Performance. Bundle size, time to interactive, frame rate, memory leaks, battery drain. 60fps target. Oldest supported device gets the same test as newest.
Crash & Bug Hunt. Edge cases, impossible state machines, division by zero, empty arrays, save corruption, localStorage failures. Every way a player can break what you built.
Code Quality. console.log count (target: 0). TypeScript any abuse. Dead code. Does it build clean? TODO comments that became permanent.
Six reports. File:line references on everything. Not "you might have a soft-lock." Exact state, exact reproduction, exact fix.
Here's the full prompt. Copy everything below. Open Claude Code in your game directory. Paste it. Type "proceed."
# PLAYTEST READINESS AUDIT REQUEST
I need a comprehensive pre-launch engineering audit for my game,
similar to what Sierra On-Line, Dope Wars, and Gangbusters developers would do.
**Context:**
- Game: [YOUR GAME NAME]
- Genre: [GENRE]
- Platform: [Web/iOS/Android/Desktop]
- Status: [X months/years of development]
- Goal: Prepare for playtest and launch
Run all 5 audits sequentially, creating detailed markdown reports for each:
## AUDIT 1: UI/UX Polish (Sierra-Style)
Focus: Pixel-perfect interface, consistency, mobile-first design
Analyze:
1. Visual consistency across all screens
2. Color palette usage and brand coherence
3. Typography hierarchy and readability
4. Interaction feedback (hover, active, disabled states)
5. Touch targets (44x44px minimum for mobile)
6. Animation quality (60fps, GPU-accelerated)
7. Information hierarchy (critical content above fold)
8. Mobile optimization (iPhone SE to iPhone Pro Max)
9. Accessibility (WCAG AA, color contrast)
10. Screen real estate efficiency (maximize content, minimize chrome)
Deliverable: SIERRA_UI_AUDIT_REPORT.md
Critical/High/Medium/Low severity. File:line references. Before/after recommendations.
## AUDIT 2: Game Economy/Loop Balance (Dope Wars-Style)
Focus: Risk/reward, addictive core loop, economic pressure
Analyze:
1. Core gameplay loop — map the full cycle, timing, rewards
2. Economic balance — starting capital, costs, profit margins
3. Risk/reward proportionality — harder = more reward?
4. Progression pacing — early/mid/late game difficulty curve
5. Dominant strategies — is one approach obviously best?
6. Resource pressure — are choices meaningful or obvious?
7. "One more turn" factor — what drives replay?
8. Soft-lock scenarios — can players reach unwinnable states?
9. Exploit potential — money duplication, price manipulation, infinite loops
10. Time pressure mechanics — loans, deadlines, decay, consequences
Deliverable: ECONOMIC_BALANCE_REPORT.md
Profitability analysis by strategy. Risk/reward matrix. Progression curve.
Balance recommendations with specific value changes.
## AUDIT 3: Performance Profiling (60fps Mobile)
Focus: Load time, frame rate, battery, memory
Analyze:
1. Bundle size — total, gzipped, breakdown by category
2. Time to Interactive — target <2s on 4G
3. Code splitting — lazy loading opportunities
4. React optimization — memo, useCallback, useMemo coverage
5. CSS performance — GPU vs CPU rendering, compositor layers
6. Animation frame rate — 60fps target, drop detection
7. Memory usage and leak detection
8. Battery drain estimation
9. Network usage (if online features exist)
10. Device compatibility — oldest to newest supported devices
Deliverable: PERFORMANCE_PROFILING_REPORT.md
Current vs target metrics. Bottleneck identification.
Optimization priorities. Estimated gains per fix.
## AUDIT 4: Crash & Bug Hunt (Sierra-Style Testing)
Focus: Edge cases, soft-locks, crashes, data corruption
Analyze:
1. Error handling — try-catch coverage, React error boundaries, global handlers
2. Data boundary testing — 0, negative, null, undefined, NaN everywhere
3. State machine validation — impossible states, race conditions
4. Array operations — empty arrays, out-of-bounds access
5. Division by zero protection — economy calculations, ratios
6. localStorage/save corruption — what happens on bad data?
7. Game-over triggers — are all end states handled?
8. Soft-lock scenarios — 0 resources, stuck states, no valid moves
9. User input validation — all entry points sanitized
10. Edge case test scenarios — document each one with reproduction steps
Deliverable: BUG_HUNT_REPORT.md
Critical/High/Medium/Low priority. Reproduction steps for each.
Soft-lock scenarios to test during playtest.
Fix recommendations with file:line references.
## AUDIT 5: Production Code Quality
Focus: Technical debt, maintainability, launch readiness
Analyze:
1. console.log statements — count and location (target: 0 in production)
2. TypeScript strictness — any usage, @ts-ignore count (target: 0)
3. TODO/FIXME comments — categorize by urgency, estimate resolution
4. Code organization — architecture, module separation, circular deps
5. Test coverage — what exists, what's missing, what's critical
6. Linting/formatting — ESLint config, consistent enforcement
7. Dead code and duplication — unused imports, copy-paste patterns
8. Security vulnerabilities — if applicable
9. Documentation — README accuracy, JSDoc on complex systems
10. Build output — warnings, bundle analysis, tree-shaking effectiveness
Deliverable: PRODUCTION_CODE_AUDIT.md
Code quality score (1-10). Technical debt inventory.
Launch readiness verdict. Post-launch recommendations.
## FINAL DELIVERABLE
Create PLAYTEST_READINESS_SUMMARY.md:
- Overall scores per audit area (1-10)
- Top 10 critical issues ranked by priority
- Estimated fix time: playtest-ready vs launch-ready
- Success metrics for launch day
- Clear go/no-go recommendation
Standards: Sierra On-Line polish. Dope Wars loop addiction. Gangbusters player agency.
Tone: Specific. "Soft-lock at line 247 in economy.ts when gold === 0" not "possible soft-lock."
Include file:line references for every issue.
Expect 5-10 critical issues. 10-20 hours of actual fixes.
The economy audit is the one that will surprise you. You've played your own game a thousand times with developer knowledge. Claude maps every strategy tree cold. Finds the path to 10x profit in six turns that you never noticed because you never needed to min-max your own game.
Soft-locks are the playtest killer. One tester hits an impossible state and the session's over. Find them before they do.
The first report is hard to read.
Good. Finding it two weeks before playtest beats finding it during.
