learn-claude-code and build-your-claude-code-from-scratch
About learn-claude-code
shareAI-lab/learn-claude-code
Bash is all you need - A nano Claude Code–like agent, built from 0 to 1
Implements a minimal Claude-powered coding agent through direct model integration rather than prompt orchestration, emphasizing tool access (shell, file I/O, API calls) and dynamic knowledge retrieval as the core "harness" architecture. The project deliberately rejects framework-heavy approaches, instead teaching how to build lightweight execution environments that let LLMs autonomously reason and act—applicable across any domain requiring perception, reasoning, and action loops. Core technical patterns cover tool composition, context management, and subagent isolation to keep the model's decision-making clean and scalable.
About build-your-claude-code-from-scratch
woodx9/build-your-claude-code-from-scratch
Build a Claude Code–like CLI coding agent from scratch.
Implements a progressive multi-chapter learning framework for agentic AI systems, from basic tool calling (native function calls and XML formats) through advanced patterns like ReAct loops, real-time streaming with character-by-character output, intelligent conversation history compression with token tracking, and sub-agent delegation with async execution. Architecture centers on the "Quick Star" CLI agent that integrates with Claude/OpenAI-compatible APIs via OpenRouter, including native MCP (Model Context Protocol) server support for dynamic external tool loading, while demonstrating practical code generation workflows through interactive debugging and file management.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work