Thank you for your interest in contributing! This repository contains an Agent Skill following the Agent Skills open format, which means it has specific structural requirements beyond typical documentation.
Agent Skills are modular capabilities that extend AI agents' functionality. They package instructions, metadata, and resources in a specific format that agents can discover and use automatically. This isn't just documentation—it's a structured skill that AI coding assistants load on-demand.
Learn more:
- Agent Skills Overview - Understanding the format
- Anthropic's Agent Skills Documentation - Official specification and best practices
- GitHub Copilot Agent Skills - Copilot-specific guidance
Contributors who aren't familiar with the Agent Skills format may inadvertently:
- Break the YAML frontmatter structure
- Use incorrect markdown formatting that agents can't parse
- Add content that doesn't follow agent-friendly patterns
- Miss important metadata requirements
That's why we strongly recommend using AI assistance for contributions.
The best way to contribute is to use Claude with the skill-creator skill. This ensures your contribution follows the Agent Skills format correctly.
How to use it:
-
Install the skill-creator skill in your AI tool (Claude.ai, VS Code with Copilot, Cursor, Gemini CLI, etc.)
# If using skills.sh npx skills add https://github.com/anthropics/skills --skill skill-creator -
Ask Claude to help with your contribution:
Use the skill-creator skill to help me update the Swift Testing Agent Skill. I want to [describe your contribution]. -
Review the generated changes to ensure technical accuracy for Swift Testing
-
Submit your PR with the AI-generated improvements
Benefits:
- Proper YAML frontmatter structure
- Agent-friendly markdown formatting
- Correct file organization
- Follows best practices from Anthropic's guidelines
If you don't have access to the skill-creator skill, you can still use Claude (or other AI assistants) to help:
-
Share the Agent Skills documentation with Claude:
-
Ask Claude to review your changes for Agent Skills format compliance
-
Request that Claude verify:
- YAML frontmatter is valid
- Markdown structure is agent-friendly
- Content follows progressive disclosure patterns
If you prefer to contribute manually, please carefully review:
- The Agent Skills specification: https://agentskills.io/home
- Anthropic's best practices: https://platform.claude.com/docs/en/agents-and-tools/agent-skills/best-practices
- Existing file structure in this repository
Key requirements:
- YAML frontmatter must be valid and complete
- Use clear, actionable headings
- Keep instructions concise and agent-friendly
- Follow the existing reference file patterns
- Test your changes with an AI agent if possible
We welcome contributions in these areas:
- Incorrect Swift Testing advice
- Outdated code examples
- Broken links or references
- Typos or clarity improvements
- Additional Swift Testing patterns
- New error scenarios and solutions
- Swift Testing framework updates
- Real-world examples
- Better decision trees
- Clearer triage guidance
- Updated agent behavior rules
- README improvements
- Installation instructions
- Usage examples
All contributions should meet these standards:
- Follows Swift Testing best practices
- Includes working code examples
- Cites official Swift documentation when relevant
- Avoids deprecated patterns
- Considers F.I.R.S.T. principles
- Valid YAML frontmatter (if modifying SKILL.md)
- Agent-friendly markdown structure
- Clear, actionable instructions
- Proper file organization
- Progressive disclosure (don't overload main SKILL.md)
- Test code examples compile (Swift 6 mode preferred)
- Verify advice with an AI agent if possible
- Check that links work
-
Fork the repository and create a new branch
git checkout -b feature/your-contribution
-
Make your changes (preferably using the skill-creator skill or Claude)
-
Test your changes:
- Verify code examples compile
- Check markdown formatting
- Test with an AI agent if possible
-
Commit with a clear message:
git commit -m "Add guidance for Swift Testing confirmation patterns" -
Submit a pull request with:
- Clear description of what you changed and why
- Rationale for the change (link to Swift documentation, real-world scenario, etc.)
- Mention if you used skill-creator or Claude to ensure format compliance
- Test results if applicable
-
Respond to feedback - We may request changes to ensure accuracy or format compliance
To test your changes locally:
-
Clone the repository:
git clone https://github.com/bocato/swift-testing-agent-skill.git cd swift-testing-agent-skill -
Install the skill in your AI tool (see README.md)
-
Test with your AI agent:
Use the swift-testing skill and [test your specific change]
- Agent Skills Home - Format overview
- Anthropic's Agent Skills Guide - Official documentation
- Agent Skills Best Practices - Writing effective skills
- GitHub Copilot Agent Skills - Copilot documentation
- Gemini CLI Documentation - Gemini CLI guide
- skill-creator skill - AI-assisted skill authoring
- Swift Testing Documentation - Official Apple documentation
- Martin Fowler's Mocks Aren't Stubs - Test double taxonomy
- PointFree's SnapshotTesting - Snapshot testing library
- Be respectful and constructive
- Focus on technical accuracy
- Welcome newcomers and help them learn
- Assume good intentions
- About Agent Skills format: See agentskills.io or Anthropic's docs
- About Swift Testing: Check the references folder or Apple's documentation
- About this skill: Open an issue for discussion
By contributing, you agree that your contributions will be licensed under the same MIT License that covers this project.
Thank you for helping make Swift Testing more accessible to AI agents and developers!