Skip to content

Skills

App Framework Studio ships a set of AI agent skills that can be installed into any AI coding assistant (GitHub Copilot, Gemini, Cursor, and others) that supports the ai-agent-skills standard.

Which skill pack?

DataRobot's official skill pack for coding assistants is datarobot-agent-skills. Its datarobot-agent-assist skill builds a fully integrated agentic application — MCP server, agent, backend API, and frontend, all in the box together and tested to work as a whole. It also covers agent observability, the Workload API, and application CI/CD, alongside a smaller set of skills for the platform's AI/ML APIs.

npx ai-agent-skills install datarobot-oss/datarobot-agent-skills

The skills in this repository are for when you want just one piece, or a customized combination you assemble yourself: just an agent, just a FastAPI app, a non-default recipe, or a component wired up by hand outside what the integrated bundle scaffolds. They also cover authoring and documenting af-component-* components.

You want to… Skill pack Install
Build a fully integrated agentic application (MCP + agent + API + frontend, bundled and tested together), instrument it, or set up its CI/CD datarobot-agent-skills npx ai-agent-skills install datarobot-oss/datarobot-agent-skills
Build just one piece, a customized recipe, or author/document a component App Framework Studio (this repository) npx ai-agent-skills install datarobot-community/app-framework

Both packs can be installed in the same assistant, and the datarobot-app-framework skill includes routing guidance for requests that belong in the official pack instead.

Installing skills

Install all skills from this repository using npx ai-agent-skills:

npx ai-agent-skills install datarobot-community/app-framework

This registers the following skills in the AI assistant:

Skill Description
datarobot-app-framework Build and deploy applications on DataRobot using the App Framework component system — FastAPI applications, LLM integrations, and agentic workflows.
datarobot-app-framework-doc-update Generate and intelligently merge README documentation for a component using its copier-module.yaml schema and the Documentation Style Specification.

Available skills

datarobot-app-framework

Guides an AI assistant through scaffolding, configuring, and deploying App Framework recipes using af-component-* building blocks. Covers agents (CrewAI, LangGraph, LlamaIndex), FastAPI backends, React frontends, and LLM integrations.

Trigger the skill by asking the assistant to:

  • Build or deploy an application on DataRobot using the App Framework.
  • Add a component (dr component add, uvx copier copy).
  • Wire components together or configure an agent.

datarobot-app-framework-doc-update

Generates a structured README.generated.md scaffold from a component's copier-module.yaml, then intelligently merges it with the existing README.md — preserving human-written content, applying the Documentation Style Specification, and using the template's authoring hints to fill any gaps.

Trigger the skill by asking the assistant to:

  • Update or regenerate the README for a component repository.
  • Generate documentation from copier-module.yaml.

Running locally before pushing to GitHub

# Step 1. Generate the scaffold from the app-framework-studio root.
cd ~/code/app-framework-studio
uv run --project tools/af_component_doc_update af-component-doc-update ~/code/YOUR_COMPONENT_DIR

# Step 2. Invoke the skill in Copilot Chat.

#skills/datarobot-app-framework-doc-update/SKILL.md Update the README for ~/code/YOUR_COMPONENT_DIR. For Step 1, use the local variant: cd ~/code/app-framework-studio && uv run --project tools/af_component_doc_update af-component-doc-update ~/code/YOUR_COMPONENT_DIR