Overview
This repository offers a curated collection of AI agent "skills" designed for real-world software engineering. Moving beyond superficial "vibe coding," these skills focus on robust development practices, drawing from decades of engineering wisdom. They are built to be small, adaptable, and compatible with any AI model, empowering you to tackle complex applications with confidence.
Key Features
- Practical Engineering Focus: Directly addresses common AI development pitfalls like misalignment, verbosity, and code quality issues with proven engineering principles.
- Small, Composable Skills: Each skill is designed to be independent and easily integrated, allowing for flexible customization and adaptation to your specific workflow.
- Seamless Integration: Offers two installation methods: a managed, auto-updating plugin for Claude Code, or direct file copies for tinkerers who prefer full control.
- Core Skills for Alignment & Clarity:
/grill-meand/grill-with-docs: Revolutionize your planning process by ensuring deep understanding and alignment with the AI before coding begins. The latter also builds a shared project vocabulary and documents key decisions.- Shared Language Development: Leverages a project
CONTEXT.mdto drastically reduce AI verbosity, improve codebase navigation, and enhance token efficiency.
- Robust Feedback Loops:
/tdd: Implements a strict red-green-refactor cycle to ensure code quality and predictable development./diagnosing-bugs: Provides a structured, phase-gated approach to debugging complex issues.
- Architectural Integrity: Skills like
/improve-codebase-architectureand/to-specencourage proactive attention to design, combating the rapid entropy often seen in AI-generated codebases. - Extensive Reference Guide: A detailed breakdown of user-invoked and model-invoked skills, categorized by "Engineering" and "Productivity," with clear explanations of their purpose and usage.
Typical Use Cases
- Kickstarting New Projects: Use
/setup-matt-pocock-skillsfollowed by/grill-with-docsto establish a solid foundation, define project vocabulary, and align your AI agent with your vision from day one. - Improving Code Quality: Integrate
/tddinto your development workflow to ensure comprehensive test coverage and reduce regressions. Leverage/diagnosing-bugsto systematically tackle challenging defects. - Refactoring and Architecture: Employ
/improve-codebase-architectureto identify and address opportunities for deepening your codebase's design, leading to more maintainable and understandable software. - Enhancing Developer-AI Collaboration: Utilize
/grill-meor/grill-with-docsfor any significant task to ensure the AI understands your intent precisely, preventing costly misunderstandings. - Streamlining Workflow: Employ productivity skills like
/handofffor seamless transitions between AI sessions or/to-questionnaireto efficiently gather information from domain experts.