bmad-testarch-test-review

Point the workflow at a test file, directory, or suite and it parses structure, fixtures, assertions, and waits, then scores each file against the criteria registry. The output is a markdown report with a 0-100 score, letter grade, and file:line findings with recommended fixes.

It checks hard waits, tautological or missing assertions, conditional logic, shared state, oversized files, missing test IDs and priority markers, network-first violations, and flakiness patterns. Each criterion has a pinned severity from Critical to Low defined in the criteria registry.

Yes. Setting headless to true skips the greeting and interactive menu, executes Create mode directly, and never prompts. You supply review_files as the authoritative review set and optionally context_files, and the run resolves everything from configuration.

No. Coverage analysis is explicitly out of scope; the review only audits the quality of existing tests and never generates tests. Coverage mapping and coverage gate decisions are routed to the separate trace workflow.

Scoring starts at 100, deducts 10 per Critical, 5 per High, 2 per Medium, and 1 per Low violation, then adds up to 30 bonus points for patterns like fixtures and network-first. The highest severity caps the score, and the grade maps from the capped value.

Yes. The registry includes Maestro-specific rows for flow assertions, secrets in flows, and optional outcome steps, plus Pact rows for worker parallelism and pool isolation. Stack detection loads the matching knowledge fragments only when relevant.