HN Showcase
Back to browse
GitHub Repository

Self-guided workshop to build your own CLI coding agent like Claude Code from scratch in Python.

14 starsPython

Build Your Own CLI Coding Agent in Python

by primaprashant·Feb 24, 2026·3 points·0 comments
Visit ProjectView on HN

AI Analysis

●●SolidShip ItCozySolve My Problem

Hands-on agent-building workshop converted to self-paced—learn by coding, not reading.

Strengths
  • •Excellent pedagogical design: skeleton repo + phased hints + reference implementations lets learners struggle productively
  • •Deliberately disables AI assistance to force deep understanding—refreshingly anti-AI in an AI-coded era
  • •Concrete deliverable: working CLI agent that participants can use and extend on real codebases
Weaknesses
  • •Educational content, not a product—similar workshops and agent tutorials are abundant (Anthropic docs, various GitHub repos)
  • •3–5 hour time commitment is significant; completion rate likely low compared to copy-paste quickstarts
Category
Education
Target Audience

Python developers wanting to understand how AI coding agents work internally; engineers learning LLM app architecture

Post Description

Ran a hands-on workshop in Tokyo where ~50 engineers built a CLI coding agent from scratch in Python and now, I've converted it into a self-paced exercise.

GitHub Repo: https://github.com/primaprashant/alduin

This should help in getting a better understanding of what goes on inside the coding agents. You will iteratively implement the core agent loop present in all coding agents like Claude Code and Codex starting from a skeleton repo with a basic input loop and no LLM. Over 7 phases, you will add the Anthropic API and implement tools (read file, edit file, bash). Each phase has hints and a reference implementation if you get stuck.

Should take ~3-5 hours and in the end, you'll have your own coding agent which you can use on other project and codebases as well.

Feedback and PRs welcome. Happy to answer any questions.

Similar Projects