Setup Ml Project

Organization

probabl-ai

setup-ml-project

Ask only the pending setup pieces. When Git is selected, commit the current files before setup writes only if that check finds files to preserve. Trigger when the user asks to set up or bootstrap a workspace, or when a pipeline stage finds status.setup.pending non-empty. The question tells the user to select each piece they want. It starts by saying that before any modeling, the workspace should be set up. Do not say the order the pieces will run. Do not offer a piece that is already done or declined. Then, in that same opening phase, ask the environment manager, whether we manage the env, the import name, and git autocommit — only for pieces the user selects, and answers that are not already recorded. Persist those answers. An empty folder or a clean tree records skipped and does not ask. Then run only what the user selects. Git may ask once, just before the first commit, which unknown files to keep. Return to the skill that dispatched this turn.

Overview

Publisher	probabl-ai
Repository	`skore-skills`
Skill name	`setup-ml-project`
Stars	144
Forks	9
Bundled files	1
License	BSD-3-Clause
Links	Skill source Repository Homepage

Markdown instructions
A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.
Works with any LLM
AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.
1 bundled files
Scripts, templates, and references the model can read while it works. Files are read-only and never executed.
Open source
Published by probabl-ai on GitHub. Read the source before you install it.

Installation

Install the Setup Ml Project AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

Open the app and go to Plugins → Skills.
Choose "Install from GitHub".
Paste the skill folder URL below and confirm.
https://github.com/probabl-ai/skore-skills/tree/main/skills/setup-ml-project
Enable the skill in any chat where you want it available.

Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills

git clone --depth 1 https://github.com/probabl-ai/skore-skills.git /tmp/skore-skills
mkdir -p .claude/skills
cp -r /tmp/skore-skills/skills/setup-ml-project .claude/skills/setup-ml-project

Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Setup Ml Project in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Setup Ml Project on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Setup Ml Project is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

Loaded only when it is needed
The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.
Switch models mid-chat
Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Set Up ML Project

Ordering only. Never run a skipped skill's steps from memory.

Human-facing prose

Details: setup-workspace references/human_facing_prose.md. Questions use data-science labels only (Python environment, workspace layout, editable install, Git, environment manager, whether we manage the Python environment, Python import name, automatic commits). Do not put skill ids, G-* names, or the wrapper CLI in the question.

Lookup

Run status and env detect first. Take the first row that matches. Do not stop after reading the table.

`setup.pending`	Then
empty	step 9. Do not ask.
every pending piece's skill id is false	name the missing skills in one line, then step 9
non-empty	step 3 multi-select, then step 4, then the stop row, then the choice table

Stop row: editable is selected, has_src is false, and workspace is not selected → one-line stop. Do not ask the remaining choices. Do not scaffold from this skill.

Choice table. Apply every row that matches. Skip a row whose piece was not selected, was not on the board, or whose value is already recorded. Ask and persist before any write.

Selected piece	Recorded state	Action
Python environment	detected manager is not `"none"`, not ambiguous, not a mismatch, and `policy.env_manager` is unset	persist that manager. Do not ask.
Python environment	`env_manager` is `"none"` and `policy.env_manager` is unset, or `ambiguous`, or `mismatch`	ask, then `policy set env_manager`
Python environment	`policy.env.managed` is null	ask whether we manage the env (default yes), then persist
Workspace layout	fresh or manager-only, `policy.package` unset, `src/<pkg>/` does not name it	ask the import name (folder name is the default), then `policy set package`
Git	`policy.git.autocommit` is null	ask `on` or `off` once, then persist

Then step 7 when Git is selected, then step 8 in order env, workspace, editable, git, then step 9.

Sequence

Run python -m skore_skills status and python -m skore_skills env detect. status.skills is a per-id dict, never a boolean. Read status.setup. Each piece is done, declined, or missing. setup.pending is the ordered list of missing pieces. Do not env init, scaffold, install a package, or git init in this opening phase.
If setup.pending is empty, ask nothing. Go to step 9. done and declined pieces are already answered. Do not offer them. Do not ask their environment manager, whether we manage the env, import name, or automatic commits again.
AskUserQuestion with allow_multiple for setup.pending only. Start by saying that before any modeling, the workspace should be set up. Then the status facts each box rests on — detected manager, has_src, git presence, and which pieces are already recorded. Stay within 2–4 lines. Do not say the order the selected pieces will run. A file link is an addition, never the context. Include a box only when that id is true and the piece is missing. Option labels (user-visible, no ids):
- Python environment — setup.pending contains env and status.skills.setup-python-env is true
- Workspace layout — setup.pending contains workspace and status.skills.setup-workspace is true
- Editable install — status.skills.add-python-package is true, status.setup.editable is missing, and (has_src is true or workspace is on this board)
- Git — setup.pending contains git and status.skills.setup-git is true
Map selected labels to setup-python-env, setup-workspace, add-python-package, setup-git when loading.

