Domino Skills extend coding assistants with platform-specific actions. Inside a Domino workspace, your agent automatically gains access to specific skills, including running jobs, registering models, tracking experiments, deploying apps, and more.
Domino Skills are currently optimized for Claude Code. They might not work out of the box with other coding agents.
How skills work
Skills load automatically when a coding assistant starts inside a Domino workspace. The agent detects the Domino project and loads available skills from the domino-claude-plugin.
You don’t need to install or configure skills manually. To use a skill, reference the skill by name when prompting your agent. For example:
Use the modeling assistant skill. Analyze the dataset in this project
and train a few models to predict diabetes.
Available skills
The following skills are included in the Domino Standard Environment:
Development & modeling
| Skill | Description |
|---|
| modeling-assistant | AI-assisted model development with MCP servers |
| experiment-tracking | Track ML experiments using MLflow-based Experiment Manager |
| domino-experiment-setup | Set up MLflow experiment tracking for traditional ML |
| environments | Create and customize Domino Compute Environments |
| workspaces | Manage interactive development environments |
| projects | Work with Domino Projects, Git integration, and collaboration |
Jobs & orchestration
| Skill | Description |
|---|
| jobs | Create, run, and manage Domino Jobs |
| flows | Orchestrate multi-step ML workflows using Domino Flows (Flyte) |
| launchers | Create parameterized web forms for self-service job execution |
| distributed-computing | Work with Spark, Ray, and Dask clusters |
Data
| Skill | Description |
|---|
| datasets | Work with Domino Datasets for versioned storage |
| data-connectivity | Connect to external data sources such as S3, Azure, etc. |
| domino-data-sdk | Use domino-data SDK for data access and Feature Store |
Deployment & monitoring
| Skill | Description |
|---|
| model-endpoints | Deploy and monitor model API endpoints |
| model-monitoring | Monitor deployed models with drift detection and alerting |
| app-deployment | Deploy web applications to Domino |
GenAI & LLM
| Skill | Description |
|---|
| ai-gateway | Access external LLM providers through Domino AI Gateway |
| genai-tracing | Trace and evaluate GenAI applications and agents |
| domino-trace-setup | Set up GenAI tracing for agents and LLM applications |
Applications & SDK
| Skill | Description |
|---|
| domino-app-init | Initialize new Domino-ready web apps such as Vite+React, Streamlit, Dash, and Flask |
| domino-ui-design | Build Domino-styled web apps matching the Domino Design System |
| python-sdk | Programmatically interact with Domino using python-domino SDK |
| domino-debug-proxy | Debug Domino proxy and routing issues for web applications |
Use skills in practice
When you ask your coding agent to perform a Domino-specific task, the agent selects and activates the appropriate skill. For example:
-
“Train a model on the diabetes dataset and track the experiment” activates the modeling-assistant and experiment-tracking skills.
-
“Deploy this model as an API endpoint” activates the model-endpoints skill.
-
“Create a Streamlit app for this project” activates the domino-app-init skill.
-
“Set up a Spark cluster and run this distributed job” activates the distributed-computing skill.
Modeling assistant walkthrough has a complete example of the modeling-assistant skill in action.
Next steps
