by 237Zidane
Grand Tour Skill: orchestrated developer library for /grand-tour
# Add to your Claude Code skills
git clone https://github.com/237Zidane/grand-tour-skillGuides for using ai agents skills like grand-tour-skill.
See how grand-tour-skill compares with popular alternatives.
grand-tour-skill is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by 237Zidane. Grand Tour Skill: orchestrated developer library for /grand-tour. It has 86 GitHub stars.
grand-tour-skill's catalog security scan is still queued. You can run an instant dependency and prompt-injection check now with the "Scan for vulnerabilities" button above.
Clone the repository with "git clone https://github.com/237Zidane/grand-tour-skill" and add it to your Claude Code skills directory (see the Installation section above).
grand-tour-skill is primarily written in HTML. It is open-source under 237Zidane on GitHub, so you can review or fork the full source.
Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh grand-tour-skill against similar tools.
No comments yet. Be the first to share your thoughts!
Available after the catalog security scan passes.
⚠️ Third-Party Software Notice
This skill is third-party open-source software developed and hosted independently on GitHub. SkillsLLM is an informational directory and does not control or maintain the underlying repository.
Any security checks, ratings, or warnings displayed by SkillsLLM are automated and limited in scope. They do not constitute a security certification or guarantee that the software is safe, error-free, or free from malicious code, vulnerabilities, compromised dependencies, or prompt-injection risks.
Review the source code, permissions, dependencies, and configuration before installing or running any third-party skill. Use is at your own risk. To the maximum extent permitted by applicable law, SkillsLLM is not liable for losses arising from third-party software.
The deep catalog scan for this skill is still queued. Run an instant dependency check now instead.

Grand Tour Skill is a comprehension skill that conducts an and over a repository. It reads the full project, traces a real execution path through the actual code, and then requires you to explain the system in your own words until your answers hold up. Use it to close the gap between building a system and being able to describe how it works, or to onboard a new owner to a they did not write.
-grounded analysis: Parses the full repository — source tree, module boundaries, and git history — before generating any narrative. No hallucinated architecture; every claim traces to a file, symbol, or commit.
Single-metaphor modeling: Selects one structural metaphor that matches the repo's actual shape (event loop, pipeline, layered onion, state machine) and uses it consistently across every explanation. Metaphor is derived, not templated.
Live trace execution: Walks one representative code path end-to-end using toy inputs, naming each function, module, and data structure it touches as it goes. Produces a call sequence you can follow with your finger.
Pattern identification pass: Names the design patterns present in the code (observer, strategy, factory, adapter, etc.) so instinctive choices get their proper vocabulary — useful for interviews and design discussions.
Audience inference: Detects authorship from commit metadata. If the visitor didn't write the repo, the tour addresses them as an inheriting owner rather than a builder. No manual mode flag required.
Three output modes: live (default) emits a guided-tour page then runs a six-round chat scavenger hunt. print emits a 14–18 page workbook with exercises, writing space, and a fold-under answer key. grade ingests photographed workbook pages and produces a talking-points card built from your own passing sentences.
Retrieval-based quizzing: Most scavenger-hunt rounds are Explain how — listener, budget, answer in your own words. Scoring is split into accurate / fitted / confident, with a retry prompt after each round.
Self-contained output artifacts: Live-mode tour pages are single-file HTML with fonts embedded. Print-mode booklets render via headless Chromium. No network dependency at view time.
Security and bug surfacing: Issues found during analysis land in a dedicated late-section, phrased as things to change. Security findings and confirmed bugs are never trimmed for length.
Additive asset model: SKILL.md alone is sufficient to load the skill; missing assets/ and examples/ are fetched on demand, so a partial copy still functions.
Invoke the skill with /grand-tour (or just ask for a grand tour of the repo) and it runs a guided, then quizzes you on it in chat.
Live mode (default). Reads the whole repo, picks one metaphor that fits its shape, traces a single real execution path through the actual code using toy data, names the patterns you built by instinct, then opens a six-round scavenger hunt. Each round gives you a listener and a budget; you answer in your own words and get graded on accurate / fitted / confident before being asked to try again.
