@bugzy-ai/bugzy 1.8.0 → 1.9.0

package/templates/init/.bugzy/runtime/test-execution-strategy.md

@@ -1,535 +1,535 @@
-# Test Execution Strategy
-
-## Overview
-
-This document defines the test execution strategy for the Bugzy QA system, explaining **which tests to run when** and **why**. All agents and commands executing tests should reference this document to make intelligent decisions about test selection.
-
-## Test Architecture Pattern
-
-Our test suite follows a **pyramid testing strategy** where test cases can have multiple test implementations:
-
-- **Comprehensive E2E Tests**: Full user journey from start to finish (tagged `@smoke`)
-- **Granular Component Tests**: Isolated validation of specific capabilities (tagged `@component` or `@fast`)
-
-**Example**: TC-003 (Project Settings Update) has:
-- 1 comprehensive E2E test covering all 20 manual steps
-- 2 granular tests covering navigation (steps 1-8) and field display (steps 1-10)
-
-## Test Tiers
-
-### Tier 1: Smoke Tests (`@smoke`)
-
-**Purpose**: Quick validation of critical user journeys
-
-**Coverage**:
-- Comprehensive E2E tests only
-- One test per test case
-- 100% manual test case coverage
-- **Zero redundancy**
-
-**Execution Time**: ~2-5 minutes for full smoke suite
-
-**When to Use**:
-- **Default for most test runs**
-- Every commit and pull request
-- Quick validation after code changes
-- Before starting new development work
-- Continuous integration (CI) on every push
-- Post-deployment verification
-
-**Command**:
-```bash
-npx playwright test --grep "@smoke"
-```
-
-**Why This Tier**:
-- Fast feedback loop
-- Full coverage of critical paths
-- No wasted execution time on redundant tests
-- Ideal for frequent validation
-
----
-
-### Tier 2: Component Tests (`@component` or `@fast`)
-
-**Purpose**: Isolated component validation for debugging
-
-**Coverage**:
-- Granular tests for specific capabilities only
-- Navigation, form display, API calls, etc.
-- Subset coverage (e.g., just steps 1-8 for navigation)
-- Skips comprehensive E2E tests
-
-**Execution Time**: ~1-3 minutes (faster than smoke)
-
-**When to Use**:
-- Debugging specific component failures
-- After refactoring a specific component
-- When you need to verify a single capability without running full E2E
-- Rapid iteration during development
-- Targeted validation after component-level bug fixes
-
-**Command**:
-```bash
-npx playwright test --grep "@component"
-# OR
-npx playwright test --grep "@fast"
-```
-
-**Why This Tier**:
-- Ultra-fast feedback for specific areas
-- Avoids running unnecessary setup/teardown
-- Ideal for focused development work
-
----
-
-### Tier 3: Full Regression (All Tests)
-
-**Purpose**: Complete validation including redundant component tests
-
-**Coverage**:
-- Smoke tests (comprehensive E2E)
-- Component tests (granular isolation)
-- **Intentional redundancy** for diagnostic value
-
-**Execution Time**: ~10-15 minutes
-
-**When to Use**:
-- Weekly scheduled CI runs
-- Before major releases or deployments
-- After significant refactoring across multiple areas
-- When comprehensive validation is required
-- Pre-merge validation for large features
-- Monthly regression testing cycles
-
-**Command**:
-```bash
-npx playwright test
-# OR explicitly
-npx playwright test --grep ".*"
-```
-
-**Why This Tier**:
-- **Fast failure isolation**: If Settings page breaks, component test fails in 15 seconds vs E2E failing after 30 seconds
-- **Component-level confidence**: Proves each capability works independently
-- **Refactoring safety**: If you refactor save logic, component tests prove navigation/display still work
-- **Comprehensive validation**: Catches edge cases that isolated tests might miss
-
-**Understanding Redundancy**:
-- Example: TC-003 login steps (1-5) run in all 3 tests
-- **Cost**: Extra 30 seconds execution time
-- **Benefit**: Instant failure location (know exactly which component broke)
-- **Trade-off**: Worth it for regression scenarios, not for daily validation
-
----
-
-## Default Behavior
-
-### When User Provides No Selector
-
-**Default Action**: Run `@smoke` tests
-
-**Rationale**:
-- Provides fast feedback (2-5 min vs 10-15 min)
-- Full coverage of critical user journeys
-- Zero redundancy = no wasted time
-- Suitable for 90% of test execution scenarios
-
-**User Communication**:
-```
-Running smoke tests (@smoke) for quick validation.
