duplication-detect（重复检测）

Code Duplication Detection & DRY Refactoring

I'll analyze your codebase for duplicate code blocks, identify similar patterns across files, and suggest DRY (Don't Repeat Yourself) refactoring strategies based on obra principles.

Detection Capabilities:

• Exact code duplication (copy-paste detection)
• Similar code patterns (semantic duplication)
• Repeated logic across files
• Duplicated constants and magic numbers
• Repeated test patterns

Supported Languages:

• JavaScript/TypeScript
• Python
• Go
• Java
• Generic text-based detection

Token Optimization

This skill uses duplication detection-specific patterns to minimize token usage:

1. Source Directory Detection Caching (500 token savings)

Pattern: Cache project structure and source directories

• Store structure in .duplication-structure-cache (1 hour TTL)
• Cache: source directories, file extensions, excluded paths
• Read cached structure on subsequent runs (50 tokens vs 550 tokens fresh)
• Invalidate on directory structure changes
• Savings: 91% on repeat duplication checks

2. Bash-Based Duplication Tool Execution (1,800 token savings)

Pattern: Use jscpd/PMD directly via bash

• JavaScript: jscpd --format json (300 tokens)
• Python: pylint --duplicate-code (300 tokens)
• Generic: simian or custom grep-based detection (400 tokens)
• Parse JSON output with jq
• No Task agents for duplication detection
• Savings: 90% vs Task-based duplication analysis

3. Sample-Based Duplication Reporting (900 token savings)

Pattern: Report first 10 duplication instances only

• Show top 10 duplications by severity (600 tokens)
• Count remaining duplications without details
• Full report via --all flag
• Savings: 65% vs reporting every duplication

4. Template-Based DRY Refactoring Recommendations (800 token savings)

Pattern: Use predefined DRY patterns

• Standard strategies: extract function, extract constant, inheritance, composition
• Pattern-based recommendations for duplication types
• No creative refactoring generation
• Savings: 80% vs LLM-generated DRY strategies

5. Incremental Duplication Checks (1,000 token savings)

Pattern: Check only changed files via git diff

• Analyze files modified since last commit (500 tokens)
• Check if changes introduce new duplication
• Full codebase analysis via --full flag
• Savings: 75% vs full codebase duplication detection

6. Grep-Based Similar Pattern Discovery (700 token savings)

Pattern: Find potential duplications with grep

• Grep for repeated patterns: function signatures, constant values (300 tokens)
• Flag files with high similarity
• Run full tool only on flagged file pairs
• Savings: 70% vs running tool on all file combinations

7. Cached Duplication Baseline (600 token savings)

Pattern: Store baseline duplication report

• Cache initial duplication report
• Compare new runs against baseline to detect new duplications
• Focus on newly introduced duplications
• Savings: 80% by focusing on deltas

8. Threshold-Based Filtering (400 token savings)

Pattern: Filter out small duplications

• Default: 6+ lines of duplication
• Skip trivial duplications (imports, boilerplate)
• Adjustable threshold
• Savings: 70% by filtering noise

Real-World Token Usage Distribution

Typical operation patterns:

• Check recent changes (git diff scope): 1,200 tokens
• Show top 10 duplications: 1,400 tokens
• Full codebase analysis (first time): 3,000 tokens
• Cached baseline comparison: 800 tokens
• Refactoring recommendations (top 5): 1,800 tokens
• Most common: Incremental checks on recent changes

Expected per-analysis: 1,500-2,500 tokens (60% reduction from 3,500-6,000 baseline) Real-world average: 1,100 tokens (due to incremental checks, sample-based reporting, cached baseline)

Arguments: $ARGUMENTS - optional: minimum duplication threshold (default: 6 lines) or specific directories

DRY refactoring strategies:

• Extract function/method (most common)
• Extract constant/configuration
• Inheritance or composition
• Higher-order functions
• Template method pattern
• Strategy pattern for variations

Phase 1: Tool Setup & Configuration

First, I'll set up duplication detection tools:

#!/bin/bash
# Duplication Detection - Setup & Configuration

echo "=== Code Duplication Detection ==="
echo ""

