Claude Code
https://code.claude.com/docs/en/setup
Terminal-based agent (CLI-first, with extensions); excels at deep reasoning, complex multi-file refactors, planning/executing tasks autonomously.
Claude Code dashboard
In Claude Code, esp skills, most users are experienced programmers. So they extended the framework in many ways, until now Claude Code integrates with many other tools.
Claude Code
Claude Skills
Claude Co-work
Opus
Sonnet
OpenAI
chatGPT Plus
Codex App
OpenAI → Codex (or Codex CLI / agent features in ChatGPT)
Agent-native platform; strong for rapid prototyping, end-to-end tasks. Integrated into tools like Cursor or standalone CLI variants.
https://m.youtube.com/watch?v=SS5DYx6mPw8
ai coding rules
https://docs.google.com/document/d/12ATcyjCEKh8T-MPDZ-VMiQ1XMa9FUvvk2QazrsKoiR8
1. Project Requirements Document (PRD), a "knowledge base" for AI coding models so AI models car refer to them.
2. App flow document, works like Google map for AI.
3. Tech stack document, lists all the technologies, dependencies, packages and APIs I use in the project.
4. Frontend guidelines document, provides complete styling guidelines that AI can follow when designing UI screens.
5. Backend structure document, the backbone of your app. It provides details about your database tables, storage bucket and user authentication. I use Supabase for database and storage and Clerkdev for Authentication.
6. Cursor rules or Windsurfrules, rules and regulations that AI model needs to follow to code your app. If you want AI to write commit message for GitHub or update any doc after each successful feature implementation.
7. Implementation plan from your knowledge base, provides complete 50-step plan to code the app end-to-end.
attach documents, or files, RAG, System and task instructions
MCP Servers
https://github.com/ahujasid/blender-mcp
claude blender mcp
https://github.com/orgs/freeCodeCamp/repositories
Not copy their course, but rather use GitHub repositories as source documents for AI-coding of same subject course materials.