setup-workspace
Skill8 starsChanged 49 days agoMITView on GitHub
Onboards a new user. Interviews them about their profile and company, its legal setup, and their preferences, then fills the identity line in AGENTS.md and the forms in knowledge/counsel-brief.md and knowledge/preferences.md. Run at first launch; re-run any time to redo onboarding.
8 other projects use a version of this.
Adopted by
Most recently taken up by
- gfunc/skillstheir own version, first seen Nov 16, 2026
- probabl-ai/skillstheir own version, first seen Nov 9, 2026
- Richicinschi/studio-ostheir own version, first seen Nov 8, 2026
- Cassius2584/claude-uni-workspacetheir own version, first seen Oct 22, 2026
- The-AI-Thinker/claude-code-pm-startertheir own version, first seen Oct 15, 2026
- probabl-ai/skore-skillstheir own version, first seen Oct 15, 2026
Other versions8 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 8
- probabl-ai/skills★ 145same name
- probabl-ai/skore-skills★ 145same name
- The-AI-Thinker/claude-code-pm-starter★ 13same name
- Cassius2584/claude-uni-workspace★ 5same name
- timscheuerai/gtm-architecture-template★ 2same name
- gfunc/skills★ 1same name
- hunterbohm/solo-staff★ 0same name
- Richicinschi/studio-os★ 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.
The file
We split this skill into the rules it gives an agent
2 rules. 0 are common, 1 is distinctive, and 1 is very specific to this project.
0
Common
Most of the 8 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 16
If the user supplies anything, read it and draft what you can, then continue with the interview.
- Very specific: line 49
Don't create knowledge/<topic>.md files — those appear when work produces them.
“Common” is measured: kept by at least 6 in 10 of the 8 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Set up workspace
Fill the template files from an interview with the user. The user supplies material and answers your questions in plain language. You ask the questions, do the writing and fill the templates.
Start: ask for material
Your first message asks for a website link, a pitch deck, or other existing context
documents on the company, to be linked or dropped in desk/.
If the user supplies anything, read it and draft what you can, then continue with the interview. If the user doesn't supply anything or skips this step, continue with the interview.
How to run the interview
- Say what's coming — six rounds, one per heading of
knowledge/counsel-brief.md, then preferences — and announce each as you go: "Round 2 of 6 — key people". Open with: name, role, company, what their work responsibilities are. - The bracketed keywords under each heading note what belongs in that section — they are not a script. Compose plain, easy questions from them. Group what a person would answer in one breath. Skip anything an earlier answer or the material already covered.
- Make it easy and engaging to answer. Where an answer is a choice or a
multi-select, offer options with room for an answer in the user’s own words (in Claude Code:
AskUserQuestion) instead of asking an open-ended question, and include "none", "not sure", or "skip" where appropriate. Do not turn a skipped or uncertain answer into a statement that something does not exist. Otherwise use your judgement: the aim is a conversation, not a form. - The user may skip any question. Ask at most one follow-up per round. If something is still missing after that, leave it out of the file and move on. Not every detail needs to be filled.
- Short answers are enough; draft from them. After each round, show the draft and get a confirmation or a correction. Keep their phrasing over the template's.
What to fill
AGENTS.md— the one bracketed line at the top: who they are, their role, their company. One sentence, in their words.knowledge/counsel-brief.md— the six sections. Replace the keywords with the answers, written as prose in the user's own words, and leave out anything unanswered rather than marking it. Set the initial-fill date and each section's As of date to the setup date. Don't createknowledge/<topic>.mdfiles — those appear when work produces them.knowledge/preferences.md— fill the Language line. Then ask whether they already have working preferences, an AI style guide or a house style note; if so, adapt the defaults to it. Otherwise ask: "What do you find yourself fixing in every AI draft?" and add the answer in their words. The shipped defaults are usable as they stand.
Finish
List each file you changed, one line each, and say which questions are still unanswered.
Then explain briefly how the workspace works — this is the first time they'll see it:
- Anything they bring in is either matter work or a quick ask.
matters/holds work they'll come back to — one folder each, and itsbrief.mdsays where things stand, so the user can resume work without re-explaining the background.desk/is everything else: a dropped file, a one-off task.knowledge/is the workspace's durable memory. It grows out of the work: when something gets settled or turns out wrong,/file-itrecords it — dated, with what it rests on — so it doesn't need repeating.
Then propose one concrete first step — open the first real matter, whatever is actually on their plate today. Starting it is as simple as saying "new matter …".
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.