# Create analysis directory
mkdir -p .claude/duplication-analysis
ANALYSIS_DIR=".claude/duplication-analysis"
TIMESTAMP=$(date +%Y%m%d-%H%M%S)
REPORT="$ANALYSIS_DIR/duplication-report-$TIMESTAMP.md"
MIN_LINES="${1:-6}"  # Minimum lines for duplication detection

echo "Configuration:"
echo "  Minimum duplication threshold: $MIN_LINES lines"
echo "  Analysis directory: $ANALYSIS_DIR"
echo ""

# Detect project structure
echo "Detecting project structure..."
SOURCE_DIRS=""

# Common source directories
for dir in src lib app components pages services utils helpers; do
    if [ -d "$dir" ]; then
        SOURCE_DIRS="$SOURCE_DIRS $dir"
        echo "  ✓ Found: $dir/"
    fi
done

# Python-specific
for dir in tests test __tests__; do
    if [ -d "$dir" ]; then
        echo "  ✓ Found test directory: $dir/"
    fi
done

if [ -z "$SOURCE_DIRS" ]; then
    echo "  Using current directory"
    SOURCE_DIRS="."
fi

echo ""

Phase 2: Install Detection Tools

I'll install and configure duplication detection tools:

echo "=== Installing Duplication Detection Tools ==="
echo ""

install_jscpd() {
    # JSCPD - Multi-language copy-paste detector
    if ! command -v jscpd >/dev/null 2>&1 && ! npm list -g jscpd >/dev/null 2>&1; then
        echo "Installing jscpd (copy-paste detector)..."
        npm install -g jscpd 2>/dev/null || npm install --save-dev jscpd

        if [ $? -eq 0 ]; then
            echo "✓ jscpd installed"
        else
            echo "⚠️  Failed to install jscpd - using basic detection"
            return 1
        fi
    else
        echo "✓ jscpd already installed"
    fi

    # Create jscpd configuration
    cat > "$ANALYSIS_DIR/.jscpd.json" << JSCPDCONFIG
{
    "threshold": $MIN_LINES,
    "reporters": ["json", "html"],
    "ignore": [
        "**/*.min.js",
        "**/node_modules/**",
        "**/dist/**",
        "**/build/**",
        "**/.git/**",
        "**/coverage/**",
        "**/__pycache__/**",
        "**/venv/**",
        "**/.venv/**"
    ],
    "format": ["javascript", "typescript", "python", "go", "java"],
    "absolute": false,
    "output": "$ANALYSIS_DIR/jscpd-report"
}
JSCPDCONFIG

    echo "✓ jscpd configuration created"
    return 0
}

# Install tool
if install_jscpd; then
    USE_JSCPD=true
else
    USE_JSCPD=false
    echo "ℹ️  Will use basic grep-based detection"
fi

echo ""

Phase 3: Detect Code Duplication

I'll analyze the codebase for duplicated code:

echo "=== Analyzing Code Duplication ==="
echo ""

run_jscpd_analysis() {
    echo "Running jscpd analysis..."
    echo ""

    jscpd $SOURCE_DIRS \
        --config "$ANALYSIS_DIR/.jscpd.json" \
        2>&1 | tee "$ANALYSIS_DIR/jscpd-log.txt"

    if [ $? -eq 0 ]; then
        echo ""
        echo "✓ jscpd analysis complete"

        # Check if duplications found
        if [ -f "$ANALYSIS_DIR/jscpd-report/jscpd-report.json" ]; then
            DUPLICATIONS=$(jq '.statistics.total.duplications' "$ANALYSIS_DIR/jscpd-report/jscpd-report.json" 2>/dev/null)
            PERCENTAGE=$(jq '.statistics.total.percentage' "$ANALYSIS_DIR/jscpd-report/jscpd-report.json" 2>/dev/null)

            echo ""
            echo "Duplication Statistics:"
            echo "  Total duplications: $DUPLICATIONS"
            echo "  Duplication percentage: $PERCENTAGE%"
            echo ""

            # Extract top duplications
            echo "Top Duplicated Blocks:"
            jq -r '.duplicates[] |
                "\(.format) - \(.lines) lines duplicated in \(.fragment) (\(.firstFile.name):\(.firstFile.start) and \(.secondFile.name):\(.secondFile.start))"' \
                "$ANALYSIS_DIR/jscpd-report/jscpd-report.json" 2>/dev/null | head -10
        fi
    else
        echo "⚠️  jscpd analysis failed - see $ANALYSIS_DIR/jscpd-log.txt"
        return 1
    fi
}

run_basic_duplication_detection() {
    echo "Running basic duplication detection..."
    echo ""

