skills.rest
Garsdal
GarsdalCOMMUNITY
Explore all skills by this author →
GitHub

dojo-onboard

Generate PROGRAM.md and SETUP.md and verified load_data.py and evaluate.py for Python ML projects.

3|1|Updated Mar 10, 2026
Search Tags:#code scaffolding#regression#dojo#ml research#experiment automation#tool verification#evaluation contract
One-click install
npx skills add https://github.com/Garsdal/Dojo --skill dojo-onboard
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: dojo-onboard
Source: https://github.com/Garsdal/Dojo/tree/main/.claude/skills/dojo-onboard
Command: npx skills add https://github.com/Garsdal/Dojo --skill dojo-onboard

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It streamlines onboarding an existing Python ML codebase into Dojo by turning your current data loading and evaluation logic into a frozen, verifiable research contract.

Core Features & Use Cases

  • Conversation-based scaffolding: Reads your repository to tailor setup and asks only targeted questions needed to write PROGRAM.md and SETUP.md.
  • Verified connector generation: Runs dojo domain setup to generate load_data.py and evaluate.py and verifies them against your actual data before freezing.
  • Safety through the freeze gate: Iterates on SETUP.md when verification fails, without bypassing contract verification.
  • Use case: Set up Dojo for a real regression pipeline where you already have data loaders and a metric, so the agent can run trustworthy experiments inside your evaluation contract.

Quick Start

Ask the AI to onboard your repo by running /dojo-onboard in Claude Code from your project directory.

Frequently Asked Questions about dojo-onboard

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I onboard an existing Python ML project into Dojo for experiment automation?▼

Onboarding a Python ML project into Dojo involves generating PROGRAM.md and SETUP.md, then producing verified, frozen load_data.py and evaluate.py tools to enforce a fixed evaluation contract for experiment automation.

What is a frozen evaluation contract in ML research and when do I need it?▼

A frozen evaluation contract in ML research locks your data loading and evaluation logic into verified tools, ensuring trustworthy experiment automation when you already have custom data loaders and metrics.

What happens if tool verification fails during Dojo domain setup?▼

If tool verification fails during Dojo domain setup, the freeze gate iterates on SETUP.md without bypassing contract verification, ensuring load_data.py and evaluate.py are validated against actual data.

How many questions does Dojo onboarding ask before scaffolding the .dojo configuration?▼

Dojo onboarding detects your current repository state and asks only 2 to 4 targeted questions needed to write PROGRAM.md and SETUP.md before non-interactively scaffolding the .dojo configuration.

Related Skills

Explore More Education & Research Skills→
shariqazeemshariqazeem
1

dojo-system

Generate Cairo Dojo contract systems for game logic and on-chain model state.

Community
Advanced
ShiienShiien
7

experiment-dse

Automate hyperparameter design-space exploration and generate DSE reports.

Community
Advanced
dojoenginedojoengine
37

dojo-wallets

Provision burner wallets and connectors for Dojo.js games.

Official
Intermediate
djscrew1738djscrew1738

npm

Pipe confirmation responses to npm package installation prompts.

Community
Basic