and chaos testing
A plugin of testing skills: from requirements and test cases to API, mobile and web runs backed by traceable evidence.
Thirteen testing skills from requirements to report: strategy, cases, API checks, Playwright and flaky-test triage.
Seven QA agents and ten skills for test automation: Playwright in TypeScript, Selenium in Java, API, accessibility and ISTQB.
MCP server, agent and skills for end-to-end UI tests where the agent drives the computer through vision on cloud machines
A library of 169 QA skills per language: test design, API, UI, regression, load, security and LLM testing.
README in short
Test files sit next to the SQL and describe data, query and expectation. Real BigQuery and short-lived datasets replace an emulator and mocks.