    # Find similar function signatures
    echo "Detecting similar function signatures..."

    # JavaScript/TypeScript functions
    if find $SOURCE_DIRS -name "*.js" -o -name "*.jsx" -o -name "*.ts" -o -name "*.tsx" 2>/dev/null | grep -q .; then
        grep -rh "^function\|^const.*=.*=>.*{$\|^export function" \
            --include="*.js" --include="*.jsx" --include="*.ts" --include="*.tsx" \
            $SOURCE_DIRS 2>/dev/null | \
            sort | uniq -d | head -10 > "$ANALYSIS_DIR/duplicate-functions.txt"

        if [ -s "$ANALYSIS_DIR/duplicate-functions.txt" ]; then
            echo "  Found duplicate function signatures (JavaScript/TypeScript):"
            cat "$ANALYSIS_DIR/duplicate-functions.txt" | sed 's/^/    /'
            echo ""
        fi
    fi

    # Python functions
    if find $SOURCE_DIRS -name "*.py" 2>/dev/null | grep -q .; then
        grep -rh "^def \|^async def " \
            --include="*.py" \
            $SOURCE_DIRS 2>/dev/null | \
            sort | uniq -d | head -10 > "$ANALYSIS_DIR/duplicate-python-functions.txt"

        if [ -s "$ANALYSIS_DIR/duplicate-python-functions.txt" ]; then
            echo "  Found duplicate function signatures (Python):"
            cat "$ANALYSIS_DIR/duplicate-python-functions.txt" | sed 's/^/    /'
            echo ""
        fi
    fi

    # Find magic numbers (potential constants)
    echo "Detecting magic numbers (repeated literals)..."

    find $SOURCE_DIRS -type f \( -name "*.js" -o -name "*.ts" -o -name "*.py" \) 2>/dev/null | \
        xargs grep -oh "[0-9]\{2,\}" 2>/dev/null | \
        sort | uniq -c | sort -rn | head -10 > "$ANALYSIS_DIR/magic-numbers.txt"

    if [ -s "$ANALYSIS_DIR/magic-numbers.txt" ]; then
        echo "  Most common numeric literals:"
        cat "$ANALYSIS_DIR/magic-numbers.txt" | sed 's/^/    /'
        echo ""
    fi

    # Find repeated strings
    echo "Detecting repeated string literals..."

    find $SOURCE_DIRS -type f \( -name "*.js" -o -name "*.ts" -o -name "*.py" \) 2>/dev/null | \
        xargs grep -oh "\"[^\"]\{10,\}\"" 2>/dev/null | \
        sort | uniq -c | sort -rn | head -10 > "$ANALYSIS_DIR/repeated-strings.txt"

    if [ -s "$ANALYSIS_DIR/repeated-strings.txt" ]; then
        echo "  Most common string literals:"
        cat "$ANALYSIS_DIR/repeated-strings.txt" | sed 's/^/    /'
        echo ""
    fi
}

# Run appropriate analysis
if [ "$USE_JSCPD" = true ]; then
    if ! run_jscpd_analysis; then
        echo "Falling back to basic detection..."
        run_basic_duplication_detection
    fi
else
    run_basic_duplication_detection
fi

Phase 4: Generate DRY Refactoring Strategies

I'll provide specific refactoring patterns for eliminating duplication:

echo ""
echo "=== Generating DRY Refactoring Strategies ==="
echo ""

cat > "$ANALYSIS_DIR/dry-refactoring-patterns.md" << 'DRYPATTERNS'
# DRY Refactoring Patterns

Based on obra YAGNI/DRY principles: Don't Repeat Yourself

---

## Pattern 1: Extract Function

**Problem:** Same code block repeated in multiple places

**Solution:** Extract to a reusable function

### Before (JavaScript)

// File 1 function processOrderA(order) { if (!order.customer || !order.customer.email) { throw new Error('Invalid customer'); } if (!order.items || order.items.length === 0) { throw new Error('Empty order'); } // ... process order }

