mattpocock/skills

This is the open-source repository behind GitHub's official help and documentation website, where anyone can contribute edits, corrections, or improvements to the written guides that millions of developers rely on daily. GitHub publishes its own instruction manuals publicly, allowing its user community to help keep documentation accurate and up to date.

why it matters: With over 68,000 forks and 3,000 contributors, this project signals that even the world's largest developer platform treats its documentation as a community-owned product — a model that reduces internal maintenance costs while building deep user loyalty. For founders, it's a reminder that open-sourcing support and documentation content can turn users into contributors and dramatically scale your knowledge base without proportional headcount.

LLVM is the foundational technology that turns code written by developers into software that computers can actually run — it's the engine underneath many of the world's most popular programming languages and compilers, including Apple's Swift and the Clang C++ compiler. With nearly 40,000 GitHub stars and close to 9,000 contributors, it's one of the most widely used open-source infrastructure projects in existence.

why it matters: If you're building a new programming language, a code analysis tool, or any software that needs to run fast on multiple platforms, LLVM is the industry-standard foundation that saves years of low-level engineering work. Companies like Apple, Google, and Meta rely on it, meaning understanding and contributing to LLVM can be a significant competitive advantage for teams building developer tools or language runtimes.

Code-graph-rag is an AI-powered tool that lets developers ask plain-language questions about large, complex codebases and get accurate answers — think of it as a smart search engine that truly understands your entire software project, not just keywords. It works across multiple programming languages and builds a relationship map of your code so an AI assistant can explain, find, and even suggest edits across the whole system.

why it matters: As codebases grow, onboarding new engineers and maintaining productivity becomes a major bottleneck and cost center — this tool attacks that problem directly by making any codebase instantly navigable for both humans and AI agents. For founders and technical leaders, it signals a maturing market for 'AI coding assistants that actually understand context,' which is becoming a key competitive differentiator in developer tooling.

Codex is an AI-powered assistant that runs directly on your computer and helps write, edit, and manage code through a simple terminal interface — essentially a smart coding helper that works from your command line without needing a browser or separate app. It's OpenAI's lightweight, locally-run version of their coding AI, giving developers a fast way to get AI assistance without leaving their existing workflow.

why it matters: With over 113,000 stars, this is one of the most watched open-source projects on GitHub right now, signaling massive developer appetite for AI tools that work locally rather than purely in the cloud — a meaningful shift for teams concerned about privacy, speed, or cost. For founders and product teams, this raises the bar for any coding tool or developer-facing product, as AI assistance is quickly becoming a baseline expectation rather than a premium feature.