Phy Bundle Size AuditSAFE

🧠 Curated collection of 1209+ best OpenClaw skills — weekly updated by MyClaw.ai

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: phy-bundle-size-audit
description: JavaScript bundle size auditor and budget enforcer. Parses webpack stats JSON, Vite bundle report, Rollup output, or Next.js build output to identify largest chunks, flag chunks that exceed configurable size budgets, detect duplicate dependencies across chunks, find treeshaking failures (packages that should be side-effect-free but aren't), and generate CI fail-gate commands. Outputs a GitHub PR annotation-ready summary with per-chunk size, gzip estimate, and actionable optimization suggestions. Zero external API — pure local file analysis. Triggers on "bundle too large", "webpack stats", "chunk size", "bundle budget", "treeshaking", "bundle analyzer", "/bundle-size-audit".
license: Apache-2.0
metadata:
  author: PHY041
  version: "1.0.0"
  tags:
    - webpack
    - vite
    - bundler
    - performance
    - treeshaking
    - developer-tools
    - javascript
    - nextjs
    - rollup
---

# Bundle Size Auditor

You added one import. Your Lighthouse score dropped 12 points. Your LCP went from 1.8s to 3.1s.

This skill parses your bundler's output — webpack stats JSON, Vite build report, or Next.js `.next/` directory — identifies the largest chunks, finds which packages are bloating them, detects treeshaking failures, and gives you the exact code changes to fix it.

**Supports webpack, Vite, Rollup, Next.js. Zero external API.**

---

## Trigger Phrases

- "bundle too large", "my bundle is huge", "reduce bundle size"
- "webpack stats", "analyze webpack output", "chunk size"
- "bundle budget", "size limit", "bundle regression"
- "treeshaking not working", "dead code in bundle"
- "which package is largest", "find large dependencies"
- "Next.js bundle", "Vite build report"
- "/bundle-size-audit"

---

## How to Provide Input

```bash
# Option 1: Webpack stats JSON (most detailed)
webpack --profile --json > stats.json
/bundle-size-audit stats.json

# Option 2: Vite build (auto-detected)
vite build
/bundle-size-audit dist/

# Option 3: Next.js build (auto-detected)
next build
/bundle-size-audit .next/

# Option 4: Rollup stats
rollup -c --bundleConfigAsCjs
/bundle-size-audit rollup-stats.json

# Option 5: Set size budgets
/bundle-size-audit --max-chunk 250kb --max-initial 500kb

# Option 6: Focus on a specific chunk or package
/bundle-size-audit --chunk main --package lodash

# Option 7: Generate CI fail-gate only
/bundle-size-audit --ci-config
```

---

## Step 1: Detect Build Artifacts

```bash
python3 -c "
import os, json
from pathlib import Path

# Priority order for detection
checks = [
    ('webpack stats', ['stats.json', 'webpack-stats.json', 'build/stats.json']),
    ('Next.js', ['.next/build-manifest.json', '.next/server/pages-manifest.json']),
    ('Vite', ['dist/', 'dist/assets/']),
    ('Rollup', ['dist/', 'build/']),
    ('Create React App', ['build/static/js/', 'build/asset-manifest.json']),
]

found = []
for name, paths in checks:
    for p in paths:
        if os.path.exists(p):
            found.append((name, p))
            break

if found:
    for name, path in found:
        size = 0
        if os.path.isfile(path):
            size = os.path.getsize(path)
            print(f'Detected {name}: {path} ({size:,} bytes)')
        else:
            # Directory — count JS files
            js_files = list(Path(path).rglob('*.js'))
            total = sum(f.stat().st_size for f in js_files)
            print(f'Detected {name}: {path}/ ({len(js_files)} JS files, {total/1024:.0f} KB total)')
else:
    print('No build artifacts found.')
    print('Run your build first:')
    print('  webpack --profile --json > stats.json')
    print('  vite build')
    print('  next build')
"
```

---

## Step 2: Parse Bundle Data

### For Webpack Stats JSON

```python
import json
from pathlib import Path

def parse_webpack_stats(stats_path):
    """Parse webpack stats.json into chunk/asset summary."""
    with open(stats_path) as f:
        stats = json.load(f)

    assets = []
    for asset in stats.get('assets', []):
        name = asset['name']
        size = asset['size']
        # Only JS and CSS assets
        if not any(name.endswith(ext) for ext in ['.js', '.css', '.mjs']):
            continue
        assets.append({
            'name': name,
            'size': size,
            'size_kb': round(size / 1024, 1),
            'gzip_estimate_kb': round(size / 1024 * 0.3, 1),  # ~30% compression typical
            'chunks': asset.get('chunkNames', []),
            'initial': asset.get('isOverSizeLimit', False),
        })

    # Sort by size desc
    assets.sort(key=lambda x: x['size'], reverse=True)

    # Module analysis — find largest modules
    modules = []
    for module in stats.get('modules', []):
        if module.get('size', 0) > 10 * 1024:  # >10KB only
            modules.append({
                'name': module.get('name', 'unknown'),
                'size': module.get('size', 0),
                'size_kb': round(module.get('size', 0) / 1024, 1),
                'reasons': [r.get('moduleName', '') for r in module.get('reasons', [])[:3]],
            })
    modules.sort(key=lambda x: x['size'], reverse=True)

    return assets, modules[:50]  # top 50 modules


def parse_webpack_chunks(stats_path):
    """Identify chunks and their composition."""
    with open(stats_path) as f:
        stats = json.load(f)

    chunks = []
    for chunk in stats.get('chunks', []):
        chunks.append({
            'id': chunk.get('id'),
            'names': chunk.get('names', []),
            'size': chunk.get('size', 0),
            'size_kb': round(chunk.get('size', 0) / 1024, 1),
            'initial': chunk.get('initial', False),
            'entry': chunk.get('entry', False),
            'module_count': len(chunk.get('modules', [])),
            'files': chunk.get('files', []),
        })

    chunks.sort(key=lambda x: x['size'], reverse=True)
    return chunks
```

### For Vite / CRA (dist/ directory)

```python
import os
from pathlib import Path

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

What the source does

Filesystem

none-observed

Network

none-observed

Shell

none-observed

Dependencies

pinned

Secrets in source

none-found

Findings (1)

<key>k5Ey9KFlkqpj+SDkUw+5ED9lTA3En/qUi0zdrydUCH3kMWTE3Eh65NXnFCaxlY2omY2JHnlEoK7Li7oOEvM7eG5VPdcO/sFlMfoCRdnLYdepJ+uLzYwOWR8W4yQVve/clxVFTVRL4DFleKInGdpAxIbHZT2yi4ADAMENls1N1XSLojRuqXePXDeAT/4Mv4TTx0s

Gates applied: no_behavioural_pass.