// File 2 function processOrderB(order) { if (!order.customer || !order.customer.email) { throw new Error('Invalid customer'); } if (!order.items || order.items.length === 0) { throw new Error('Empty order'); } // ... different processing }

### After

// utils/validation.js export function validateOrder(order) { if (!order.customer?.email) { throw new Error('Invalid customer'); } if (!order.items?.length) { throw new Error('Empty order'); } }

// File 1 import { validateOrder } from './utils/validation';

function processOrderA(order) { validateOrder(order); // ... process order }

// File 2 import { validateOrder } from './utils/validation';

function processOrderB(order) { validateOrder(order); // ... different processing }

**DRY Improvement:** 8 duplicated lines → 1 function call

---

## Pattern 2: Extract Configuration/Constants

**Problem:** Same literals repeated across codebase

**Solution:** Centralize in configuration

### Before (Python)

Multiple files with repeated values

def calculate_shipping(): if weight < 10: return 5.99 return 9.99

def check_free_shipping(total): return total >= 50.00

def apply_discount(): if quantity >= 5: return 0.10 return 0

### After

config/constants.py

class ShippingConfig: FREE_SHIPPING_THRESHOLD = 50.00 LIGHT_PACKAGE_WEIGHT = 10 LIGHT_PACKAGE_COST = 5.99 HEAVY_PACKAGE_COST = 9.99

class DiscountConfig: BULK_QUANTITY = 5 BULK_DISCOUNT = 0.10

services/shipping.py

from config.constants import ShippingConfig, DiscountConfig

def calculate_shipping(weight): if weight < ShippingConfig.LIGHT_PACKAGE_WEIGHT: return ShippingConfig.LIGHT_PACKAGE_COST return ShippingConfig.HEAVY_PACKAGE_COST

def check_free_shipping(total): return total >= ShippingConfig.FREE_SHIPPING_THRESHOLD

def apply_discount(quantity): if quantity >= DiscountConfig.BULK_QUANTITY: return DiscountConfig.BULK_DISCOUNT return 0

**DRY Improvement:** Magic numbers eliminated, single source of truth

---

## Pattern 3: Template Method Pattern

**Problem:** Similar algorithms with slight variations

**Solution:** Use template method or strategy pattern

### Before (TypeScript)

class PDFReport { generate() { this.loadData(); this.formatHeader(); this.formatPDFContent(); this.addPDFFooter(); this.savePDF(); }

private loadData() { /* same logic */ } private formatHeader() { /* same logic */ } private formatPDFContent() { /* PDF-specific */ } private addPDFFooter() { /* PDF-specific */ } private savePDF() { /* PDF-specific */ } }

class ExcelReport { generate() { this.loadData(); this.formatHeader(); this.formatExcelContent(); this.addExcelFooter(); this.saveExcel(); }

private loadData() { /* DUPLICATE logic */ } private formatHeader() { /* DUPLICATE logic */ } private formatExcelContent() { /* Excel-specific */ } private addExcelFooter() { /* Excel-specific */ } private saveExcel() { /* Excel-specific */ } }

### After

abstract class Report { // Template method generate() { this.loadData(); this.formatHeader(); this.formatContent(); this.addFooter(); this.save(); }

// Common implementations protected loadData() { // Shared logic }

protected formatHeader() { // Shared logic }

// Abstract methods for subclasses protected abstract formatContent(): void; protected abstract addFooter(): void; protected abstract save(): void; }

class PDFReport extends Report { protected formatContent() { /* PDF-specific */ } protected addFooter() { /* PDF-specific */ } protected save() { /* PDF-specific */ } }

class ExcelReport extends Report { protected formatContent() { /* Excel-specific */ } protected addFooter() { /* Excel-specific */ } protected save() { /* Excel-specific */ } }

**DRY Improvement:** Shared logic in base class, variations in subclasses

---

## Pattern 4: Higher-Order Functions

**Problem:** Similar operations with different behaviors

**Solution:** Use higher-order functions or callbacks

### Before (JavaScript)

function processUsersForEmail(users) { const results = []; for (const user of users) { if (user.email && user.active) { results.push(user); } } return results; }

function processUsersForSMS(users) { const results = []; for (const user of users) { if (user.phone && user.active) { results.push(user); } } return results; }

function processUsersForPush(users) { const results = []; for (const user of users) { if (user.deviceToken && user.active) { results.push(user); } } return results; }