/grand-tour
Print mode. Renders a 14–18 page field guide and workbook: the guide, eight exercises with two-draft writing space and a printed self-check, plus an answer key that gives the ingredients of a good answer behind a fold-under stop page.
/grand-tour print
Grade mode. Photograph your filled workbook pages, attach them, and run:
/grand-tour grade
It reads your handwriting, grades your second drafts, checks the self-stamps, and builds a talking-points card from your own sentences wherever they passed.
Requirements. Live mode needs nothing extra: the tour page is a single HTML file with fonts embedded. Print mode renders the booklet with headless Chromium (Playwright works) and uses pdftotext and pdftoppm from Poppler to fill page references and check layout. Grade mode takes the attached photos.
Installation layout. The skill is the whole folder: SKILL.md plus assets/field-guide/ (style, stylesheet, fonts) and examples/ (reference pages both modes copy markup from). If only SKILL.md lands somewhere, it fetches the rest from the repo.
/grand-tour [mode]
| Mode | Default | Output |
|---|---|---|
live |
yes | Guided-tour page plus six-round scavenger hunt in chat |
print |
— | 14–18 page field guide and workbook |
grade |
— | Grade filled workbook pages and build a talking-points card |
live is used when no mode is supplied. grade requires attached photos of filled workbook pages.
liveEach scavenger hunt round presents a listener and a budget. Most rounds are Explain how. Responses are graded on three axes — accurate, fitted, confident — then the user is asked to retry.
Requirements: none beyond the skill files. The output page is a single HTML file with fonts embedded via assets/field-guide/components/fonts-inline.css.
printRenders a 14–18 page booklet containing:
The answer key gives the ingredients of a good answer, not a memorizable paragraph.
Requirements: headless Chromium (Playwright works) for rendering. pdftotext and pdftoppm (Poppler) for page-reference filling and layout.
gradeReads handwriting from photographed workbook pages, grades second drafts, verifies self-applied stamps, and builds a talking-points card using the user's own sentences wherever they passed.
Requirements: attached photos of filled workbook pages.
Tour pages and booklets copy their markup from the reference files in examples/. Style rules, component vocabulary, and print/screen mechanics live in assets/field-guide/. Fonts are declared through fonts.css (booklets) or fonts-inline.css (single-file tour pages); the shared stylesheet is bundle.css.
Give me a grand tour of this repo.
grand-tour print
grand-tour grade (with photos of your filled pages attached)
Every example below is a fragment you can paste into a prompt. Each one names a target and a mode; swap in your own repo path.
Give me a grand tour of./services/billing.
The skill reads the tree, picks a metaphor for the system's shape, then traces a single call through the real code with a small toy payload. The chat portion follows as six rounds. Rounds look like:
Round 3 — Explain how
Listener: a backend engineer who has never seen this service.
Budget: 60 seconds.
Question: walk me from an incoming webhook to the row that gets written.
You answer in prose. The skill grades each answer on accurate / fitted / confident, marks what slipped, and asks for a second attempt.
grand-tour print
Produces a 14–18 page field guide plus workbook: the, eight exercises with two-draft writing space and a self-check stamp on each page, and an answer key behind a fold-under stop page. The key lists the ingredients of a good answer, not a paragraph to copy.
Target a subdirectory when the repo is large:
grand-tour print./packages/runtime
grand-tour grade
Attach photos of your filled workbook pages. The skill reads the handwriting, grades your second drafts, verifies the stamps you gave yourself, and emits a talking-points card built from your sentences wherever they passed.
Narrow the tour to one subsystem or one commit range. Useful when you only need to speak to the part you just changed:
Give me a grand tour of the diff between HEAD~20 and HEAD.
Give me a grand tour of./src/queue, print mode, and treat./.tour/notes.md as prior context.
The skill inspects git history to decide whether to address you as the builder or the new owner. If you are onboarding to someone else's code, say so explicitly to skip the guess:
I didn't write this. Give me a grand tour of./ingest as the new maintainer.
Modes compose. Print first, grade later against the same repo:
grand-tour print./ap