Make It Pretty

Make It Pretty

I'll improve code readability while preserving exact functionality.

Token Optimization Strategy

Target Reduction: 65% (3,000-4,500 → 1,050-1,575 tokens)

Core Optimization Patterns

1. Focused Scope (Save 60-70%)

# ❌ AVOID: Reading entire codebase
Glob "**/*.{js,ts,jsx,tsx,py}" && Read all files

# ✅ PREFER: Focused file analysis
# User-specified file
Read "/path/to/specific-file.ts"

# OR recently changed code
git diff --name-only HEAD~3 | head -5
Read changed-files-only

2. Template-Based Formatting (Save 50-60%)

# ❌ AVOID: Manual formatting analysis
Read file → analyze patterns → apply improvements

# ✅ PREFER: Template-based patterns
# Use cached style patterns from .claude/cache/make-it-pretty/
1. Load project style guide (if cached)
2. Apply consistent naming conventions
3. Use formatter configs (prettier, black, gofmt)

3. Git Diff for Changed Code (Save 70-80%)

# ❌ AVOID: Analyzing all project files
Glob "**/*" → Read all → identify ugly code

# ✅ PREFER: Focus on recent changes
git diff HEAD~3 --name-only --diff-filter=AM | head -5
Read only-changed-files

4. Bash-Based Formatter Execution (Save 80-90%)

# ❌ AVOID: Manual formatting via Edit tool
Read → Edit line-by-line → Write

# ✅ PREFER: Use existing formatters
# JavaScript/TypeScript
npx prettier --write "src/**/*.{js,ts}"

# Python
black . --line-length 88

# Go
gofmt -w .

# Rust
cargo fmt

5. Progressive Improvements (Save 50-60%)

# ❌ AVOID: All improvements at once
Fix naming + structure + types + complexity in one pass

# ✅ PREFER: One aspect at a time
Pass 1: Naming (variables, functions)
Pass 2: Structure (extraction, grouping)
Pass 3: Types (annotations, specificity)
Pass 4: Cleanup (unused code, redundancy)

6. Early Exit When Clean (Save 90%)

# ❌ AVOID: Analyzing already-clean code
Read all files → check each → report "already clean"

# ✅ PREFER: Quick validation check
1. Check formatter config exists
2. Run formatter in check mode
3. Exit if no changes needed

# Example
npx prettier --check "src/**/*.ts" && exit 0

Implementation Checklist

Before Starting:

• Ask user for specific file(s) to improve
• If no file specified, use git diff --name-only HEAD~3
• Limit to 3-5 files per session
• Check if formatters are configured

During Execution:

• Create git checkpoint: git stash push -m "pre-prettify"
• Use Bash for formatter execution (prettier, black, gofmt)
• Load cached style patterns from .claude/cache/make-it-pretty/
• Apply one improvement type per pass
• Run tests after each pass if available

After Completion:

• Cache new style patterns discovered
• Report improvements made (naming, structure, types)
• Suggest remaining improvements for future sessions

Caching Strategy

Cache Location: .claude/cache/make-it-pretty/

Cached Data:

1. style-guide.json - Project naming conventions

   • Variable naming patterns (camelCase, snake_case)
   • Function naming patterns
   • File naming patterns
   • Indentation preferences (tabs vs. spaces)

2. type-patterns.json - Type annotation standards

   • Common type definitions
   • Generic type usage
   • Type import patterns

3. formatter-config.json - Detected formatters

   • Prettier config
   • ESLint config
   • Black config
   • Language-specific formatters

Cache Invalidation:

• Clear cache when formatter configs change
• Refresh after major refactoring
• User can force refresh with --no-cache flag

Usage Patterns

Standard Usage:

# Improve specific file
/make-it-pretty src/components/UserProfile.tsx

# Improve recently changed files
/make-it-pretty --recent

# Focus on specific aspect
/make-it-pretty --naming src/utils/helpers.ts
/make-it-pretty --structure src/api/client.ts
/make-it-pretty --types src/models/user.ts

# Dry run (preview changes)
/make-it-pretty --dry src/components/*.tsx

Focus Area Flags:

• --naming - Only improve variable/function names
• --structure - Only reorganize code structure
• --types - Only improve type annotations
• --dry - Preview changes without applying
• --no-cache - Force refresh of style patterns

Expected Token Usage

Optimized Pattern:

• File scope identification: 50-100 tokens
• Load cached style patterns: 100-200 tokens
• Bash formatter execution: 100-200 tokens
• Targeted improvements: 500-800 tokens
• Validation & report: 200-300 tokens
• Total: 1,050-1,575 tokens (65% reduction)

Comparison (Previous Unoptimized):

• Read entire codebase: 1,000-1,500 tokens
• Manual pattern analysis: 800-1,200 tokens
• Line-by-line edits: 600-1,000 tokens
• Comprehensive improvements: 400-600 tokens
• Full testing: 200-300 tokens
• Total: 3,000-4,500 tokens

Token Savings: 1,950-2,925 tokens (65% reduction)

Optimization Status

• ✅ Optimized (Phase 2 Batch 3D-F, 2026-01-26)
• ✅ Focused scope (user-specified or git diff)
• ✅ Template-based formatting patterns
• ✅ Bash-based formatter execution
• ✅ Progressive improvements
• ✅ Early exit when clean
• ✅ Caching style patterns

Strategic Thinking Process

1. Code Analysis

   • What makes this code "ugly" or hard to read?
   • Are there complex nested structures that need simplification?
   • Is the naming unclear or inconsistent?
   • Are there repeated patterns that could be abstracted?

2. Safety Considerations

   • Which changes are purely cosmetic (safe)?
   • Which might affect behavior (risky)?
   • Are there implicit dependencies on current structure?
   • Could renaming break external references?

3. Improvement Strategy

   • Priority 1: Clear naming (variables, functions, files)
   • Priority 2: Reduce complexity (extract functions, simplify logic)
   • Priority 3: Remove redundancy (DRY principle)
   • Priority 4: Improve type safety (if applicable)

4. Validation Approach

   • How can I ensure functionality remains identical?
   • What tests exist to verify behavior?
   • Should I add temporary logging to verify flow?

Based on this analysis, I'll proceed safely:

Safety First:

• Create git checkpoint before changes
• Use Write tool to create backups
• Track all modifications systematically

I'll identify files to beautify based on:

• Files you specify, or if none specified, analyze the entire application
• Recently modified code
• Our conversation context

Improvements I'll Make:

• Variable and function names for clarity
• Code organization and structure
• Remove unused code and clutter
• Simplify complex expressions
• Group related functionality
• Fix loose or generic type declarations
• Add missing type annotations where supported
• Make types more specific based on usage

My Approach:

1. Analyze current code patterns and type usage
2. Apply consistent naming conventions
3. Improve type safety where applicable
4. Reorganize for better readability
5. Remove redundancy without changing logic

Quality Assurance:

• All functionality remains identical
• Tests continue to pass (if available)
• No behavior changes occur
• Clear commit messages for changes

Important: I will NEVER:

• Add "Co-authored-by" or any Claude signatures
• Include "Generated with Claude Code" or similar messages
• Modify git config or user credentials
• Add any AI/assistant attribution to the commit

This helps transform working code into maintainable code without risk.