### After

function processUsers(users, channel) { const validators = { email: user => user.email && user.active, sms: user => user.phone && user.active, push: user => user.deviceToken && user.active };

return users.filter(validators[channel]); }

// Or more flexible with custom validator function processUsers(users, isValid) { return users.filter(isValid); }

// Usage const emailUsers = processUsers(users, user => user.email && user.active); const smsUsers = processUsers(users, user => user.phone && user.active); const pushUsers = processUsers(users, user => user.deviceToken && user.active);

**DRY Improvement:** 3 functions → 1 configurable function

---

## Pattern 5: Composition Over Duplication

**Problem:** Repeated utility combinations

**Solution:** Compose smaller utilities

### Before (Go)

// Repeated validation patterns func ValidateUser(user User) error { if user.Email == "" { return errors.New("email required") } if !strings.Contains(user.Email, "@") { return errors.New("invalid email") } if user.Age < 18 { return errors.New("must be 18+") } return nil }

func ValidateAdmin(admin Admin) error { if admin.Email == "" { return errors.New("email required") } if !strings.Contains(admin.Email, "@") { return errors.New("invalid email") } if admin.Age < 18 { return errors.New("must be 18+") } if admin.Role == "" { return errors.New("role required") } return nil }

### After

// Composable validators func requireEmail(email string) error { if email == "" { return errors.New("email required") } return nil }

func validateEmailFormat(email string) error { if !strings.Contains(email, "@") { return errors.New("invalid email") } return nil }

func requireAdult(age int) error { if age < 18 { return errors.New("must be 18+") } return nil }

func requireRole(role string) error { if role == "" { return errors.New("role required") } return nil }

// Compose validators func ValidateUser(user User) error { validators := []func() error{ func() error { return requireEmail(user.Email) }, func() error { return validateEmailFormat(user.Email) }, func() error { return requireAdult(user.Age) }, } return runValidators(validators) }

func ValidateAdmin(admin Admin) error { validators := []func() error{ func() error { return requireEmail(admin.Email) }, func() error { return validateEmailFormat(admin.Email) }, func() error { return requireAdult(admin.Age) }, func() error { return requireRole(admin.Role) }, } return runValidators(validators) }

func runValidators(validators []func() error) error { for _, validate := range validators { if err := validate(); err != nil { return err } } return nil }

**DRY Improvement:** Reusable validators, composable validation

---

## Pattern 6: Extract Test Helpers

**Problem:** Repeated test setup/teardown

**Solution:** Create test utilities

### Before

// test/user.test.js describe('User tests', () => { it('should create user', () => { const db = createDatabase(); const user = { name: 'John', email: 'john@example.com' }; // ... test logic db.close(); });

it('should update user', () => { const db = createDatabase(); const user = { name: 'John', email: 'john@example.com' }; // ... test logic db.close(); }); });

// test/order.test.js describe('Order tests', () => { it('should create order', () => { const db = createDatabase(); const user = { name: 'John', email: 'john@example.com' }; // ... test logic db.close(); }); });

### After

// test/helpers/testUtils.js export function setupTestDatabase() { const db = createDatabase(); return { db, cleanup: () => db.close() }; }

export function createTestUser(overrides = {}) { return { name: 'John', email: 'john@example.com', ...overrides }; }

// test/user.test.js import { setupTestDatabase, createTestUser } from './helpers/testUtils';

describe('User tests', () => { let db;

beforeEach(() => { ({ db } = setupTestDatabase()); });

afterEach(() => { db.close(); });

it('should create user', () => { const user = createTestUser(); // ... test logic (cleaner!) });

it('should update user', () => { const user = createTestUser({ name: 'Jane' }); // ... test logic }); });

**DRY Improvement:** Shared test utilities, less boilerplate

---

## obra YAGNI/DRY Principles

### YAGNI: You Aren't Gonna Need It

- Don't add functionality until necessary
- Avoid premature abstraction
- Wait for duplication before extracting

### DRY: Don't Repeat Yourself

- Every piece of knowledge should have a single representation
- Avoid copy-paste programming
- Extract when you see duplication 2-3 times

