nemotron-add-pattern

Generate Nemotron step pattern files with frontmatter and validate via pytest.

1.8k|373|Updated Oct 3, 2025
One-click install
npx skills add https://github.com/NVIDIA-NeMo/Nemotron --skill nemotron-add-pattern
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: nemotron-add-pattern
Source: https://github.com/NVIDIA-NeMo/Nemotron/tree/main/skills/nemotron-add-pattern
Command: npx skills add https://github.com/NVIDIA-NeMo/Nemotron --skill nemotron-add-pattern

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Contributors need a reliable way to add new cross-cutting decision patterns for Nemotron steps without breaking the repo’s required frontmatter format, catalog regeneration, or pattern tests.

Core Features & Use Cases

  • Guided pattern authoring: Ensures new pattern files follow the required frontmatter fields, body sections, and naming conventions under src/nemotron/steps/patterns/.
  • Catalog and validation automation: Regenerates src/nemotron/steps/PATTERNS.md and runs uv run pytest tests/steps -q to prevent inconsistent wiring.
  • Scope and trigger calibration: Prompts for pattern scope (steps or global []), concrete triggers, and confidence selection (high, medium, experimental) before writing files.

Quick Start

Invoke the skill as /nemotron-add-pattern from the Nemotron repo root and answer the five orientation questions so it creates the correct src/nemotron/steps/patterns/{id}.md, regenerates PATTERNS.md, and passes uv run pytest tests/steps -q.

Frequently Asked Questions about nemotron-add-pattern

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

FAQPage Schema
How do I add a cross-cutting decision pattern for ML step planning?

To add a decision pattern, you create a structured markdown file under src/nemotron/steps/patterns with required frontmatter fields, regenerate the PATTERNS.md catalog, and validate the changes using pytest.

What is a cross-cutting decision pattern in an ML workflow?

A cross-cutting decision pattern encodes recurring ML decisions with concrete triggers, a defined step scope, and a calibrated confidence value for downstream step planning to reuse.

Do I need to run pytest after creating a new markdown pattern file?

Yes, you must run pytest validation after creating a new pattern file to prevent inconsistent wiring and ensure the new pattern passes required structural and catalog tests.

Can I set a global scope for a Nemotron step planning pattern?

Yes, you can set the pattern scope to apply to specific steps or leave it global by using an empty array, ensuring the pattern triggers apply correctly across your ML workflow.

What confidence levels can I assign when authoring a pattern?

You can assign high, medium, or experimental confidence levels when authoring a pattern, calibrating how downstream step planning processes the encoded ML decision triggers.

Why does adding a new decision pattern require regenerating PATTERNS.md?

Regenerating PATTERNS.md is required to update the pattern catalog with the newly added markdown file, preventing inconsistent wiring and ensuring downstream planning can discover the pattern.