Skip to content

Contributing

Thank you for your interest in contributing to MCP Sub-Agents! This guide will help you understand how to contribute effectively.

Ways to Contribute

1. Report Issues

Found a bug or have a suggestion? Open an issue:

When reporting issues, include:

  • Clear description of the problem
  • Steps to reproduce
  • Expected vs actual behavior
  • Agent name and command used
  • Error messages (if any)
  • Environment details (OS, Node.js version)

2. Improve Documentation

Help improve the documentation:

  • Fix typos and grammar
  • Add examples
  • Clarify explanations
  • Add missing information
  • Improve code examples

3. Add New Agents

Create new specialist or savant agents:

  • Identify need for new agent
  • Follow agent template
  • Ensure comprehensive coverage
  • Include examples
  • Test thoroughly

4. Create Workflow Commands

Add new workflow commands:

  • Identify common workflow patterns
  • Create structured guidance
  • Include checklists
  • Provide time estimates
  • Test with real scenarios

5. Enhance Existing Agents

Improve existing agents:

  • Add new patterns
  • Update for latest versions
  • Improve examples
  • Clarify guidance
  • Fix errors

Getting Started

1. Fork the Repository

# Fork on GitHub, then clone
git clone https://github.com/YOUR_USERNAME/mcp-sub-agents.git
cd mcp-sub-agents

2. Create a Branch

git checkout -b feature/your-feature-name
# or
git checkout -b fix/your-bug-fix

3. Make Changes

Edit the relevant files and test your changes.

4. Commit Changes

git add .
git commit -m "Description of changes"

5. Push and Create PR

git push origin feature/your-feature-name

Then create a Pull Request on GitHub.

Adding New Agents

Agent Template

Create a new markdown file in the appropriate directory:

---
name: agent-name
description: Brief description of agent expertise and when to use it
tools: Read, Write, Edit, Bash
model: sonnet
---

You are a [Agent Type] specializing in [Domain].

## Core Expertise

### Category 1
- Technology A - Key features
- Technology B - Key features

### Category 2
- Technology C - Key features

## When to Use This Agent

Use this agent when you need:
- Use case 1
- Use case 2
- Use case 3

## Best Practices

- Best practice 1
- Best practice 2

## Code Examples

### Example 1: [Description]

```[language]
// Code example

Example 2: [Description]

```[language] // Code example

## Common Patterns

Pattern descriptions and examples.

Agent Placement

Savant Agents

Place in agents/savants/:

agents/savants/savant-[stack]-name.md

Example: agents/savants/savant-fullstack-js.md

Stack-Specific Specialists

Place in appropriate stack directory:

agents/[stack]/specialist-name.md

Example: agents/javascript/fullstack-nextjs.md

Cross-Cutting Specialists

Place in agents/cross-cutting/:

agents/cross-cutting/specialist-name.md

Example: agents/cross-cutting/security.md

Agent Naming Conventions

  • Savants: savant-[stack]-[focus] (e.g., savant-fullstack-js)
  • Specialists: [technology]-[focus] (e.g., fullstack-nextjs)
  • Cross-cutting: [domain] (e.g., security, performance)

Agent Requirements

New agents must include:

  • Clear, specific name in frontmatter
  • Comprehensive description
  • Core expertise areas
  • When to use guidance
  • Best practices
  • Code examples (at least 3)
  • Common patterns
  • 800+ lines of quality guidance

Adding Workflow Commands

Command Template

Create a new markdown file in .claude/commands/:

---
description: Brief description of what this workflow command does
---

# Workflow Command Title

Brief overview of the command.

## Purpose

What this command helps you accomplish.

## Workflow

Step-by-step workflow:

1. **Step 1 Name**
    - Description
    - What agent to use
    - Expected output

2. **Step 2 Name**
    - Description
    - What agent to use
    - Expected output

## Checklist

- [ ] Requirement 1
- [ ] Requirement 2
- [ ] Requirement 3

## Time Estimate

Estimated time: [X-Y] minutes

## Example Usage

Example of running this command with expected results.

