App Analyzer & Optimizer (2026 Edition)

English | Bahasa Indonesia

---

English

Description

Deep application analysis and optimization skill. Performs architectural audits, bottleneck detection, bundle analysis, database query profiling, and AI-assisted code review across Next.js, React, Node.js, Go, Python, and Rust applications.

Trigger Conditions

• Auditing an existing codebase for architectural issues or technical debt.
• Detecting performance bottlenecks (slow API routes, large bundles, N+1 queries).
• Analyzing bundle size and suggesting code splitting opportunities.
• Reviewing code quality (complexity, duplication, dead code).
• Using AI-assisted tools to accelerate code review and analysis.

Analysis Protocol (5 Phases)

Phase 1: Architecture Discovery

# Map the entire project structure
find . -type f -name "*.ts" -o -name "*.tsx" -o -name "*.go" | head -100
# Count lines per file (find largest files)
find . -name "*.ts" -exec wc -l {} + | sort -rn | head -20
# Find circular dependencies (TypeScript/JS)
npx madge --circular --extensions ts,tsx src/

Phase 2: Bundle Analysis (Next.js / Vite)

# Next.js bundle analyzer
ANALYZE=true next build

# Or install dedicated tool
npx @next/bundle-analyzer

# Vite bundle visualization
npx vite-bundle-visualizer

# Check for duplicate dependencies
npx depcheck
npx bundle-phobia-cli check package.json

Key bundle red flags:

• Any single chunk > 500KB (uncompressed).
• Importing entire libraries (import _ from 'lodash' vs import debounce from 'lodash/debounce').
• Moment.js (replace with date-fns or Temporal).
• node_modules leaking into client bundle.

Phase 3: Database Query Analysis

-- PostgreSQL: Find slow queries
SELECT query, mean_exec_time, calls, total_exec_time
FROM pg_stat_statements
ORDER BY mean_exec_time DESC
LIMIT 20;

-- Find missing indexes (sequential scans on large tables)
SELECT schemaname, tablename, attname, n_distinct, correlation
FROM pg_stats
WHERE tablename = 'your_table'
ORDER BY n_distinct DESC;

-- Detect N+1 queries: Enable query logging
-- In Drizzle:
const db = drizzle(client, { logger: true });
-- In Prisma:
DATABASE_URL="...?connection_limit=5" prisma studio

Phase 4: AI-Assisted Code Review

Use AI tools to accelerate analysis:

Tool	Purpose
CodeRabbit	Automated PR review, pattern detection
Sourcegraph Cody	Codebase-wide semantic search and explanation
GitHub Copilot	Inline suggestions and refactoring
Cursor / Windsurf	AI IDE with full-repo context

AI review prompts for Gemini Agent:

Analyze all files in src/features/ and identify:
1. Functions longer than 30 lines
2. Duplicated business logic across files
3. Missing error handling in async functions
4. Components that directly call APIs (violating separation of concerns)

Phase 5: Performance Profiling

Frontend (React / Next.js):

# React DevTools Profiler — identify render bottlenecks
# Chrome DevTools > Performance tab > Record interaction

# Lighthouse CI — automated CWV tracking
npm install -g @lhci/cli
lhci autorun

Backend (Node.js):

# Built-in Node.js profiler
node --prof server.js
node --prof-process isolate-*.log > processed.txt

# Clinic.js — flamegraph, bubble chart, doctor
npm install -g clinic
clinic doctor -- node server.js
clinic flame -- node server.js

Go:

# Built-in pprof profiler
go tool pprof http://localhost:6060/debug/pprof/profile?seconds=30
go tool pprof -http=:8080 profile.out