### When to Extract

1. **Three strikes rule**: Duplicate 3 times → extract
2. **Clear pattern**: Similar code structure appears
3. **High maintenance cost**: Changes require updates in multiple places
4. **Business logic**: Domain rules should be centralized

### When NOT to Extract

1. **Premature abstraction**: Only seen once or twice
2. **Coincidental duplication**: Similar code, different concepts
3. **Temporary code**: Prototypes, experiments
4. **Over-engineering**: Abstraction more complex than duplication

DRYPATTERNS

echo "✓ DRY refactoring patterns guide created"

Phase 5: Generate Duplication Report

I'll create a comprehensive report with prioritized refactoring opportunities:

echo ""
echo "=== Generating Duplication Report ==="
echo ""

# Count duplications
TOTAL_DUPLICATIONS=0
DUPLICATION_PERCENTAGE=0

if [ -f "$ANALYSIS_DIR/jscpd-report/jscpd-report.json" ]; then
    TOTAL_DUPLICATIONS=$(jq '.statistics.total.duplications' "$ANALYSIS_DIR/jscpd-report/jscpd-report.json" 2>/dev/null)
    DUPLICATION_PERCENTAGE=$(jq '.statistics.total.percentage' "$ANALYSIS_DIR/jscpd-report/jscpd-report.json" 2>/dev/null)
fi

cat > "$REPORT" << EOF
# Code Duplication Analysis Report

**Generated:** $(date)
**Minimum Lines Threshold:** $MIN_LINES lines
**Total Duplications Found:** $TOTAL_DUPLICATIONS
**Duplication Percentage:** $DUPLICATION_PERCENTAGE%

---

## Summary

Code duplication violates the DRY (Don't Repeat Yourself) principle and leads to:
- Higher maintenance costs (fix bugs in multiple places)
- Increased bug likelihood (miss updating one location)
- Larger codebase (more code to understand)
- Lower code quality (copy-paste programming)

**Duplication Guidelines (obra principles):**
- **0-5%:** Excellent - minimal duplication
- **5-10%:** Good - acceptable for large projects
- **10-20%:** Moderate - consider refactoring
- **20%+:** High - significant technical debt

---

## Duplication Statistics

EOF

if [ "$USE_JSCPD" = true ] && [ -f "$ANALYSIS_DIR/jscpd-report/jscpd-report.json" ]; then
    cat >> "$REPORT" << EOF
### Overall Statistics
- Total Lines Analyzed: $(jq '.statistics.total.lines' "$ANALYSIS_DIR/jscpd-report/jscpd-report.json" 2>/dev/null)
- Duplicated Lines: $(jq '.statistics.total.clones' "$ANALYSIS_DIR/jscpd-report/jscpd-report.json" 2>/dev/null)
- Duplication Blocks: $TOTAL_DUPLICATIONS
- Duplication Percentage: $DUPLICATION_PERCENTAGE%

### Top Duplicated Files

EOF
    jq -r '.statistics.formats[] |
        "\(.format):\n  Files: \(.sources)\n  Duplications: \(.duplications)\n  Percentage: \(.percentage)%\n"' \
        "$ANALYSIS_DIR/jscpd-report/jscpd-report.json" 2>/dev/null >> "$REPORT"

    cat >> "$REPORT" << EOF

### Largest Duplication Blocks

EOF
    jq -r '.duplicates[:10] | .[] |
        "**\(.lines) lines** duplicated in \(.format)\n- Location 1: \(.firstFile.name):\(.firstFile.start)\n- Location 2: \(.secondFile.name):\(.secondFile.start)\n"' \
        "$ANALYSIS_DIR/jscpd-report/jscpd-report.json" 2>/dev/null >> "$REPORT"

    cat >> "$REPORT" << EOF

**Full HTML Report:** Open \`$ANALYSIS_DIR/jscpd-report/html/index.html\` in browser

EOF
else
    cat >> "$REPORT" << EOF
### Basic Analysis Results

EOF
    if [ -f "$ANALYSIS_DIR/duplicate-functions.txt" ] && [ -s "$ANALYSIS_DIR/duplicate-functions.txt" ]; then
        cat >> "$REPORT" << EOF
**Duplicate Function Signatures (JavaScript/TypeScript):**
\`\`\`
$(cat "$ANALYSIS_DIR/duplicate-functions.txt")
\`\`\`