The question's last line is exactly: Select each one you want. Write nothing after that line. Do not add a done or declined piece because the layout already looks unfinished. If every pending piece's skill id is false, skip the ask, name the missing skills in one line, and go to step 9.
After the answer, and not in the question, persist each piece the user does not select: python -m skore_skills policy set setup.<piece> declined (env, workspace, editable, or git). Do not persist declined for a box that was not on the board.
Editable selected, has_src false, and workspace not selected → one-line stop. Do not ask the remaining choices. Do not scaffold from this meta.
Ask the remaining choices now, before any write. Skip a question when the user does not select its piece, it was not on the board, or the value is already recorded. Each ask states in 2–4 lines what the answer authorizes. Order:
- Environment manager — the user selects Python environment. If env_manager is not "none", ambiguous is false, mismatch is false, and policy.env_manager is unset, persist that detected manager. Do not ask. Otherwise ask when (env_manager is "none" and policy.env_manager is unset, or ambiguous is true, or mismatch is true). Ask with AskUserQuestion, using recommended order. PATH is not permission. Do not curl | sh. Persist python -m skore_skills policy set env_manager <manager>.
- Whether we manage the env — the user selects Python environment and policy.env.managed is null. Default yes. Persist policy set env.managed true or false.
- Python import name — the user selects workspace, the layout is fresh or manager-only (no src/ and no journal/), policy.package is unset, and src/<pkg>/ does not already name it. Folder name is the default option. “You pick” / “go fast” does not resolve it. Persist policy set package <pkg>.
- Automatic commits — the user selects Git and policy.git.autocommit is null. Ask once: should later stages persist with git commit (on) or never (off)? That answer consents to the first commit. It does not authorize tracking unknown hidden files or review paths. Persist policy set git.autocommit on or off. That answer does not decide the pre-change snapshot.
Do not ask notebooks, the documentation site, tabular library, or where reports go.
Pre-change snapshot, before any setup write. Run this only when Git was selected, status.skills.setup-git is true, and policy.git.snapshot is null. Load setup-git and do only its pre-change snapshot section, then return here. Do not env init, scaffold, or git ignore-merge in this step. Do not ask the snapshot question in this skill. The loaded section asks only when git snapshot returns ask. clean and nothing_to_commit, including an empty folder, persist skipped and do not ask. If Git was not selected, git.snapshot is already set, or setup-git is not installed, skip this step. A missing setup-git is one line, then continue. When the check returns ask, the commit is still offered if autocommit is off.
Load selected skills only, in this order: env → workspace → editable (has_src) → git. Load <id> only if status.skills.<id> is true; else one-line skip. Do not invent that skill's steps. The loaded skills must not ask again for a choice persisted in step 6. After each loaded skill returns, persist python -m skore_skills policy set setup.<piece> done. For editable, do that only when env add --editable --execute exits 0. A non-zero exit leaves setup.editable missing. Do not mark a skipped piece, or a piece the user does not select, done.
Return to the skill that dispatched this turn (a lifecycle stage, or triage waiting on a lifecycle request). Do not open the entry board and do not start exploratory data analysis or a pipeline from here. When the user asked only to set up or bootstrap, status again and load triage-ml-task only if status.skills.triage-ml-task is true; else stop. That handoff may ask what to do next; it is not a setup question.

Stop conditions

Do not env init, scaffold, or install a package before step 8. Do not git init before step 7.
Do not offer a done or declined piece again.
Do not ask tabular library, skore mode, notebooks, or site.
Do not ask which hidden files or review paths to keep; setup-git asks that once, before the first commit.
Do not install sklearn, skrub, or pandas. The selected setup-python-env skill installs plain skore during bootstrap; do not install or configure Skore directly from this coordinator.
Do not write experiment or pipeline bodies.
Do not commit except by loading setup-git.
Do not abort setup because one skill is missing.
Do not invent a missing skill's procedure.
Do not pixi add / uv add from this coordinator; add-python-package owns install.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

evals/evals.json31 KB

Frequently asked questions

What does the Setup Ml Project AI skill do?

Why use Setup Ml Project on TypingMind?

Because you install it once and use it with any model. Setup Ml Project is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Setup Ml Project in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/probabl-ai/skore-skills/tree/main/skills/setup-ml-project. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Setup Ml Project?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Setup Ml Project?

As many as you like. As long as a model supports skills, you can use Setup Ml Project with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Setup Ml Project AI skill free?

Yes. It is published on GitHub by probabl-ai under the BSD-3-Clause license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Graincrawl

openclaw

OrganizationPopular

Granola archive: search, sync freshness, notes, transcripts, panels, SQL counts, and Graincrawl repo work.

Openclaw Changelog Update

openclaw

OrganizationPopular

Regenerate OpenClaw release changelog sections from git history before beta, stable, or extended-stable releases.

Openclaw Ci Limits

Manage OpenClaw GitHub Actions and Blacksmith CI capacity, runner-registration budgets, fanout caps, main-push single-flight, shard sizing, hosted-runner offload, queue health, and safe…

Openclaw Debugging

Debug OpenClaw model, provider, tool-surface, code-mode, streaming, and live/Crabbox behavior by choosing the right logs, probes, and proof path before changing code, including fetching stored…

Overview

Installation

Install in TypingMind

Install in another agent

Use it in TypingMind

Works with any AI model

Skill instructions

Set Up ML Project

Human-facing prose

Lookup

Sequence

Stop conditions

Bundled files

Frequently asked questions

Related AI Skills

Set up your own AI workspace now