Common Bottleneck Patterns & Fixes

Bottleneck	Symptom	Fix
N+1 queries	DB calls proportional to list length	Add .include() / JOIN or DataLoader
Missing indexes	Slow queries on filtered columns	CREATE INDEX CONCURRENTLY
Unoptimized images	Large LCP, slow page load	Next.js <Image>, WebP, lazy loading
Over-fetching	Fetching full objects when only 1 field needed	Select specific columns
No pagination	Fetching entire tables	Add LIMIT/OFFSET or cursor pagination
Blocking main thread	High INP, unresponsive UI	useTransition, web workers
No caching	Same data fetched repeatedly	Redis, React Query staleTime, HTTP cache
Bundle bloat	Large JS payload	Tree-shaking, code splitting, lazy imports

---

Bahasa Indonesia

Integrasi Orkestrasi

Terhubung dan mengorkestrasi skill domain yang relevan seperti brainstorming, zero-to-prod-orchestrator, dan project-context-mapper untuk memastikan eksekusi yang kohesif.

Deskripsi

Skill analisis dan optimasi aplikasi mendalam. Melakukan audit arsitektur, deteksi bottleneck, analisis bundle, profiling query database, dan code review berbantuan AI di aplikasi Next.js, React, Node.js, Go, Python, dan Rust.

Kondisi Pemicu

• Mengaudit codebase yang ada untuk masalah arsitektur atau hutang teknis.
• Mendeteksi bottleneck performa (rute API lambat, bundle besar, query N+1).
• Menganalisis ukuran bundle dan menyarankan peluang code splitting.
• Meninjau kualitas kode (kompleksitas, duplikasi, kode mati).
• Menggunakan tool berbantuan AI untuk mempercepat code review dan analisis.

Protokol Analisis (5 Fase)

Fase 1: Penemuan Arsitektur

Petakan seluruh struktur proyek, hitung baris per file untuk menemukan file terbesar, dan deteksi dependensi circular dengan npx madge --circular.

Fase 2: Analisis Bundle (Next.js / Vite)

Gunakan ANALYZE=true next build atau npx vite-bundle-visualizer. Tanda bahaya bundle: chunk tunggal > 500KB, mengimpor library penuh, Moment.js, atau node_modules yang bocor ke bundle klien.

Fase 3: Analisis Query Database

Gunakan pg_stat_statements untuk menemukan query lambat. Cari scan sequential pada tabel besar (tanda indeks yang hilang). Aktifkan logging query di Drizzle/Prisma untuk mendeteksi pola N+1.

Fase 4: Code Review Berbantuan AI

Gunakan CodeRabbit untuk review PR otomatis, Sourcegraph Cody untuk pencarian semantik seluruh codebase, dan Gemini Agent dengan prompt analisis spesifik untuk menemukan fungsi panjang, logika bisnis yang diduplikasi, dan error handling yang hilang.

Fase 5: Profiling Performa

• Frontend: React DevTools Profiler, Lighthouse CI.
• Backend Node.js: node --prof, Clinic.js untuk flamegraph.
• Go: go tool pprof untuk CPU dan memory profiling.

Pola Bottleneck Umum & Perbaikan

Bottleneck	Gejala	Perbaikan
Query N+1	Panggilan DB proporsional dengan panjang daftar	Tambahkan .include() / JOIN atau DataLoader
Indeks yang hilang	Query lambat pada kolom yang difilter	CREATE INDEX CONCURRENTLY
Gambar tidak dioptimalkan	LCP besar, halaman lambat	Next.js <Image>, WebP, lazy loading
Over-fetching	Mengambil objek penuh saat hanya 1 field dibutuhkan	Pilih kolom spesifik
Tanpa paginasi	Mengambil seluruh tabel	Tambahkan LIMIT/OFFSET atau cursor pagination
Memblokir thread utama	INP tinggi, UI tidak responsif	useTransition, web workers
Tanpa caching	Data yang sama diambil berulang kali	Redis, React Query staleTime, HTTP cache
Bundle membengkak	Payload JS besar	Tree-shaking, code splitting, lazy imports

Orchestration & Integration

• Integrates with performance-web-vitals and production-ready-hardener.