name: 'step-03-quality-evaluation' description: 'Orchestrate parallel quality dimension checks (4 subprocesses)'
Launch 4 parallel subprocesses to evaluate test quality dimensions:
Coverage is intentionally excluded from this workflow and handled by trace.
{communication_language}Generate unique timestamp:
const timestamp = new Date().toISOString().replace(/[:.]/g, '-');
Prepare context for all subprocesses:
const subprocessContext = {
test_files: /* from Step 2 */,
knowledge_fragments_loaded: ['test-quality'],
timestamp: timestamp
};
Subprocess A: Determinism
./step-03a-subprocess-determinism.md/tmp/tea-test-review-determinism-${timestamp}.jsonSubprocess B: Isolation
./step-03b-subprocess-isolation.md/tmp/tea-test-review-isolation-${timestamp}.jsonSubprocess C: Maintainability
./step-03c-subprocess-maintainability.md/tmp/tea-test-review-maintainability-${timestamp}.jsonSubprocess D: Performance
./step-03e-subprocess-performance.md/tmp/tea-test-review-performance-${timestamp}.json⏳ Waiting for 4 quality subprocesses to complete...
✅ All 4 quality subprocesses completed successfully!
const outputs = ['determinism', 'isolation', 'maintainability', 'performance'].map(
(dim) => `/tmp/tea-test-review-${dim}-${timestamp}.json`,
);
outputs.forEach((output) => {
if (!fs.existsSync(output)) {
throw new Error(`Subprocess output missing: ${output}`);
}
});
🚀 Performance Report:
- Execution Mode: PARALLEL (4 subprocesses)
- Total Elapsed: ~max(all subprocesses) minutes
- Sequential Would Take: ~sum(all subprocesses) minutes
- Performance Gain: ~60-70% faster!
Pass the same timestamp value to Step 3F (do not regenerate it). Step 3F must read the exact temp files written in this step.
Load next step: {nextStepFile}
The aggregation step (3F) will:
Proceed to Step 3F when:
Do NOT proceed if any subprocess failed.
Master Rule: Parallel subprocess execution is MANDATORY for performance.