agentleFS
Sign inStart free
JonathanBeck1JonathanBeck1/claude-studio-toolkit/skills/studio-onboard/SKILL.md

studio-onboard

Skill0 starsChanged 28 days agoMITView on GitHub

First-session onboarding ritual for a new collaborator on a studio repo that uses this toolkit. Walks through environment verification, context reading, and a structured briefing filled from the repo's own CLAUDE.md and README. Triggers on "I'm new", "onboard me", "where do I start", "first time here", "just joined", "getting set up", "new collaborator", "new engineer", any variation of "help me get oriented", "what is this repo", or signals that the user has never worked in this repo before. Do NOT use for returning collaborators, mid-session re-orientation, or quick "where is X" lookups — this is the full first-session ritual only.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. chuanglibs/Claude-Code-Game-Studios-Codex★ 1same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository3 other files it gives its agents.

The file

Show its rules and text

We split this skill into the rules it gives an agent

13 rules. About 0 are common, 8 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈8

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Greet: Read the Opening section of tour.md to the user, with the studio name filled from CLAUDE.md / README.md.

    line 16
  2. Distinctive:

    Run git rev-parse --show-toplevel — confirm we're at the repo root.

    line 19
  3. Very specific:

    Run node --version — compare against the engines field of the app's package.json.

    line 20
  4. Very specific:

    Check that node_modules/ exists where the app lives.

    line 21
  5. Distinctive:

    Anchor context: Read these in order:

    line 24
  6. Distinctive:

    Survey current direction: Read the "Current direction" section of CLAUDE.md if it has one — that is the live trajectory.

    line 29
  7. Very specific:

    Brief the human: Read sections 2–7 of tour.md to the user, in order, filling every bracketed prompt from what you just read.

    line 31
  8. Very specific:

    First-day prompt: Read section 8 of tour.md.

    line 33
  9. Distinctive:

    Do NOT install dependencies, start the dev server, or modify any files WITHOUT asking.

    line 39
  10. Distinctive:

    Do NOT commit, push, or branch during onboarding.

    line 40
  11. Distinctive:

    Do NOT shorten the ritual because the user seems experienced.

    line 41
  12. Distinctive:

    Do NOT inject personal opinions about the codebase or studio direction.

    line 42
  13. Distinctive:

    Do NOT fill a bracketed prompt from memory or guesswork.

    line 43

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

studio-onboard

Deterministic first-session ritual for a new contributor. Replaces ad-hoc orientation with a structured walk-through: environment check, context reading, briefing, first-day prompt. Everything Claude says comes from the repo's own documents — the tour is a template, not a script about any particular studio.

When to use

Invoke EXCLUSIVELY on the first session for a new contributor. For a returning person who wants a quick refresher, point them at /CLAUDE.md instead and do not run the full ritual.

The ritual

  1. Greet. Read the Opening section of tour.md to the user, with the studio name filled from CLAUDE.md / README.md.
  2. Verify environment. In order:
    • Run git rev-parse --show-toplevel — confirm we're at the repo root.
    • Run node --version — compare against the engines field of the app's package.json.
    • Check that node_modules/ exists where the app lives.
    • Most repo checklists make install + a dev-server smoke test the first thing a new collaborator does. If this one does, OFFER it: "Run npm install and start the dev server to confirm the app boots?" Run it only on an explicit yes, report the URL it serves, and stop the server once the check passes. Never run either unilaterally, and never substitute your own command for the one the checklist names.
  3. Anchor context. Read these in order:
    • /CLAUDE.md (auto-loaded but re-read to prime the conversation)
    • /README.md
    • The standards, taste, and brand documents those two files point at.
  4. Survey current direction. Read the "Current direction" section of CLAUDE.md if it has one — that is the live trajectory. Archived plans and retired docs are history only; never build from them.
  5. Brief the human. Read sections 2–7 of tour.md to the user, in order, filling every bracketed prompt from what you just read. Paraphrase naturally to match their apparent experience level. Do not skip sections. Do not inject personal opinions.
  6. First-day prompt. Read section 8 of tour.md. Wait for the user's answer. Route them to the matching skill / subagent / file using the tour's routing guide.
  7. Confirm the context actually loads. Most repo checklists close by having the new collaborator verify that CLAUDE.md is being auto-loaded — typically by asking the repo's own bar question ("what's the bar?" in the studio's vocabulary) and checking the answer comes back in the repo's own terms, unprompted. Run that check: put the question to the user to ask, and if the answer does not land in the repo's language, CLAUDE.md is not loading. Say so plainly and tell them to fix that before trusting any later session — do not paper over it by quoting the file you read manually in step 3.

What not to do

  • Do NOT install dependencies, start the dev server, or modify any files WITHOUT asking. Install and the dev-server smoke test are an offered step (step 2) when the repo's checklist calls for them — offer, wait for a yes, then run. Everything else stays hands-off.
  • Do NOT commit, push, or branch during onboarding. Apart from the offered install / dev-server check, the session changes nothing.
  • Do NOT shorten the ritual because the user seems experienced. The full ritual is what makes the first session deterministic.
  • Do NOT inject personal opinions about the codebase or studio direction. Stick to what the files say.
  • Do NOT fill a bracketed prompt from memory or guesswork. If the repo doesn't document it, say so.

After the ritual

Point the user at /CLAUDE.md (and its "Current direction" section, if any) as their ongoing reference. Make it clear: next time they sit down, no ritual — just work.

Files

  • SKILL.md — this file (the script)
  • tour.md — the briefing template Claude fills and reads during steps 1, 5, and 6
  • evals/triggers.json — should/shouldn't-trigger regression set for the description

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free