closestfriend

Grand Tour — AI skill for Claude Code

AI community

A Claude skill that tours a codebase you built with AI until you can explain how it works: traces one real run, names your patterns, makes you say it back.

How to install Grand Tour

This entry records only its repository, not the path inside it, so there is no exact command to give. Open closestfriend/grand-tour and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Grand Tour does

A Claude skill that tours a codebase you built with AI until you can explain how it works: traces one real run, names your patterns, makes you say it back. Live tour, print workbook, or grade mode. Companion to explain-diff.

Alternatives in AI

  • Omniget — Udemy & Hotmart course downloader, YouTube downloader (yt-dlp GUI, 1,800+ sites) + desktop app for AI agents 13.9k ★
  • Ralph For Claude Code — by Frank Bria - An autonomous AI development framework that enables Claude Code to work iteratively on project 8k ★
  • Deepreasoning — A high-performance LLM inference API and Chat UI that integrates DeepSeek R1's CoT reasoning traces with Anthr 5.4k ★

README

The Grand Tour

A Claude skill that walks you through a codebase you built, usually with an AI, until you can **explain how it works**: to a user, a collaborator, an interviewer, or an engineer pairing with you.

![Six pages from an example field guide: the cover, the map, the traced run, the exercises, and the answer key](examples/preview.png)

Why

Building with an AI often leaves you with a system you understand from the inside and still can't talk about. You know what it does, you had the ideas, you can extend it. Then someone asks "so how does it work?" and the answer comes out vague, or it comes out as *why* you built it rather than *what it does*.

The Grand Tour closes that gap. It reads the whole repo, picks one metaphor that fits its shape, traces one real run through the actual code with toy data, gives you the proper names for patterns you built by instinct, and then makes you say it back, in your own words, until it holds up.

It works just as well for onboarding someone to a team's codebase: it reads the git history, notices the visitor didn't write it, and addresses them as the new owner instead of the builder. Nobody has to say which they are.

It's written to you, as the person who built (or now owns) the thing. It isn't a code review. Problems get their own section near the end, phrased as things you'll want to change, and every real one is there: security findings and confirmed bugs are never cut to save space.

Three modes

Mode What you get
live (default) A long guided-tour page, then a six-round scavenger hunt in chat. Most rounds are Explain how: a listener, a budget, and you answer in your own words. You get graded on accurate / fitted / confident, then asked to try again.
print A 14–18 page field guide and workbook to take somewhere with a pen. The guide, eight exercises with two-draft writing space and a printed self-check, and an answer key behind a fold-under stop page that gives the *ingredie