-For comprehensive regression including component tests, use 'all'.
-For specific areas, provide test file pattern or tag.
-```
-
-### When User Says "All"
-
-**Action**: Run full regression suite (all tests)
-
-**User Communication**:
-```
-Running full regression suite including component tests.
-This includes intentional redundancy for diagnostic value and will take longer (~10-15 min).
-```
-
----
-
-## Context-Based Selection Logic
-
-Parse user intent from their natural language request:
-
-### Fast Validation Context
-
-**Keywords**: "quick", "validate", "sanity check", "verify", "check"
-
-**Action**: Run `@smoke` tests
-
-**Examples**:
-- "Quick check if login still works"
-- "Validate the settings page"
-- "Sanity test before I start work"
-
----
-
-### Comprehensive Validation Context
-
-**Keywords**: "full", "comprehensive", "regression", "all", "everything", "complete"
-
-**Action**: Run all tests
-
-**Examples**:
-- "Full regression before release"
-- "Comprehensive validation of the system"
-- "Run everything to be sure"
-
----
-
-### Debugging/Focused Context
-
-**Keywords**: "debug", "check [specific area]", "[component name]", "just [feature]"
-
-**Action**: Run specific test file or pattern
-
-**Examples**:
-- "Debug the settings page" → Run `TC-003` or pattern `settings`
-- "Check authentication only" → Run pattern `auth` or `TC-001`
-- "Just test the navigation" → Run `@component` tests
-
----
-
-### Ambiguous Context
-
-**When**: No clear keywords or selector provided
-
-**Action**: Default to `@smoke` and inform user of options
-
-**User Communication**:
-```
-No test selection specified. Running smoke tests (@smoke) by default.
-
-Available options:
-- '@smoke' - Quick validation (2-5 min, default)
-- '@component' - Component tests only (1-3 min)
-- 'all' - Full regression (10-15 min)
-- '[pattern]' - Specific tests matching pattern
-```
-
----
-
-## Tag Taxonomy
-
-### Primary Tags
-
-| Tag | Purpose | Coverage | Speed |
-|-----|---------|----------|-------|
-| `@smoke` | Critical E2E journeys | Comprehensive, 100% manual test coverage | Medium (30-60s/test) |
-| `@component` | Isolated component validation | Partial, specific capability only | Fast (10-20s/test) |
-| `@fast` | Alias for `@component` | Same as @component | Fast |
-| `@regression` | Full regression suite | All tests (smoke + component) | Slow (full suite) |
-
-### Domain Tags (for filtering by feature area)
-
-| Tag | Feature Area |
-|-----|--------------|
-| `@authentication` | Login, logout, session management |
-| `@project-management` | Project CRUD, settings |
-| `@test-management` | Test cases, test runs, activity feed |
-| `@settings` | Settings pages and persistence |
-| `@navigation` | Navigation and routing |
-| `@data-persistence` | Data saving and loading |
-
-### Combining Tags
-
-**Multiple tags on same test**:
-```typescript
-test('should update settings and persist @smoke @project-management @settings', ...)
