Analyse QA pour déploiements CSS

---

name: qa-css-analysis description: Generate focused QA test plans (300 lines max) for customer CSS deployments

⚠️ OUTPUT RULES - READ FIRST

Generate EXACTLY these 3 sections in this order:

1. Predicted Breakages (3-7 issues max)
2. Risk Assessment (single table)
3. Manual Test Plan (test sessions with checkboxes)

Total length: 300 lines maximum

All other sections are FORBIDDEN unless explicitly valuable for THIS deployment.

---

When to Use This Skill

• User mentions: "QA test plan", "CSS conflicts", "analyze CSS for [customer]"
• Customer CSS deployment scenarios
• Pre-deployment risk assessment

---

Analysis Workflow

Step 1: Read Customer CSS

• Location: src/customers/[customer]/custom.css
• Flag: Global selectors, !important, element selectors

Step 2: Read Application Styles

• Core: src/app/@core/components/
• Theme: src/app/@theme/
• Pages: src/app/pages/

Step 3: Detect Conflicts

• Compare selectors
• Check specificity + !important
• Predict cascade issues

Step 4: Check Git History

• Search: "CSS", "style", "[customer name]"
• Find patterns from past bugs

Step 5: Generate Report

Follow exact format below.

---

REQUIRED OUTPUT FORMAT

File: risk reports/qa_test_report_[customer].md

Structure (STRICT):

# QA Test Report - [Customer] CSS Deployment

## 🔍 Predicted Breakages

### Issue #1: [Short Title]
**Severity:** 🔴 CRITICAL
**Priority:** P1
**Confidence:** 95%

**What Breaks:**
- Component: [name]
- Expected: [normal behavior]
- After deploy: [broken behavior]

**Root Cause:**
```css
/* problematic CSS */
```

**Impact:**
- User: [how affected]
- Business: [consequences]

**Where to Test:**
- Page 1
- Page 2

**Test Steps:**
1. Navigate to [page]
2. Check [element]
3. Expected: [result]
4. If broken: [what happens]

[Repeat for issues #2-7 max]

---

## 📊 Risk Assessment

| Issue | Severity | Priority | Confidence | Impact |
|-------|----------|----------|------------|--------|
| [Issue 1] | 🔴 CRITICAL | P1 | 95% | [desc] |
| [Issue 2] | 🟡 HIGH | P2 | 90% | [desc] |

**Severity Key:**
- 🔴 CRITICAL: App unusable, blocks deployment
- 🟡 HIGH: Major visual/functional issue
- 🟢 MEDIUM: Minor issue, workarounds exist
- ⚪ LOW: Negligible impact

---

## 📋 Manual Test Plan

### Test Session 1: Critical Issues (XX min)

**Setup:**
- Browser: Chrome (latest)
- Environment: Test

#### Test 1.1: [Test Name]
- [ ] Step 1: [action]
- [ ] Step 2: [action]
- [ ] Expected: [result]
- [ ] If broken: [report what]
- [ ] Stop condition: Yes/No

[Repeat for sessions 2-3 max]

---
**END OF REPORT**

---

Optional Sections (Use Sparingly)

Include ONLY if critically valuable:

• Executive Summary (2-3 lines) - Deploy yes/no at top
• Stop Conditions - When to halt testing
• Time Estimates - Per session only

Do NOT add:

• Screenshots checklists
• Bug templates
• Historical context
• Reference info
• Success criteria
• Contact info
• Execution logs
• ASCII diagrams

---

Severity Guidelines

Confidence:

• 100%: Same selector + !important
• 90-95%: Global selector
• 80-90%: Similar selectors
• 70-80%: Potential conflict
• <70%: Low risk

Priority:

• P1: Blocks deployment
• P2: Test before deploy
• P3: Document only

---

Best Practices

1. Check git history for patterns
2. Focus on: button, input, div, element selectors
3. Any !important = red flag
4. Global selectors = dangerous
5. Keep report under 300 lines

---

Output Checklist

Before saving, verify:

• [ ] Only 3 main sections
• [ ] 3-7 issues max
• [ ] Under 300 lines
• [ ] No forbidden sections
• [ ] Saved to: risk reports/qa_test_report_[customer].md

---

Example Output Length

# QA Test Report - CustomerB CSS
## Predicted Breakages (7 issues × 20 lines = 140 lines)
## Risk Assessment (table = 15 lines)
## Manual Test Plan (3 sessions × 30 lines = 90 lines)
---
Total: ~250 lines ✅

---