Testing Orchestration Patterns

┌─────────────────────────────────────────────────────────────┐
│                                                             │
│   Confidence through verification.                          │
│   Generate, execute, analyze — all in parallel.             │
│                                                             │
└─────────────────────────────────────────────────────────────┘

Load when: Test generation, test execution, coverage analysis, test maintenance, E2E testing Common patterns: Coverage-Driven Generation, Parallel Test Suites, Broken Test Triage

Table of Contents

1. Test Generation
2. Test Execution
3. Coverage Analysis
4. Test Maintenance
5. E2E Testing

---

Test Generation

Pattern: Coverage-Driven Generation

User Request: "Add tests for the UserService"

Phase 1: EXPLORE
└─ Explore agent: Understand UserService methods, dependencies

Phase 2: FAN-OUT (Parallel test writing)
├─ Agent A: Unit tests for method group 1
├─ Agent B: Unit tests for method group 2
├─ Agent C: Integration tests for external dependencies
└─ Agent D: Edge cases and error scenarios

Phase 3: PIPELINE
└─ General-purpose agent: Verify tests pass, check coverage

Pattern: Behavior-First

User Request: "Test the checkout flow"

Phase 1: EXPLORE
└─ Explore agent: Map checkout flow steps and branches

Phase 2: PIPELINE (Generate by behavior)
├─ General-purpose agent: Happy path tests
├─ General-purpose agent: Error path tests
└─ General-purpose agent: Edge case tests

Phase 3: BACKGROUND
└─ Background agent: Run tests, report results

---

Test Execution

Pattern: Parallel Test Suites

User Request: "Run all tests"

Phase 1: FAN-OUT (Parallel suites)
├─ Background agent: Unit tests
├─ Background agent: Integration tests
├─ Background agent: E2E tests
└─ Background agent: Performance tests

Phase 2: REDUCE
└─ General-purpose agent: Aggregate results, identify failures

Pattern: Targeted Execution

User Request: "Test the changes I made"

Phase 1: EXPLORE
└─ Explore agent: Identify changed files and affected tests

Phase 2: FAN-OUT
├─ Background agent: Run directly affected tests
└─ Background agent: Run dependent module tests

Phase 3: PIPELINE
└─ General-purpose agent: Report results, suggest additional tests

---

Coverage Analysis

Pattern: Gap Identification

User Request: "Improve test coverage"

Phase 1: BACKGROUND
└─ Background agent: Run coverage report

Phase 2: EXPLORE
└─ Explore agent: Identify critical uncovered paths

Phase 3: FAN-OUT (Prioritized gap filling)
├─ Agent A: Tests for critical uncovered module 1
├─ Agent B: Tests for critical uncovered module 2
└─ Agent C: Tests for error handlers

Phase 4: PIPELINE
└─ General-purpose agent: Re-run coverage, verify improvement

---

Test Maintenance

Pattern: Broken Test Triage

User Request: "Fix failing tests"

Phase 1: BACKGROUND
└─ Background agent: Run tests, capture failures

Phase 2: FAN-OUT (Parallel diagnosis)
├─ Agent A: Diagnose failure group 1
├─ Agent B: Diagnose failure group 2
└─ Agent C: Diagnose failure group 3

Phase 3: FAN-OUT (Parallel fixes)
├─ Agent A: Fix test group 1
├─ Agent B: Fix test group 2
└─ Agent C: Fix test group 3

Phase 4: PIPELINE
└─ Background agent: Verify all tests pass

Pattern: Test Refactoring

User Request: "Clean up test duplication"

Phase 1: EXPLORE
└─ Explore agent: Find duplicate test patterns

Phase 2: PLAN
└─ Plan agent: Design shared fixtures, helpers, patterns

Phase 3: FAN-OUT
├─ Agent A: Extract shared fixtures
├─ Agent B: Refactor test file group 1
└─ Agent C: Refactor test file group 2

Phase 4: PIPELINE
└─ Background agent: Verify tests still pass

---

E2E Testing

Pattern: User Journey Testing

User Request: "Add E2E tests for user registration"

Phase 1: EXPLORE
└─ Explore agent: Map registration flow, identify test scenarios

Phase 2: PIPELINE (Sequential scenarios)
├─ General-purpose agent: Happy path registration
├─ General-purpose agent: Validation error scenarios
├─ General-purpose agent: Duplicate email handling
└─ General-purpose agent: Email verification flow

Phase 3: BACKGROUND
└─ Background agent: Run E2E suite, capture screenshots

---

Task Management for Testing

Structure testing work as tasks with clear dependencies:

# Create testing tasks
npx cc-mirror tasks create --subject "Identify testing scope" --description "Analyze what needs testing..."
npx cc-mirror tasks create --subject "Generate unit tests" --description "Tests for module A..."
npx cc-mirror tasks create --subject "Generate integration tests" --description "Tests for API endpoints..."
npx cc-mirror tasks create --subject "Run test suite" --description "Execute all tests, capture results..."
npx cc-mirror tasks create --subject "Fix failures" --description "Address any failing tests..."
npx cc-mirror tasks create --subject "Verify all pass" --description "Final test run to confirm..."

# Dependencies
npx cc-mirror tasks update 2 --add-blocked-by 1
npx cc-mirror tasks update 3 --add-blocked-by 1
npx cc-mirror tasks update 4 --add-blocked-by 2,3
npx cc-mirror tasks update 5 --add-blocked-by 4
npx cc-mirror tasks update 6 --add-blocked-by 5

# Parallel test generation (sonnet for well-structured work)
Task(subagent_type="general-purpose", prompt="Task 2: Generate unit tests...",
     model="sonnet", run_in_background=True)
Task(subagent_type="general-purpose", prompt="Task 3: Generate integration tests...",
     model="sonnet", run_in_background=True)

Test Execution Best Practices

1. Always run in background for long test suites
2. Parallelize independent suites (unit, integration, e2e)
3. Fail fast - stop on first failure for quick feedback
4. Capture artifacts - screenshots, logs, coverage reports
5. Report actionable results - file:line for failures

---

─── ◈ Testing ───────────────────────────