-```
-
-**Run tests matching multiple criteria**:
-```bash
-# Smoke tests in project management area
-npx playwright test --grep "@smoke.*@project-management"
-
-# All settings-related tests (smoke + component)
-npx playwright test --grep "@settings"
-```
-
----
-
-## Time and Coverage Trade-offs
-
-### Execution Time Comparison
-
-| Tier | Tests Run | Time | Use Frequency |
-|------|-----------|------|---------------|
-| Smoke (`@smoke`) | ~6 tests | 2-5 min | Daily (90%) |
-| Component (`@component`) | ~8 tests | 1-3 min | As needed (5%) |
-| Full Regression (all) | ~15 tests | 10-15 min | Weekly (5%) |
-
-### Coverage vs Speed Matrix
-
-```
-High Coverage ┐
-              │
-              │        ┌─ Full Regression
-              │        │  (High coverage, slower)
-              │        │
-              │   ┌────┘
-              │   │
-              │   │ Smoke Tests
-              │   │ (High coverage, faster)
-              │   │
-              ├───┼─────────────────────►
-              │   │                    Fast
-              │   │
-              │   └─ Component Tests
-              │      (Partial coverage, fastest)
-              │
-Low Coverage  ┘
-```
-
-### Recommendation by Scenario
-
-| Scenario | Recommended Tier | Rationale |
-|----------|------------------|-----------|
-| Developer commit | Smoke | Fast feedback, full coverage |
-| Pull request CI | Smoke | Balance speed and confidence |
-| Pre-deployment | Full Regression | Comprehensive validation |
-| Debugging component | Component or specific file | Fastest feedback for focused area |
-| Weekly CI schedule | Full Regression | Diagnostic redundancy value |
-| Post-hotfix | Smoke | Quick validation of critical paths |
-
----
-
-## Example Scenarios
-
-### Scenario 1: Developer Makes Code Change
-
-**Context**: Developer updates Settings page form validation
-
-**User Request**: `/run-tests` (no selector)
-
-**Agent Decision**:
-- Read this strategy file
-- No selector provided → Default to `@smoke`
-- Run: `npx playwright test --grep "@smoke"`
-- Result: 6 tests in ~3 min, full E2E coverage, zero redundancy
-
-**Communication**:
-```
-Running smoke tests (@smoke) for quick validation...
-✓ 6/6 tests passed in 3m 12s
-All critical user journeys validated successfully.
-```
-
----
-
-### Scenario 2: Pre-Release Validation
-
-**User Request**: `/run-tests all`
-
-**Agent Decision**:
-- Explicit "all" keyword → Run full regression
-- Run: `npx playwright test`
-- Result: 15 tests in ~12 min, includes redundant component tests
-
-**Communication**:
-```
-Running full regression suite including component tests...
-This includes intentional redundancy for diagnostic value (~10-15 min).
-
-✓ 15/15 tests passed in 12m 34s
-Full regression validation complete:
-- 6 comprehensive E2E tests (smoke)
-- 9 component-level tests
-All tests passed.
-```
-
----
-
-### Scenario 3: Debugging Settings Page Failure
-
-**User Request**: `/run-tests settings`
-
-**Agent Decision**:
-- File pattern provided → Match specific tests
-- Run: `npx playwright test settings`
-- Result: Runs TC-003 (3 tests: 1 smoke + 2 component)
-
-**Communication**:
-```
-Running tests matching pattern "settings"...
-Found: tests/specs/project-management/TC-003-project-settings-update.spec.ts
-
-Results:
-✓ should update project settings and persist changes @smoke
-✗ should navigate to settings page @component
-  └─ Error: Navigation timeout
-✓ should display project name and description @component
-
-Failure isolated to navigation component. Settings form display works correctly.
-```
-
----
-
-### Scenario 4: Quick Validation Request
-
-**User Request**: "Can you quickly check if the login flow still works?"
-
-**Agent Decision**:
-- Keyword "quickly" → Use `@smoke`
-- Implicit focus on "login" → Could run specific test or full smoke
-- Run: `npx playwright test --grep "@smoke"` (include all smoke, not just login)
-- Result: 6 tests including login in ~3 min
-
-**Communication**:
-```
-Running smoke tests (@smoke) for quick validation of all critical flows including login...
-
-✓ 6/6 tests passed in 3m 8s
-All critical user journeys validated, including:
-- TC-001: Login and navigation ✓
-- TC-003: Project settings update ✓
-- TC-004: Test cases page access ✓
-- TC-005: Test runs page access ✓
-- TC-006: Session management ✓
-- TC-008: Activity feed access ✓
-```
-
----
-
-## Understanding Test Redundancy
-
-### What is Redundant Coverage?
-
-Some test suites contain multiple tests that exercise the same functionality:
-
-**Example: TC-003 (Project Settings Update)**
-
-```
-Test 1 (E2E, @smoke): Steps 1-20 (login → navigate → update → verify)
-Test 2 (Component, @component): Steps 1-8 (login → navigate)  ← Overlaps with Test 1
-Test 3 (Component, @component): Steps 1-10 (login → navigate → display) ← Overlaps with Test 1
-```
-
-**Overlap**: Steps 1-8 run in all 3 tests
-
-### Why Allow Redundancy?
