Claude Code Plugins

Community-maintained marketplace

Feedback

best-practices

@claimhawk/shebang
0
0

Generate technology-specific best practices and anti-patterns documentation. Use when setting up a new project or when the user asks about coding standards for a specific language or framework.

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 best-practices
description Generate technology-specific best practices and anti-patterns documentation. Use when setting up a new project or when the user asks about coding standards for a specific language or framework.

Best Practices Generator

Creates customized best practices and anti-patterns documentation based on the project's technology stack.

When to Use

  • After /new creates a project
  • When user asks about best practices for a language
  • When detecting a new technology in the project
  • When user wants coding standards documented

Detection

Detect technology from:

  • File extensions (.py, .ts, .swift, .go, .rs)
  • Config files (package.json, pyproject.toml, Cargo.toml)
  • Framework markers (next.config.js, vite.config.ts)

Output Files

Create in .shebang/ directory:

best-practices.md

# Best Practices: [Technology]

## Code Organization
- [Patterns for this technology]

## Error Handling
- [Error patterns for this language]

## Testing
- [Testing strategies]

## Performance
- [Performance considerations]

## Security
- [Security practices]

anti-patterns.md

# Anti-Patterns: [Technology]

## Code Smells
- [Common mistakes]

## Performance Anti-Patterns
- [Things that hurt performance]

## Security Vulnerabilities
- [Common security issues]

## Testing Anti-Patterns
- [Testing mistakes]

Technology-Specific Content

Python

  • Type hints everywhere
  • Use dataclasses/Pydantic
  • Avoid bare except
  • Use context managers
  • pytest over unittest

TypeScript

  • Strict mode always
  • No any types
  • Prefer interfaces
  • Use discriminated unions
  • Proper error boundaries in React

Swift

  • Use optionals properly
  • Avoid force unwrapping
  • Protocol-oriented design
  • Structured concurrency

Go

  • Error handling patterns
  • Context propagation
  • Interface composition
  • Goroutine management

Rust

  • Ownership patterns
  • Error handling with Result
  • Lifetime management
  • Unsafe usage guidelines

Workflow

  1. Detect tech - Scan project files
  2. Load base template - From docs/testing/ if available
  3. Customize - Add project-specific patterns
  4. Write files - Create in .shebang/
  5. Report - Tell user what was created

Example

Agent detects: Python + FastAPI + PostgreSQL

Creates:
- .shebang/best-practices.md (Python/FastAPI patterns)
- .shebang/anti-patterns.md (Common FastAPI mistakes)

Includes:
- Async/await patterns
- Dependency injection
- Database session handling
- Request validation