gds-test-review

Run a structured test review that counts tests by type, measures pass rates and durations, detects flaky and slow tests, and maps feature coverage. The workflow produces a report with prioritized recommendations for immediate, short-term, and long-term improvements.

Analyze recent test run history for inconsistent pass/fail results across the last 10 runs. Common causes include hard-coded waits, shared static state, and timing dependencies, which should be replaced with explicit waits and proper setup/teardown.

Coverage should be mapped by feature priority: P0 areas like the core loop and save/load need the deepest testing, while P2/P3 features can have lighter coverage. The review flags features with no tests and prioritizes gaps by risk and player impact.

Yes, the review supports Unity, Unreal, and Godot through a configurable game_engine setting, with an auto mode that detects the engine. The quality criteria and anti-pattern checks apply across engines.

Common anti-patterns include hard-coded waits like Thread.Sleep, shared static test state, testing private implementation details, missing cleanup of instantiated objects, and tests without assertions. Each has a recommended fix such as explicit waits or setup/teardown methods.

The workflow recommends three cadences: a quick weekly check of pass rates and flaky tests by QA, a monthly full review of coverage and quality by a tech lead, and a quarterly deep dive into infrastructure and strategy by the team.