-
-**Diagnostic Value**:
-- If Test 2 fails but Test 1 passes → Issue is isolated to navigation
-- If Test 3 fails but Tests 1 & 2 pass → Issue is in form display
-- If Test 1 fails but Tests 2 & 3 pass → Issue is in update/save logic (steps 11-20)
-
-**Speed Benefit**:
-- Test 2 fails in 15 seconds (just navigation)
-- Test 1 fails in 30 seconds (full E2E including navigation)
-- **Saved time**: 15 seconds to identify failure point
-
-### When is Redundancy Worth It?
-
-**Smoke Runs**: NOT worth it (90% of executions)
-- Goal: Fast validation
-- Trade-off: Skip redundant tests, accept slower failure diagnosis
-
-**Full Regression Runs**: Worth it (10% of executions)
-- Goal: Comprehensive validation with diagnostic value
-- Trade-off: Accept 30 extra seconds for faster failure isolation
-- Context: Pre-release, weekly CI, post-refactoring
-
----
-
-## Integration with Other Systems
-
-### Test Generation (`/generate-test-cases`)
-
-When generating new test cases, the generator should:
-1. Create comprehensive E2E test with `@smoke` tag
-2. Optionally create component tests with `@component` or `@fast` tags
-3. Apply appropriate domain tags (`@authentication`, `@project-management`, etc.)
-
-### CI/CD Pipelines
-
-Recommended pipeline configuration:
-
-```yaml
-# On every PR - SMOKE ONLY
-on_pull_request:
-  npx playwright test --grep "@smoke"
-  # Fast validation: 2-5 min
-
-# Nightly regression - ALL TESTS
-on_schedule:
-  npx playwright test
-  # Comprehensive validation: 10-15 min
-
-# Pre-deployment - SMOKE (fast gate)
-on_deployment:
-  npx playwright test --grep "@smoke"
-  # Critical path validation: 2-5 min
-```
-
-### Issue Tracking
-
-When tests fail, triage results should consider:
-- Smoke test failures → High priority (critical path broken)
-- Component test failures (when smoke passes) → Medium priority (isolated issue)
-- Component test failures (when smoke also fails) → Confirms root cause
-
----
-
-## Decision Tree for Agents
-
-```
-User provides selector?
-├─ Yes
-│  ├─ Selector is "all" or "everything"?
-│  │  └─ Run: npx playwright test (Full Regression)
-│  ├─ Selector is tag (starts with @)?
-│  │  └─ Run: npx playwright test --grep "[tag]"
-│  ├─ Selector is file pattern (e.g., "auth", "settings")?
-│  │  └─ Run: npx playwright test [pattern]
-│  └─ Selector is specific file path?
-│     └─ Run: npx playwright test [file-path]
-└─ No selector provided
-   ├─ User request contains context keywords?
-   │  ├─ "quick", "validate", "sanity" → Run: @smoke
-   │  ├─ "full", "comprehensive", "regression" → Run: all tests
-   │  └─ "debug [area]", "check [component]" → Run: specific pattern
-   └─ No context keywords
-      └─ Default: Run @smoke and inform user of options
-```
-
----
-
-## Maintenance Notes
-
-### When to Update This Strategy
-
-This document should be updated when:
-- New test tiers are introduced
-- Tag taxonomy changes
-- Execution time significantly changes
-- New testing patterns emerge
-- CI/CD pipeline requirements change
-
-### Related Documents
-
-- `.bugzy/runtime/testing-best-practices.md` - How to write tests (test patterns)
-- `.bugzy/runtime/knowledge-base.md` - Accumulated testing insights
-- `.claude/commands/run-tests.md` - Test execution command (mechanics)
-- `playwright.config.ts` - Playwright configuration
-- Test cases in `./test-cases/` - Individual test case documentation
-
----
-
-## Summary
-
-**Key Principles**:
-1. **Default to smoke tests** for fast validation (90% of runs)
-2. **Use component tests** for focused debugging
-3. **Run full regression** only when diagnostic redundancy is needed (10% of runs)
-4. **Parse user context** to choose the right tier intelligently
-5. **Communicate clearly** about what's running and why
-
-**Remember**: The goal is to provide fast feedback most of the time, while maintaining the ability to run comprehensive validation when needed. Redundancy is intentional but conditional.
