nanparth/ai-skill-hub

This repository is a collection of "skills" for AI assistant tools, shared from someone's personal workflows. A skill in this context is a folder containing a structured instruction file (called SKILL.md) that tells an AI coding or agent tool how to handle a specific type of task. Once installed, you can invoke the skill by name and the AI follows the instructions inside it. The format is based on a shared specification called Agent Skills, which is supported by tools like Claude Code, GitHub Copilot CLI, Cursor, and OpenAI Codex. The nine skills currently included cover a range of work types. A few are oriented toward research: one researches organizations into structured profile notes, another does the same for individuals. A business interview skill helps create and analyze customer discovery interview scripts and qualitative reports. A legal diagramming skill generates visual diagrams from legal documents, with support for French-language source material. A job application skill tailors cover letters, CVs, and interview preparation materials from supplied source content. On the technical side, there are skills for writing specs and product requirement documents through an interactive drafting session, running a test-driven implementation or bug-fix process with review gates, and auditing code structure to produce refactoring roadmaps. A general note-taking skill handles creating, merging, and reformatting Markdown notes from various source material. Installing a skill depends on which AI tool you use. For terminal-based tools like Claude Code, you copy the skill folder into a hidden skills directory on your computer. For Claude in the browser, you zip the folder and upload it in settings. For VS Code with GitHub Copilot, you add the skill folder inside a project directory. The README includes animated GIF walkthroughs for each installation method, aimed at users who are not familiar with GitHub or command-line tools. The author notes these are personal tools released as-is, stable enough to share but not polished products. Customization before relying on any skill for serious work is recommended.