Adding Testing Phases

When to Use

After /create-task when you want to add UI testing/polishing phases to verify frontend work.

Execution Flow

code

1. Load existing task
2. Analyze phases → find frontend-touching ones
3. Determine granularity (flexible)
4. Generate testing phase(s) with test scenarios
5. Insert into task structure
6. Update README.md and CONTEXT.md

Step 1: Load Task

Read:

•docs/tasks/{task-name}/README.md
•docs/tasks/{task-name}/CONTEXT.md
•All phase files (0*.md)

Extract:

•App name
•Phase list with descriptions
•Success criteria (will become test scenarios)
•Which phases touch frontend

Step 2: Identify Frontend-Touching Phases

Indicators:

•Phase name: UI, Frontend, Component, Page, Polish
•Files mention: frontend/, .tsx, .vue, components/
•Success criteria: visible, click, form, navigation, user sees, displays

Mark each phase:

code

01-backend.md      → implementation
02-api.md          → implementation
03-frontend-ui.md  → FRONTEND ✓
04-polish.md       → FRONTEND ✓

Step 3: Determine Granularity

Frontend Phases	Strategy
1-4	One testing phase at end
5+	Multiple testing phases, grouped by feature

Simple case (most common): Add single testing phase after all implementation.

Step 4: Generate Testing Phase

Convert Success Criteria to Test Scenarios

From phase success criteria:

markdown

- [ ] User can see their role on admin page
- [ ] Protected action shows feedback

To test scenarios:

markdown

### Scenario 1: Role Display
**Steps:**
1. Navigate to `/admin`
2. Verify "Your Access" card visible
3. Verify role text shows correctly

**Pass:** Role displayed matches user's actual role

### Scenario 2: Protected Action Feedback
**Steps:**
1. On admin page, click "Execute Protected Action"
2. Verify feedback appears

**Pass:** Shows success or "access denied" message

Phase Template

markdown

# Phase {N}: Testing & Polish

**Type:** testing-polish
**Depends on:** Phase {N-1}

**Objective:** Verify and fix {covered features} until working

---

## Test Scenarios

### Scenario 1: {Name}
**Steps:**
1. {action}
2. {verification}

**Pass:** {clear success state}

### Scenario 2: {Name}
...

---

## Verification Checklist

- [ ] No console errors
- [ ] All API calls succeed
- [ ] Visual appearance correct

---

## Fix Protocol

For each scenario:
1. Execute via Playwright
2. If PASS → next
3. If FAIL → diagnose → fix → retest (up to 3x)
4. After 3 failures → document in KNOWN_ISSUES.md

**Reference:** `.claude/shared/playwright-testing.md`

Step 5: Insert Phase File

Write to: docs/tasks/{task-name}/{NN}-testing-polish.md

Numbering: Use next available number after last implementation phase.

Step 6: Update Task Files

Update README.md

Add to Phases Overview table:

markdown

| {N} | Testing & Polish | testing-polish |

Update CONTEXT.md

Add note:

markdown

## Testing Phases

- Phase {N}: Testing & Polish (added via /add-testing)
- Type: testing-polish
- Agent: testing-polish-agent

Output Format

code

Testing phases added: docs/tasks/{task-name}/

Analyzed:
- Total phases: {N}
- Frontend-touching: {M}

Added:
- {NN}-testing-polish.md

Test scenarios ({count}):
1. {scenario 1}
2. {scenario 2}
...

Next: /execute-task {task-name}

Checklist

code

Adding Testing Phases:
- [ ] Read README.md + CONTEXT.md + all phase files
- [ ] Identified frontend-touching phases
- [ ] Determined granularity (usually: one at end)
- [ ] Converted success criteria to test scenarios
- [ ] Wrote testing-polish phase file
- [ ] Updated README.md phases table
- [ ] Updated CONTEXT.md with note
- [ ] Reported to user