+# Test Execution Strategy
+
+## Overview
+
+This document defines the test execution strategy for the Bugzy QA system, explaining **which tests to run when** and **why**. All agents and commands executing tests should reference this document to make intelligent decisions about test selection.
+
+## Test Architecture Pattern
+
+Our test suite follows a **pyramid testing strategy** where test cases can have multiple test implementations:
+
+- **Comprehensive E2E Tests**: Full user journey from start to finish (tagged `@smoke`)
+- **Granular Component Tests**: Isolated validation of specific capabilities (tagged `@component` or `@fast`)
+
+**Example**: TC-003 (Project Settings Update) has:
+- 1 comprehensive E2E test covering all 20 manual steps
+- 2 granular tests covering navigation (steps 1-8) and field display (steps 1-10)
+
+## Test Tiers
+
+### Tier 1: Smoke Tests (`@smoke`)
+
+**Purpose**: Quick validation of critical user journeys
+
+**Coverage**:
+- Comprehensive E2E tests only
+- One test per test case
+- 100% manual test case coverage
+- **Zero redundancy**
+
+**Execution Time**: ~2-5 minutes for full smoke suite
+
+**When to Use**:
+- **Default for most test runs**
+- Every commit and pull request
+- Quick validation after code changes
+- Before starting new development work
+- Continuous integration (CI) on every push
+- Post-deployment verification
+
+**Command**:
+```bash
+npx playwright test --grep "@smoke"
+```
+
+**Why This Tier**:
+- Fast feedback loop
+- Full coverage of critical paths
+- No wasted execution time on redundant tests
+- Ideal for frequent validation
+
+---
+
+### Tier 2: Component Tests (`@component` or `@fast`)
+
+**Purpose**: Isolated component validation for debugging
+
+**Coverage**:
+- Granular tests for specific capabilities only
+- Navigation, form display, API calls, etc.
+- Subset coverage (e.g., just steps 1-8 for navigation)
+- Skips comprehensive E2E tests
+
+**Execution Time**: ~1-3 minutes (faster than smoke)
+
+**When to Use**:
+- Debugging specific component failures
+- After refactoring a specific component
+- When you need to verify a single capability without running full E2E
+- Rapid iteration during development
+- Targeted validation after component-level bug fixes
+
+**Command**:
+```bash
+npx playwright test --grep "@component"
+# OR
+npx playwright test --grep "@fast"
+```
+
+**Why This Tier**:
+- Ultra-fast feedback for specific areas
+- Avoids running unnecessary setup/teardown
+- Ideal for focused development work
+
+---
+
+### Tier 3: Full Regression (All Tests)
+
+**Purpose**: Complete validation including redundant component tests
+
+**Coverage**:
+- Smoke tests (comprehensive E2E)
+- Component tests (granular isolation)
+- **Intentional redundancy** for diagnostic value
+
+**Execution Time**: ~10-15 minutes
+
+**When to Use**:
+- Weekly scheduled CI runs
+- Before major releases or deployments
+- After significant refactoring across multiple areas
+- When comprehensive validation is required
+- Pre-merge validation for large features
+- Monthly regression testing cycles
+
+**Command**:
+```bash
+npx playwright test
+# OR explicitly
+npx playwright test --grep ".*"
+```
+
+**Why This Tier**:
+- **Fast failure isolation**: If Settings page breaks, component test fails in 15 seconds vs E2E failing after 30 seconds
+- **Component-level confidence**: Proves each capability works independently
+- **Refactoring safety**: If you refactor save logic, component tests prove navigation/display still work
+- **Comprehensive validation**: Catches edge cases that isolated tests might miss
+
+**Understanding Redundancy**:
+- Example: TC-003 login steps (1-5) run in all 3 tests
+- **Cost**: Extra 30 seconds execution time
+- **Benefit**: Instant failure location (know exactly which component broke)
+- **Trade-off**: Worth it for regression scenarios, not for daily validation
+
+---
+
+## Default Behavior
+
+### When User Provides No Selector
+
+**Default Action**: Run `@smoke` tests
+
+**Rationale**:
+- Provides fast feedback (2-5 min vs 10-15 min)
+- Full coverage of critical user journeys
+- Zero redundancy = no wasted time
+- Suitable for 90% of test execution scenarios
+
+**User Communication**:
+```
+Running smoke tests (@smoke) for quick validation.