EOF
    fi

    if [ -f "$ANALYSIS_DIR/duplicate-python-functions.txt" ] && [ -s "$ANALYSIS_DIR/duplicate-python-functions.txt" ]; then
        cat >> "$REPORT" << EOF
**Duplicate Function Signatures (Python):**
\`\`\`
$(cat "$ANALYSIS_DIR/duplicate-python-functions.txt")
\`\`\`

EOF
    fi

    if [ -f "$ANALYSIS_DIR/magic-numbers.txt" ] && [ -s "$ANALYSIS_DIR/magic-numbers.txt" ]; then
        cat >> "$REPORT" << EOF
**Repeated Numeric Literals:**
\`\`\`
$(cat "$ANALYSIS_DIR/magic-numbers.txt")
\`\`\`
Consider extracting to named constants.

EOF
    fi

    if [ -f "$ANALYSIS_DIR/repeated-strings.txt" ] && [ -s "$ANALYSIS_DIR/repeated-strings.txt" ]; then
        cat >> "$REPORT" << EOF
**Repeated String Literals:**
\`\`\`
$(cat "$ANALYSIS_DIR/repeated-strings.txt")
\`\`\`
Consider extracting to configuration.

EOF
    fi
fi

cat >> "$REPORT" << 'EOF'
---

## Recommended DRY Refactoring

### Priority 1: Extract Duplicated Functions
- Identify exact code duplicates (10+ lines)
- Extract to shared utility functions
- Consolidate in common modules

### Priority 2: Centralize Configuration
- Extract magic numbers to constants
- Create configuration files
- Use environment variables for deployment-specific values

### Priority 3: Eliminate Similar Patterns
- Identify similar code structures
- Extract common patterns
- Use higher-order functions or templates

### Priority 4: Consolidate Test Helpers
- Extract common test setup/teardown
- Create test utilities
- Share fixtures across test suites

**See detailed patterns:** `cat $ANALYSIS_DIR/dry-refactoring-patterns.md`

---

## Implementation Steps

1. **Create Git Checkpoint**

git add -A git commit -m "Pre DRY-refactoring checkpoint" || echo "No changes"

1. **Prioritize Duplications**
  - Start with largest duplication blocks
  - Focus on frequently changed code
  - Consider domain importance
2. **Apply DRY Patterns**
  - Extract function (most common)
  - Extract configuration
  - Template method pattern
  - Higher-order functions
  - Composition
3. **Three Strikes Rule**
  - Wait for 3 occurrences before extracting
  - Avoid premature abstraction
  - Balance DRY with YAGNI
4. **Verify Improvements**
  - Re-run duplication analysis
  - Ensure tests pass
  - Check for reduced code size
5. **Document Changes**
  - Update documentation
  - Note refactoring decisions
  - Share patterns with team

---

## Continuous Monitoring

### Add to CI/CD

#### Pre-commit Hook

.git/hooks/pre-commit

#!/bin/bash

Check duplication before commit

jscpd --threshold $MIN_LINES . || exit 1

#### GitHub Actions

name: Code Quality on: [push, pull_request] jobs: duplication: runs-on: ubuntu-latest steps: - uses: actions/checkout@v2 - name: Check duplication run: | npm install -g jscpd jscpd --threshold 6 .

---

## Integration with Other Skills

- **`/refactor`** - Systematic code restructuring
- **`/complexity-reduce`** - Reduce cyclomatic complexity
- **`/make-it-pretty`** - Improve code readability
- **`/review`** - Include duplication in code reviews
- **`/test`** - Ensure tests after refactoring

---

## Resources

- [obra/superpowers - YAGNI/DRY principles](https://github.com/obra/superpowers)
- [The Pragmatic Programmer - DRY Principle](https://pragprog.com/titles/tpp20/)
- [Refactoring - Martin Fowler](https://refactoring.com/)
- [JSCPD - Copy-Paste Detector](https://github.com/kucherenko/jscpd)
- [Rule of Three (refactoring)](https://en.wikipedia.org/wiki/Rule_of_three_(computer_programming))

---

**Report generated at:** $(date)

**Next Steps:**

1. Review high-duplication areas
2. Select appropriate DRY pattern
3. Implement refactoring incrementally
4. Re-run analysis to verify improvement

