atlas-setup
Skill0 starsChanged 4 months agoMITView on GitHub
Guided setup/reconfigure wizard for an Atlas-based project - walks the user through naming, canonical state, ownership, which CLI agents + agent files, the optional fleet layer, memory seeding, the first experiment, and git/remote/attribution, confirming before anything destructive or outward-facing. Use when the user says "/atlas-setup", "/onboard", "set up this project", "configure Atlas", or "change the project settings". Re-runnable (idempotent) to adjust settings later, not just first-time.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- Atlas-Authorization/atlas-claude-skillstheir own version, first seen Oct 15, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- Atlas-Authorization/atlas-claude-skills★ 0same 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 repository13 other files it gives its agents.
- AGENTS.mdAGENTS.md
- fleet/AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- .cursor/rules/atlas.mdcCursor rule
- add-repoSkill
- councilSkill
- mediateSkill
- onboardSkill
- optimizeSkill
- registry-syncSkill
- researchSkill
- snapshotSkill
- sync-repo-knowledgeSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 4 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
Start by telling the user what this will do and that nothing irreversible happens without a confirm.
- Very specific: line 21
Attribution: keep the upstream "distilled from ...
- Distinctive: line 27
Leave §3 withdrawn-claims register empty (it fills as work is retracted).
- Distinctive: line 35
Ask which CLI agents the team uses (Claude Code, Codex, Gemini, Cursor, Aider, ...).
- Very specific: line 49
Keep experiments/exp0_demo as a reference, or archive it.
- Distinctive: line 65
Keep responses tight; pipe any large output to a file and summarize.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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.
Files
/atlas-setup (alias: /onboard) - guided project setup & reconfigure
Walk the user through making this Atlas template their project - or changing settings later. Ask one focused question at a time (or small grouped sets), apply each change, confirm before anything destructive (deleting files, renaming) or outward-facing (creating/pushing a remote), and finish with a summary + a single commit. Re-runnable: if a setting is already configured, show its current value and offer to change it.
Start by telling the user what this will do and that nothing irreversible happens without a confirm. Then go section by section; skip any the user says is already fine.
1. Identity
- Project name → update
README.mdtitle + intro, and propose renaming the repo directory (confirm beforegit mv/mv). - Domain: research / development / both → adjust the README framing wording.
- Attribution: keep the upstream "distilled from ... Atlas / Curv Institute" credit (the MIT license asks reusers to retain it) and add the user's own org/author line.
2. Canonical state (CURRENT_STATE.md)
- Set the
As of:date. - Seed §1 live claims, §5 open questions with the user's real starting state; clear the example rows.
- Leave §3 withdrawn-claims register empty (it fills as work is retracted).
3. Ownership (OWNERSHIP.md)
- Fill the areas table: who/what owns which repo/experiment/surface (one owner per active area).
- Confirm the mediation channel for out-of-bounds requests (a decisions inbox dir, or the default
MEDIATION/REQ-*.mdnotes) and who the fallback owner is.
4. Agents & agent files
- Ask which CLI agents the team uses (Claude Code, Codex, Gemini, Cursor, Aider, ...).
AGENTS.mdis canonical for all; confirmCLAUDE.md/GEMINI.md/.cursor/rules/redirects are present for the ones in use (add any missing). Review.claude/settings.jsondefaults.
5. Fleet layer (optional, fleet/)
- Single machine? → offer to remove
fleet/(confirm) and drop its mention from README/AGENTS. - Multi-machine? →
cp fleet/fleet.example.yaml fleet/fleet.yamland help fill it. Remind: real IPs/MACs go only infleet.yaml(gitignored), never in the committed example.
6. Memory (memory/)
- Seed a real maintainer profile (
user) and any standing feedback (how they like to work). - Clear/replace the
memory/examples/*files; keepMEMORY.mdas the index (one line per memory).
7. First experiment
- Keep
experiments/exp0_demoas a reference, or archive it. - Scaffold the user's first experiment dir from the demo's shape:
PREREGISTRATION.md(freeze before running),pilot/run.py,gate.py,RESULTS_*.md. Don't run anything heavy without confirmation.
8. Git, remote, license
- Set repo-local
user.email/user.name. - Confirm
LICENSE+ attribution; review.gitignore. - Remote (outward-facing - confirm explicitly): offer to create a GitHub repo (
gh repo create) and push; ask public vs private-first and the org/account. Do NOT push without an explicit yes.
9. Confirm & commit
- Summarize every change made. Run the demo (
uv run ...) if experiments were touched, to confirm the repo still works. Commit with a clear message + attribution trailer. Leave any push to §8's confirm.
Rules
- One change at a time; show current value before changing it; never delete or push without a confirm.
- Keep responses tight; pipe any large output to a file and summarize.
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
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.