+For comprehensive regression including component tests, use 'all'.
+For specific areas, provide test file pattern or tag.
+```
+
+### When User Says "All"
+
+**Action**: Run full regression suite (all tests)
+
+**User Communication**:
+```
+Running full regression suite including component tests.
+This includes intentional redundancy for diagnostic value and will take longer (~10-15 min).
+```
+
+---
+
+## Context-Based Selection Logic
+
+Parse user intent from their natural language request:
+
+### Fast Validation Context
+
+**Keywords**: "quick", "validate", "sanity check", "verify", "check"
+
+**Action**: Run `@smoke` tests
+
+**Examples**:
+- "Quick check if login still works"
+- "Validate the settings page"
+- "Sanity test before I start work"
+
+---
+
+### Comprehensive Validation Context
+
+**Keywords**: "full", "comprehensive", "regression", "all", "everything", "complete"
+
+**Action**: Run all tests
+
+**Examples**:
+- "Full regression before release"
+- "Comprehensive validation of the system"
+- "Run everything to be sure"
+
+---
+
+### Debugging/Focused Context
+
+**Keywords**: "debug", "check [specific area]", "[component name]", "just [feature]"
+
+**Action**: Run specific test file or pattern
+
+**Examples**:
+- "Debug the settings page" → Run `TC-003` or pattern `settings`
+- "Check authentication only" → Run pattern `auth` or `TC-001`
+- "Just test the navigation" → Run `@component` tests
+
+---
+
+### Ambiguous Context
+
+**When**: No clear keywords or selector provided
+
+**Action**: Default to `@smoke` and inform user of options
+
+**User Communication**:
+```
+No test selection specified. Running smoke tests (@smoke) by default.
+
+Available options:
+- '@smoke' - Quick validation (2-5 min, default)
+- '@component' - Component tests only (1-3 min)
+- 'all' - Full regression (10-15 min)
+- '[pattern]' - Specific tests matching pattern
+```
+
+---
+
+## Tag Taxonomy
+
+### Primary Tags
+
+| Tag | Purpose | Coverage | Speed |
+|-----|---------|----------|-------|
+| `@smoke` | Critical E2E journeys | Comprehensive, 100% manual test coverage | Medium (30-60s/test) |
+| `@component` | Isolated component validation | Partial, specific capability only | Fast (10-20s/test) |
+| `@fast` | Alias for `@component` | Same as @component | Fast |
+| `@regression` | Full regression suite | All tests (smoke + component) | Slow (full suite) |
+
+### Domain Tags (for filtering by feature area)
+
+| Tag | Feature Area |
+|-----|--------------|
+| `@authentication` | Login, logout, session management |
+| `@project-management` | Project CRUD, settings |
+| `@test-management` | Test cases, test runs, activity feed |
+| `@settings` | Settings pages and persistence |
+| `@navigation` | Navigation and routing |
+| `@data-persistence` | Data saving and loading |
+
+### Combining Tags
+
+**Multiple tags on same test**:
+```typescript
+test('should update settings and persist @smoke @project-management @settings', ...)