EOF

echo "✓ Duplication report generated: $REPORT"

Summary

echo ""
echo "=== ✓ Duplication Analysis Complete ==="
echo ""
echo "📊 Analysis Results:"
if [ "$USE_JSCPD" = true ]; then
    echo "  Total duplications: $TOTAL_DUPLICATIONS"
    echo "  Duplication percentage: $DUPLICATION_PERCENTAGE%"
    echo "  Threshold: $MIN_LINES lines"
else
    echo "  Basic analysis complete"
    echo "  Install jscpd for detailed analysis: npm install -g jscpd"
fi
echo ""
echo "📁 Generated Files:"
echo "  - Duplication Report: $REPORT"
echo "  - DRY Patterns: $ANALYSIS_DIR/dry-refactoring-patterns.md"
[ -f "$ANALYSIS_DIR/jscpd-report/html/index.html" ] && echo "  - HTML Report: $ANALYSIS_DIR/jscpd-report/html/index.html"
echo ""
echo "🎯 Recommended Actions:"
if [ "$DUPLICATION_PERCENTAGE" != "0" ] && [ "${DUPLICATION_PERCENTAGE%.*}" -gt 10 ]; then
    echo "  ⚠️  High duplication detected ($DUPLICATION_PERCENTAGE%)"
    echo "  1. Review largest duplication blocks"
    echo "  2. Extract duplicated functions"
    echo "  3. Centralize configuration"
    echo "  4. Run tests after refactoring"
elif [ "$DUPLICATION_PERCENTAGE" != "0" ] && [ "${DUPLICATION_PERCENTAGE%.*}" -gt 5 ]; then
    echo "  Moderate duplication ($DUPLICATION_PERCENTAGE%)"
    echo "  Consider refactoring during feature work"
else
    echo "  ✓ Low duplication - excellent!"
    echo "  Continue monitoring in code reviews"
fi
echo ""
echo "💡 DRY Refactoring Patterns:"
echo "  1. Extract Function (consolidate duplicates)"
echo "  2. Extract Configuration (centralize constants)"
echo "  3. Template Method (share algorithm structure)"
echo "  4. Higher-Order Functions (parameterize behavior)"
echo "  5. Composition (build from smaller pieces)"
echo ""
echo "📏 obra YAGNI/DRY Guidelines:"
echo "  - Three strikes rule: Extract after 3 duplications"
echo "  - Avoid premature abstraction"
echo "  - Balance DRY with readability"
echo ""
echo "🔗 Integration Points:"
echo "  - /refactor - Systematic restructuring"
echo "  - /complexity-reduce - Reduce complexity"
echo "  - /review - Include in code reviews"
echo ""
echo "View report: cat $REPORT"
echo "View patterns: cat $ANALYSIS_DIR/dry-refactoring-patterns.md"
[ -f "$ANALYSIS_DIR/jscpd-report/html/index.html" ] && echo "View HTML: open $ANALYSIS_DIR/jscpd-report/html/index.html"

Safety Guarantees

What I'll NEVER do:

• Automatically refactor without analysis
• Extract after single occurrence (violates YAGNI)
• Remove code without understanding context
• Skip testing after refactoring
• Add AI attribution to commits

What I WILL do:

• Identify genuine code duplication
• Suggest proven DRY patterns
• Follow obra YAGNI principles
• Recommend incremental refactoring
• Preserve functionality
• Provide clear examples

Credits

This skill is based on:

• obra/superpowers - YAGNI and DRY principles
• The Pragmatic Programmer - DRY principle origin
• Martin Fowler - Refactoring patterns
• JSCPD - Multi-language duplication detection
• Rule of Three - When to extract abstraction

Token Budget

Target: 2,000-3,500 tokens per execution

• Phase 1-2: ~600 tokens (setup + tool installation)
• Phase 3: ~1,000 tokens (duplication detection)
• Phase 4-5: ~1,500 tokens (patterns + reporting)

Optimization Strategy:

• Use jscpd for efficient detection
• Fallback to grep-based analysis
• Template-based pattern generation
• Focus on actionable duplications
• Clear DRY refactoring guidance

This ensures thorough duplication detection with practical DRY refactoring strategies based on obra principles while respecting token limits.