Yazinc AI Toolkit
Looking for the designed homepage? Open the Yazinc AI Toolkit landing page. This repository contains the source files and installable skills.
Public AI-agent skill library created by Yasir Ikram and maintained by Yaz Inc. It provides reusable skills, patterns, templates, and scripts for Codex, Cursor, Claude Code, Antigravity, and other Agent Skills-compatible tools. The library is designed to keep growing as new, proven workflows are developed.
Easy to understand. Easy to use.
The interactive toolkit makes every skill approachable for first-time developers and experienced teams alike. Each skill card provides three clear actions:
- Read instructions to understand the complete workflow and its safety rules.
- Copy skill instructions to place the full skill guidance on the clipboard instantly.
- How to use to open a short guide with the invocation, three simple steps, and a ready-to-copy example prompt.
There is no need to memorize commands, installation folders, or agent-specific syntax. Install the toolkit once, choose a skill, copy the suggested prompt, and describe the result you need in your active project.
Smart implementation playbooks
The landing page also provides a small, curated set of outcome-based technology playbooks. Each one explains what the solution pattern is, how to apply it safely, and provides a copy-ready starter prompt with official references. Playbooks are guidance, not installable skills, so the main skills catalog remains focused and easy to navigate.
Reusable prompts: Use prompts/google-sans-website-agent.md to instruct a development agent to implement and verify Google Sans correctly.
What & why
| |
|---|
| What | Central repo for agent skills and proven dev patterns |
| Why | Start each new project from scaffolding you already trust; push learnings back after each delivery |
| Who | Developers and AI-agent users who need reusable, inspectable workflows |
For humans
- Clone and install everything:
git clone https://github.com/Yaz-inc/yazinc-ai-toolkit.git; cd yazinc-ai-toolkit; powershell -NoProfile -ExecutionPolicy Bypass -File .\scripts\install-toolkit.ps1 -Target all
- Open a new Codex, Cursor, Claude Code, or Antigravity session.
- Ask:
Use the test-software-app skill to audit this application before launch.
- To update later:
git pull; powershell -NoProfile -ExecutionPolicy Bypass -File .\scripts\install-toolkit.ps1 -Target all
For AI agents
Read AGENTS.md first. It explains the public repository rules, installation method, validation requirements, and publishing checklist.
Meta-skill for contributing: skills/extract-to-toolkit/SKILL.md
Structure
yazinc-ai-toolkit/
├── AGENTS.md ← agents read this on open
├── WORKFLOW.md ← generalize, validate, commit, publish
├── INDEX.md ← canonical skill catalog
├── projects/REGISTRY.md ← public provenance policy
├── skills/ ← reusable SKILL.md folders
├── patterns/ ← detailed implementation guidance
├── templates/ ← safe reusable examples
└── scripts/ ← installation and validation helpers
Rules
- Never commit secrets, tokens, customer data, private paths, or unrelated repository history
- One skill = one focused job
- For every new skill, update
INDEX.md, projects/REGISTRY.md, and the landing-page catalog
GitHub