## Next Steps

What to do after completing this workflow.

Command Naming

Use format: mcp-[phase]-[technology]

Examples:

  • mcp-design-architecture
  • mcp-implement-fullstack
  • mcp-review-security
  • mcp-qa-e2e

Documentation Improvements

Documentation Structure

docs/
├── index.md                    # Home page
├── getting-started/
│   ├── installation.md
│   └── quick-start.md
├── architecture/
│   ├── overview.md
│   └── workflow-commands.md
├── agents/
│   ├── savants.md
│   ├── javascript-specialists.md
│   └── cross-cutting-specialists.md
├── guides/
│   ├── workflows.md
│   └── best-practices.md
├── reference/
│   └── tech-stack.md
├── contributing.md
└── faq.md

Documentation Guidelines

  • Use clear, concise language
  • Include code examples
  • Add diagrams where helpful (Mermaid)
  • Cross-reference related pages
  • Keep formatting consistent
  • Use admonitions for important notes
  • Ensure mobile responsiveness

Mermaid Diagrams

Use Mermaid for architecture and flow diagrams:

```mermaid
graph TB
    A[Start] --> B[Process]
    B --> C[End]
```

Code Quality Standards

TypeScript

  • Use strict mode
  • Provide complete type definitions
  • Avoid any type
  • Use proper interfaces

Code Examples

  • Production-ready code
  • Include error handling
  • Add comments for complex logic
  • Follow best practices
  • Test all examples

Formatting

  • Use Prettier for formatting
  • Follow existing code style
  • Consistent indentation (2 spaces)
  • Line length max 100 characters

Testing

Test Your Changes

Before submitting:

  1. Agent Testing

    • Test agent invocation
    • Verify output quality
    • Check examples work
    • Test edge cases
  2. Command Testing

    • Run through entire workflow
    • Verify all steps work
    • Check time estimates
    • Test with real scenarios
  3. Documentation Testing

    • Build documentation locally
    • Check for broken links
    • Verify code examples
    • Test on mobile

Build Documentation Locally

# Install MkDocs
pip install mkdocs mkdocs-material

# Serve locally
mkdocs serve

# Build static site
mkdocs build

Pull Request Process

Before Submitting

  • Test your changes thoroughly
  • Update relevant documentation
  • Add examples if applicable
  • Check for typos and formatting
  • Ensure consistent style
  • Verify all links work

PR Description

Include in your PR:

  • What: What changes were made
  • Why: Reason for changes
  • How: How to test the changes
  • Related: Link to related issues

PR Template

## Description
Brief description of changes

## Type of Change
- [ ] New agent
- [ ] Workflow command
- [ ] Documentation
- [ ] Bug fix
- [ ] Enhancement

## Testing
How to test these changes

## Checklist
- [ ] Tested thoroughly
- [ ] Documentation updated
- [ ] Examples included
- [ ] Follows style guide
- [ ] No breaking changes

Code of Conduct

Be Respectful

  • Treat everyone with respect
  • Welcome newcomers
  • Accept constructive criticism
  • Focus on what's best for the community

Be Collaborative

  • Help others learn
  • Share knowledge
  • Give credit where due
  • Support fellow contributors

Be Professional

  • Use welcoming language
  • Avoid inflammatory comments
  • Stay on topic
  • Assume good intentions

Review Process

What to Expect

  1. Initial Review - Maintainer reviews within 1-3 days
  2. Feedback - Suggestions for improvements
  3. Iteration - Make requested changes
  4. Approval - PR approved and merged
  5. Release - Changes included in next release

Review Criteria

PRs are evaluated on:

  • Code quality and correctness
  • Documentation completeness
  • Test coverage
  • Consistency with existing style
  • Value added to the project

Getting Help

Resources

Questions

If you have questions:

  1. Check existing documentation
  2. Search closed issues
  3. Ask in GitHub Discussions
  4. Create a new issue

Recognition

Contributors will be:

  • Listed in CONTRIBUTORS.md
  • Mentioned in release notes
  • Acknowledged in documentation

Thank you for contributing to MCP Sub-Agents!