e2eWrite, run, and diagnose browser end-to-end tests (Playwright, Cypress, etc.) for critical user journeys: accessible locators, condition-based waits, isolated test data and accounts, trace/screenshot artifacts, and flake classification. Use when adding or fixing an e2e spec, triaging a flaky or CI-o
gap-auditRead-only audit of missing, weak, and stale tests for a repo, PR, feature, or route. Maps source files against tests with a bundled script, ranks gaps P0 to P3 by behavior and risk (never by coverage percentage), separates confirmed from inferred gaps, and proposes exact tests. Use when asked what t
harnessGenerate unit and integration tests for functions added or changed on the current branch (or a given path or symbol), run them with the /testing:run runner, and return a compact report. Manual command (/testing:harness).
runRun the project's tests, show only compact failures, and fix them in a bounded loop. Manual command (/testing:run).
unitWrite, update, and run focused unit or regression tests for changed behavior or a confirmed bug, in any language, and decide which tests to run after a change. Use for "add a test for this", "write a regression test for this fix", "write tests for this function", "which tests should I run?", or "is