+```
+
+**Run tests matching multiple criteria**:
+```bash
+# Smoke tests in project management area
+npx playwright test --grep "@smoke.*@project-management"
+
+# All settings-related tests (smoke + component)
+npx playwright test --grep "@settings"
+```
+
+---
+
+## Time and Coverage Trade-offs
+
+### Execution Time Comparison
+
+| Tier | Tests Run | Time | Use Frequency |
+|------|-----------|------|---------------|
+| Smoke (`@smoke`) | ~6 tests | 2-5 min | Daily (90%) |
+| Component (`@component`) | ~8 tests | 1-3 min | As needed (5%) |
+| Full Regression (all) | ~15 tests | 10-15 min | Weekly (5%) |
+
+### Coverage vs Speed Matrix
+
+```
+High Coverage ┐
+              │
+              │        ┌─ Full Regression
+              │        │  (High coverage, slower)
+              │        │
+              │   ┌────┘
+              │   │
+              │   │ Smoke Tests
+              │   │ (High coverage, faster)
+              │   │
+              ├───┼─────────────────────►
+              │   │                    Fast
+              │   │
+              │   └─ Component Tests
+              │      (Partial coverage, fastest)
+              │
+Low Coverage  ┘
+```
+
+### Recommendation by Scenario
+
+| Scenario | Recommended Tier | Rationale |
+|----------|------------------|-----------|
+| Developer commit | Smoke | Fast feedback, full coverage |
+| Pull request CI | Smoke | Balance speed and confidence |
+| Pre-deployment | Full Regression | Comprehensive validation |
+| Debugging component | Component or specific file | Fastest feedback for focused area |
+| Weekly CI schedule | Full Regression | Diagnostic redundancy value |
+| Post-hotfix | Smoke | Quick validation of critical paths |
+
+---
+
+## Example Scenarios
+
+### Scenario 1: Developer Makes Code Change
+
+**Context**: Developer updates Settings page form validation
+
+**User Request**: `/run-tests` (no selector)
+
+**Agent Decision**:
+- Read this strategy file
+- No selector provided → Default to `@smoke`
+- Run: `npx playwright test --grep "@smoke"`
+- Result: 6 tests in ~3 min, full E2E coverage, zero redundancy
+
+**Communication**:
+```
+Running smoke tests (@smoke) for quick validation...
+✓ 6/6 tests passed in 3m 12s
+All critical user journeys validated successfully.
+```
+
+---
+
+### Scenario 2: Pre-Release Validation
+
+**User Request**: `/run-tests all`
+
+**Agent Decision**:
+- Explicit "all" keyword → Run full regression
+- Run: `npx playwright test`
+- Result: 15 tests in ~12 min, includes redundant component tests
+
+**Communication**:
+```
+Running full regression suite including component tests...
+This includes intentional redundancy for diagnostic value (~10-15 min).
+
+✓ 15/15 tests passed in 12m 34s
+Full regression validation complete:
+- 6 comprehensive E2E tests (smoke)
+- 9 component-level tests
+All tests passed.
+```
+
+---
+
+### Scenario 3: Debugging Settings Page Failure
+
+**User Request**: `/run-tests settings`
+
+**Agent Decision**:
+- File pattern provided → Match specific tests
+- Run: `npx playwright test settings`
+- Result: Runs TC-003 (3 tests: 1 smoke + 2 component)
+
+**Communication**:
+```
+Running tests matching pattern "settings"...
+Found: tests/specs/project-management/TC-003-project-settings-update.spec.ts
+
+Results:
+✓ should update project settings and persist changes @smoke
+✗ should navigate to settings page @component
+  └─ Error: Navigation timeout
+✓ should display project name and description @component
+
+Failure isolated to navigation component. Settings form display works correctly.
+```
+
+---
+
+### Scenario 4: Quick Validation Request
+
+**User Request**: "Can you quickly check if the login flow still works?"
+
+**Agent Decision**:
+- Keyword "quickly" → Use `@smoke`
+- Implicit focus on "login" → Could run specific test or full smoke
+- Run: `npx playwright test --grep "@smoke"` (include all smoke, not just login)
+- Result: 6 tests including login in ~3 min
+
+**Communication**:
+```
+Running smoke tests (@smoke) for quick validation of all critical flows including login...
+
+✓ 6/6 tests passed in 3m 8s
+All critical user journeys validated, including:
+- TC-001: Login and navigation ✓
+- TC-003: Project settings update ✓
+- TC-004: Test cases page access ✓
+- TC-005: Test runs page access ✓
+- TC-006: Session management ✓
+- TC-008: Activity feed access ✓
+```
+
+---
+
+## Understanding Test Redundancy
+
+### What is Redundant Coverage?
+
+Some test suites contain multiple tests that exercise the same functionality:
+
+**Example: TC-003 (Project Settings Update)**
+
+```
+Test 1 (E2E, @smoke): Steps 1-20 (login → navigate → update → verify)
+Test 2 (Component, @component): Steps 1-8 (login → navigate)  ← Overlaps with Test 1
+Test 3 (Component, @component): Steps 1-10 (login → navigate → display) ← Overlaps with Test 1
+```
+
+**Overlap**: Steps 1-8 run in all 3 tests
+
+### Why Allow Redundancy?
