| name | playwright-review |
| description | Use when reviewing Playwright BDD test files for quality issues — locally staged/modified files or PR diffs. Triggers on: "review playwright code", "check test quality", "review before commit", staged test changes. |
Playwright BDD Code Review
Overview
Structured review process for Playwright BDD test code in this repository. Classifies findings P0–P3, applies framework-specific standards, and produces actionable reports.
When to Use
- Before committing test changes
- During PR review of test files
- After fixing test failures to verify fix quality
- When onboarding to understand what "good" looks like
Review Process
Step 1 — Categorize Files
Classify each changed file:
- spec —
.feature files
- page-object —
src/pages/*.ts
- step-definition —
src/steps/*.ts
- support —
src/utils/*.ts, src/services/**/*.ts, config files
Step 2 — Load Standards
Read both reference files before applying checks:
.claude/references/standards.md — framework architecture standards.claude/references/anti-patterns.md — full P0–P3 taxonomy with examples
These replace any external documentation. Always read both files fresh — do not rely on memory.
Step 3 — Apply Anti-Pattern Checks
Apply all P0–P3 checks from .claude/references/anti-patterns.md.
Regex patterns to scan for:
P0: waitForTimeout|page\.\$\(|page\.\$\$\(|innerHTML|mock-token|authenticated.*true.*catch
P1: networkidle|module-level.*let.*Page|process\.env\.\w+!
P2: isPageLoaded|LOG_LEVEL_MAP\[.*\];$|apiResponse\?.*unknown
P3: log\.debug.*initialized|Before\(async.*log\.debug|@Sample
Step 4 — Keyword Elevation
Auto-escalate any finding to P0 if the affected code touches:
auth, login, password, credentials, token, secret, payment, pii, phi, hipaa, security
Example: A P2 misleading return type in an auth step → P0.
Step 5 — Generate Report
## Playwright BDD Code Review Report
**Files reviewed:** N
**Critical (P0):** N | **High (P1):** N | **Medium (P2):** N | **Low (P3):** N
---
### 🔴 Critical Issues (P0) — Block Merge
#### [File:Line] Issue Title
**Pattern:** what was found
**Risk:** why it matters
**Fix:**
\```typescript
// Before
...
// After
...
\```
---
### 🟠 High Priority (P1) — Should Fix
...
### 🟡 Medium Priority (P2)
...
### 🔵 Low Priority (P3) — Suggestions
...
### ✅ Passed Checks
- [list what was done correctly]
Step 6 — Auto-Fix Guidance
For each P0 finding, provide exact before/after replacement. Reference the auto-fix table in .claude/references/anti-patterns.md.
Quick Reference: Priority Rules
| Priority | Action | Blocks Merge? |
|---|
| P0 | Must fix | Yes |
| P1 | Should fix (document if deferred) | Recommended |
| P2 | Fix when touching the area | No |
| P3 | Suggestion only | No |
Common Mistakes
| Mistake | Correct behaviour |
|---|
| Treating P3 as blocking | P3 = suggestion only, never block |
| Skipping standards files | Always read both reference files fresh |
| Omitting ✅ Passed Checks section | Always list what passed — builds trust |
| Reporting issues in untouched code | Only report on changed lines unless P0 |
| Not applying keyword elevation | Auth/payment/PII findings always P0 |
Integration
- Local review:
git diff --staged → apply this skill → report in conversation
- Full diff review:
git diff HEAD → apply this skill
- PR review: Get PR diff → apply this skill → post as PR comment
