skills.rest
thinking-machines-lab
thinking-machines-labOFFICIAL
Explore all skills by this author →
WebsiteGitHub

new-recipe

Scaffold a new training recipe directory with boilerplate in tinker_cookbook/recipes.

4.0k|507|Updated Jul 14, 2025
Search Tags:#cli#scaffold#recipes#training#tinker-cookbook
One-click install
npx skills add https://github.com/thinking-machines-lab/tinker-cookbook --skill new-recipe
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: new-recipe
Source: https://github.com/thinking-machines-lab/tinker-cookbook/tree/main/.claude/skills/new-recipe
Command: npx skills add https://github.com/thinking-machines-lab/tinker-cookbook --skill new-recipe

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Scaffold a new training recipe directory and boilerplate in tinker_cookbook/recipes, reducing setup time and ensuring consistency across projects.

Core Features & Use Cases

  • Scaffolds a complete recipe skeleton: Includes train.py and minimal structure following repo conventions.
  • Guides user through naming, type, and core patterns: Helps standardize recipe creation across teams.
  • Use Case: Rapidly prototype a new SL, RL, DPO, or distillation workflow with a consistent layout.

Quick Start

Run the scaffold command to create a new recipe directory under tinker_cookbook/recipes with the standard training script layout.

Frequently Asked Questions about new-recipe

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

FAQPage Schema
How do I scaffold a new training recipe in tinker_cookbook?▼

To scaffold a new training recipe, run the scaffold command to generate a recipe directory under tinker_cookbook/recipes. This creates a complete boilerplate structure including train.py, following repo conventions for fast and reliable experimentation.

What training workflows can I prototype using recipe scaffolding?▼

Recipe scaffolding supports prototyping SL, RL, DPO, and distillation workflows. The tool guides you through naming and selecting the training type, ensuring your recipe adopts the correct core patterns and consistent layout.

Does recipe scaffolding auto-generate log paths for training experiments?▼

Yes, recipe scaffolding auto-generates log paths for training experiments. This standardization ensures fast, reliable experimentation and a consistent layout across all newly created training recipes in the repository.

Can I use recipe scaffolding to standardize training recipe creation across teams?▼

Yes, you can use recipe scaffolding to standardize training recipe creation across teams. It guides users through naming, training type, and core pattern adoption to ensure CLI patterns and consistent layouts across projects.

What's included in the boilerplate when I create a new training recipe?▼

The boilerplate includes a minimal directory structure and train.py following repo conventions. This standard layout ensures your new training recipe is ready for immediate CLI pattern adoption and rapid experimentation.

Related Skills

Explore More Software Engineering Skills→
corezoidcorezoid
58

simulator-init

Initialize and configure the Simulator.Company environment with OAuth authentication and .env storage.

Official
Intermediate
kungfuaikungfuai
12

verify-training-pipeline

Validate CVLization training examples for structure, build, training, and metric logging.

Official
Intermediate
thinking-machines-labthinking-machines-lab
4.0k

hyperparams

Calculate learning rate, LoRA rank, batch size, and schedule for fine-tuning language models.

Official
Intermediate
khokskhoks
2

user_custom_strategy_intake

Scaffold custom trading strategy directories and register them in ClaudeTrading configuration.

Community
Intermediate