Claude Code Plugins

Community-maintained marketplace

Feedback

ai-code-cleanup

@89jobrien/steve
1
0

Remove AI-generated code slop from branches. Use after AI-assisted coding

Install Skill

1Download skill
2Enable skills in Claude

Open claude.ai/settings/capabilities and find the "Skills" section

3Upload to Claude

Click "Upload skill" and select the downloaded ZIP file

Note: Please verify skill by going through its instructions before using it.

SKILL.md

name ai-code-cleanup
description Remove AI-generated code slop from branches. Use after AI-assisted coding sessions to clean up defensive bloat, unnecessary comments, type casts, and style inconsistencies. Focuses on identifying and removing AI artifacts that degrade code quality.
author Joseph OBrien
status unpublished
updated 2025-12-23
version 1.0.1
tag skill
type skill

AI Code Cleanup

This skill identifies and removes AI-generated artifacts that degrade code quality, including defensive bloat, unnecessary comments, type casts, and style inconsistencies.

When to Use This Skill

  • After AI-assisted coding sessions
  • Before code reviews or merging branches
  • When cleaning up code that feels "over-engineered"
  • When removing unnecessary defensive code
  • When standardizing code style after AI generation
  • When preparing code for production

What This Skill Does

  1. Identifies AI Artifacts: Detects patterns typical of AI-generated code
  2. Removes Bloat: Eliminates unnecessary defensive code and comments
  3. Fixes Type Issues: Removes unnecessary type casts and workarounds
  4. Standardizes Style: Ensures consistency with project conventions
  5. Preserves Functionality: Maintains code behavior while improving quality
  6. Validates Changes: Ensures code still compiles and tests pass

How to Use

Clean Up Branch

Remove AI slop from this branch
Clean up the code in this pull request

Specific Cleanup

Remove unnecessary comments and defensive code from src/

Slop Patterns to Remove

1. Unnecessary Comments

Patterns:

  • Comments explaining obvious code
  • Comments inconsistent with file's documentation style
  • Redundant comments that restate the code
  • Over-documentation of simple operations

Example:

// ❌ AI-generated: Obvious comment
// Set the user's name
user.name = name;

// ✅ Clean: Self-documenting code
user.name = name;

2. Defensive Bloat

Patterns:

  • Extra try/catch blocks abnormal for that codebase
  • Defensive null/undefined checks on trusted paths
  • Redundant input validation when callers already validate
  • Error handling that can never trigger

Example:

// ❌ AI-generated: Unnecessary defensive code
function processUser(user) {
  try {
    if (user && user.name && typeof user.name === 'string') {
      return user.name.toUpperCase();
    }
    return null;
  } catch (error) {
    console.error(error);
    return null;
  }
}

// ✅ Clean: Trust the input, handle real errors
function processUser(user) {
  return user.name.toUpperCase();
}

3. Type Workarounds

Patterns:

  • Casts to any to bypass type issues
  • Unnecessary type assertions (as X)
  • @ts-ignore or @ts-expect-error without legitimate reason
  • Overly complex generic constraints

Example:

// ❌ AI-generated: Type workaround
const data = response.data as any;
const result = processData(data as ProcessedData);

// ✅ Clean: Proper typing
const data = response.data;
const result = processData(data);

4. Style Inconsistencies

Patterns:

  • Naming conventions different from rest of file
  • Formatting that doesn't match surrounding code
  • Import organization inconsistent with file patterns
  • Variable declarations inconsistent with file style

Example:

// ❌ AI-generated: Inconsistent style
const UserData = fetchUser();
var processedData = process(UserData);
let FinalResult = format(processedData);

// ✅ Clean: Consistent style
const userData = fetchUser();
const processedData = process(userData);
const finalResult = format(processedData);

5. AI Tells

Patterns:

  • Unnecessary emoji usage in code or comments
  • Overly verbose variable names
  • Redundant intermediate variables
  • "Just in case" code with no actual use case

Cleanup Process

1. Get the Diff

git diff $(git merge-base HEAD main)..HEAD --name-only

2. Analyze Each File

For each changed file:

  • Read the full file to understand existing style
  • Identify patterns from slop categories
  • Make surgical edits to remove slop
  • Preserve functionality

3. Verification

  • Ensure code still compiles/runs
  • Changes should reduce code, not add more
  • Run tests to verify behavior
  • Check for any regressions

Examples

Example 1: Comment Cleanup

Input: Remove unnecessary comments from src/auth.js

Output:

## AI Code Cleanup: src/auth.js

### Removed
- 12 redundant comments explaining obvious code
- 3 comments inconsistent with file style
- 5 over-documented simple operations

### Changes
```javascript
// Before
// Function to validate user email
// Takes email string as input
// Returns boolean indicating validity
function validateEmail(email) {
  // Check if email is not empty
  if (!email) {
    return false; // Return false if empty
  }
  // Return regex test result
  return /^[^\s@]+@[^\s@]+\.[^\s@]+$/.test(email);
}

// After
function validateEmail(email) {
  if (!email) return false;
  return /^[^\s@]+@[^\s@]+\.[^\s@]+$/.test(email);
}

Result

  • Lines reduced: 12 → 4
  • Functionality preserved
  • Tests pass

### Example 2: Defensive Bloat Removal

**Input**: Remove defensive bloat from `src/api/users.js`

**Output**:
```markdown
## AI Code Cleanup: src/api/users.js

### Removed
- 3 unnecessary try/catch blocks
- 8 redundant null checks on validated inputs
- 2 error handlers that can never trigger

### Changes
```javascript
// Before
async function getUser(userId) {
  try {
    if (!userId || typeof userId !== 'string') {
      throw new Error('Invalid userId');
    }
    const user = await db.users.findById(userId);
    if (user && user.id) {
      return user;
    }
    return null;
  } catch (error) {
    console.error(error);
    throw error;
  }
}

// After
async function getUser(userId) {
  const user = await db.users.findById(userId);
  return user || null;
}

Result

  • Code reduced: 15 lines → 3 lines
  • Functionality preserved
  • Error handling appropriate for context

## Reference Files

- **`references/REFACTORING_PLAN.template.md`** - Refactoring plan template with code smells, before/after metrics, and rollback strategy

## Best Practices

### Cleanup Guidelines

1. **Preserve Functionality**: Only remove code that doesn't affect behavior
2. **Maintain Style**: Follow existing project conventions
3. **Keep Real Errors**: Don't remove legitimate error handling
4. **Test After Changes**: Always verify code still works
5. **Incremental**: Make changes incrementally, test as you go

### What to Keep

- Legitimate error handling
- Necessary type assertions
- Helpful comments that add context
- Defensive code for untrusted inputs
- Style that matches the codebase

### What to Remove

- Obvious comments
- Unnecessary defensive code
- Type workarounds
- Style inconsistencies
- AI-generated artifacts

## Related Use Cases

- Post-AI coding cleanup
- Code review preparation
- Code quality improvement
- Style standardization
- Removing technical debt