Claude Code Plugins

Community-maintained marketplace

Feedback

code-commenter

@luohaha/oh-my-skills
0
0

Add detailed comments explaining WHY code works, its context, edge cases, and performance implications. Use when user mentions "add comments", "document", "explain code", "improve readability", or "maintainability". Supports C++, Python, Java, JavaScript, Go, Rust, SQL.

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 code-commenter
description Add detailed comments explaining WHY code works, its context, edge cases, and performance implications. Use when user mentions "add comments", "document", "explain code", "improve readability", or "maintainability". Supports C++, Python, Java, JavaScript, Go, Rust, SQL.

Code Commenter

Add high-quality comments that explain WHY code works, not just WHAT it does.

Core Principles

1. Explain WHY, not WHAT

Explain reasoning and business context, not just what the code does.

// Bad: Increment counter
counter++;

// Good: Track failed connection attempts for exponential backoff
counter++;

2. Document Non-Obvious Decisions

Explain performance trade-offs, algorithm choices, constraints.

// Use bloom filter for memory efficiency: 10x memory savings, false positives acceptable
// for duplicate detection. Hash set would be more accurate but needs 10x more RAM.

3. Highlight Edge Cases and Gotchas

// IMPORTANT: Don't call from within a transaction - opens its own tx, causing deadlock
// NOTE: This mutex must be held when accessing sharedCache (race condition otherwise)

4. Document Assumptions

// ASSUMES: Input already sanitized, max 1000 chars
// CONSTRAINT: Parser doesn't handle nested parens beyond depth 3

Comment Types

File/Module Level: Purpose, dependencies, constraints Function/Method: Purpose, parameters, side effects, performance notes Inline: Explain specific lines when non-obvious Section: Break long functions into logical sections TODO/FIXME: Track technical debt

See EXAMPLES.md for detailed examples of each type.

Language-Specific Guidelines

C/C++: Doxygen style, explain memory/thread safety Python: Docstrings (PEP 257), type hints, examples Java: JavaDoc, thread safety, exceptions
JavaScript/React: Props, state, performance notes Go: Exported symbols, concurrency, context Rust: Doc comments, ownership, lifetimes SQL: Join logic, performance, business rules

For detailed examples, see LANGUAGE_GUIDE.md.

What NOT to Comment

Don't state the obvious: x = 5; doesn't need a comment saying "set x to 5" Don't comment bad code - refactor it: Make code self-documenting with good names Don't maintain outdated comments: Delete rather than keep misleading comments

Output Format

  1. Preserve original code structure (don't reformat unless requested)
  2. Add comments naturally where they make logical sense
  3. Match existing comment style in the codebase
  4. After commenting, provide brief summary of what was documented

Example:

I've added comments to explain:
- Algorithm complexity and trade-offs
- Edge case handling for empty inputs  
- Memory management strategy
- Non-obvious performance optimizations

Best Practices

DO:

  • Explain WHY and provide context
  • Document assumptions, constraints, edge cases
  • Use clear, concise language
  • Keep comments close to relevant code
  • Update comments when code changes

DON'T:

  • State the obvious
  • Comment bad code instead of fixing it
  • Write novels - be concise
  • Leave outdated comments
  • Comment out code (use version control)