Cole Medin - AI-Coding Knowledge Bundle

June 25, 2026 · View on GitHub

A portable knowledge base of my best AI-coding videos - Claude Code, context engineering, the PIV loop, the AI layer, and the new Archon harness builder - in the Open Knowledge Format (OKF): plain markdown any AI agent can read.

How to use it

Paste this to your AI coding assistant (Claude Code, Cursor, Codex, Gemini CLI, …):

Here's a knowledge bundle: https://github.com/coleam00/cole-medin-ai-coding
Read its README and set it up so I can search over it. Then tell me what's inside.

That's it - your agent reads the rest of this README, clones the bundle, and you can start asking questions about anything I teach.


Agent setup instructions

You're being asked to make this knowledge bundle available so the user can ask questions about Cole Medin's AI-coding content. Follow these steps.

1. What this is

This repo is an Open Knowledge Format (OKF) bundle - a folder of markdown files, each with a small YAML frontmatter block, navigated via index.md. You read it directly: no database, no embeddings, no API, no special tooling. It also ships a tiny, dependency-free CLI - okf-cli.py (Python standard library only) - for fast navigation and search.

2. Set it up

Clone the repo into the current working directory:

git clone https://github.com/coleam00/cole-medin-ai-coding
cd cole-medin-ai-coding
python okf-cli.py index      # confirm it works - prints the table of contents

3. Use the CLI

  • python okf-cli.py index [subpath] - print an index (start at the root; e.g. python okf-cli.py index videos)
  • python okf-cli.py find "<query>" - ranked keyword search across the bundle
  • python okf-cli.py read <path> - print a page, e.g. python okf-cli.py read concepts/the-piv-loop

(You can also just open the markdown files directly - okf-cli.py is only a convenience.)

4. How to answer the user's questions

  1. python okf-cli.py find "<the user's topic>" (or read index.md) to locate the relevant pages.
  2. python okf-cli.py read the specific videos/<slug> or concepts/<slug> pages - only what's relevant, not the whole bundle (progressive disclosure).
  3. Follow the relative markdown links between a video and the concepts it teaches.
  4. Answer grounded in those pages, and cite the source video by its title and resource (YouTube URL) from the frontmatter.

This is read-only reference knowledge - don't modify the bundle.

What's inside

  • index.md - the table of contents (start here)
  • videos/ - one page per video (summary, key ideas, tools, link)
  • concepts/ - the cross-cutting themes that tie the videos together
  • okf-cli.py - the dependency-free navigation/search CLI
  • log.md - change history

Channel: https://www.youtube.com/@ColeMedin