+
+**Diagnostic Value**:
+- If Test 2 fails but Test 1 passes → Issue is isolated to navigation
+- If Test 3 fails but Tests 1 & 2 pass → Issue is in form display
+- If Test 1 fails but Tests 2 & 3 pass → Issue is in update/save logic (steps 11-20)
+
+**Speed Benefit**:
+- Test 2 fails in 15 seconds (just navigation)
+- Test 1 fails in 30 seconds (full E2E including navigation)
+- **Saved time**: 15 seconds to identify failure point
+
+### When is Redundancy Worth It?
+
+**Smoke Runs**: NOT worth it (90% of executions)
+- Goal: Fast validation
+- Trade-off: Skip redundant tests, accept slower failure diagnosis
+
+**Full Regression Runs**: Worth it (10% of executions)
+- Goal: Comprehensive validation with diagnostic value
+- Trade-off: Accept 30 extra seconds for faster failure isolation
+- Context: Pre-release, weekly CI, post-refactoring
+
+---
+
+## Integration with Other Systems
+
+### Test Generation (`/generate-test-cases`)
+
+When generating new test cases, the generator should:
+1. Create comprehensive E2E test with `@smoke` tag
+2. Optionally create component tests with `@component` or `@fast` tags
+3. Apply appropriate domain tags (`@authentication`, `@project-management`, etc.)
+
+### CI/CD Pipelines
+
+Recommended pipeline configuration:
+
+```yaml
+# On every PR - SMOKE ONLY
+on_pull_request:
+  npx playwright test --grep "@smoke"
+  # Fast validation: 2-5 min
+
+# Nightly regression - ALL TESTS
+on_schedule:
+  npx playwright test
+  # Comprehensive validation: 10-15 min
+
+# Pre-deployment - SMOKE (fast gate)
+on_deployment:
+  npx playwright test --grep "@smoke"
+  # Critical path validation: 2-5 min
+```
+
+### Issue Tracking
+
+When tests fail, triage results should consider:
+- Smoke test failures → High priority (critical path broken)
+- Component test failures (when smoke passes) → Medium priority (isolated issue)
+- Component test failures (when smoke also fails) → Confirms root cause
+
+---
+
+## Decision Tree for Agents
+
+```
+User provides selector?
+├─ Yes
+│  ├─ Selector is "all" or "everything"?
+│  │  └─ Run: npx playwright test (Full Regression)
+│  ├─ Selector is tag (starts with @)?
+│  │  └─ Run: npx playwright test --grep "[tag]"
+│  ├─ Selector is file pattern (e.g., "auth", "settings")?
+│  │  └─ Run: npx playwright test [pattern]
+│  └─ Selector is specific file path?
+│     └─ Run: npx playwright test [file-path]
+└─ No selector provided
+   ├─ User request contains context keywords?
+   │  ├─ "quick", "validate", "sanity" → Run: @smoke
+   │  ├─ "full", "comprehensive", "regression" → Run: all tests
+   │  └─ "debug [area]", "check [component]" → Run: specific pattern
+   └─ No context keywords
+      └─ Default: Run @smoke and inform user of options
+```
+
+---
+
+## Maintenance Notes
+
+### When to Update This Strategy
+
+This document should be updated when:
+- New test tiers are introduced
+- Tag taxonomy changes
+- Execution time significantly changes
+- New testing patterns emerge
+- CI/CD pipeline requirements change
+
+### Related Documents
+
+- `.bugzy/runtime/testing-best-practices.md` - How to write tests (test patterns)
+- `.bugzy/runtime/knowledge-base.md` - Accumulated testing insights
+- `.claude/commands/run-tests.md` - Test execution command (mechanics)
+- `playwright.config.ts` - Playwright configuration
+- Test cases in `./test-cases/` - Individual test case documentation
+
+---
+
+## Summary
+
+**Key Principles**:
+1. **Default to smoke tests** for fast validation (90% of runs)
+2. **Use component tests** for focused debugging
+3. **Run full regression** only when diagnostic redundancy is needed (10% of runs)
+4. **Parse user context** to choose the right tier intelligently
+5. **Communicate clearly** about what's running and why
+
+**Remember**: The goal is to provide fast feedback most of the time, while maintaining the ability to run comprehensive validation when needed. Redundancy is intentional but conditional.