ZenML Agent Skills

Modular AI coding agent skills for ZenML workflows. Add the marketplace to your AI coding tool and install skills that guide you through implementing ZenML features.

Quick Start

Claude Code

# Add the ZenML skills marketplace
/plugin marketplace add zenml-io/skills

# Install a skill (e.g. quick-wins)
/plugin install zenml-quick-wins@zenml

# Use it — navigate to your ZenML project and run:
/zenml-quick-wins

Plugin docs · Marketplace docs

OpenAI Codex CLI

Ask the built-in skill installer to fetch skills from this repo:

$skill-installer install the zenml skills from github.com/zenml-io/skills

Codex skills docs

Cursor

  1. Open Cursor Settings (Cmd+Shift+J / Ctrl+Shift+J)
  2. Navigate to Rules → Project Rules → Add Rule
  3. Choose Remote Rule (GitHub) and enter https://github.com/zenml-io/skills

Cursor skills docs

Available Skills

Core Skills

Skill Description Install
zenml-quick-wins Analyze your setup, recommend high-impact improvements, and implement features like metadata logging, experiment tracking, alerts, schedules/triggers, streaming events, and model governance /plugin install zenml-quick-wins@zenml
zenml-scoping Scope and decompose ML workflow ideas into realistic ZenML architectures, including pipelines, deployments, resources, and streaming boundaries /plugin install zenml-scoping@zenml
zenml-pipeline-authoring Author ZenML pipelines with steps, artifacts, Docker settings, materializers, metadata, secrets, deployments, streaming events, YAML config, and visualizations /plugin install zenml-pipeline-authoring@zenml

Migration Skills

These skills are best used when you already have real source material from another workflow system — code, YAML, JSON job specs, pipeline definitions, or project config — and you want help translating it into idiomatic ZenML rather than doing a superficial syntax swap.

A few important caveats:

  • They classify source concepts as direct, approximate, or unsupported rather than pretending every platform feature has a safe 1:1 ZenML equivalent.
  • They are most helpful when you can provide the actual workflow code/config, not just a vague description.
  • Some migrations will be partial by design: the skill may generate working ZenML code for the cleanly mappable parts and a migration report for the pieces that need redesign.
  • After a migration, the best follow-up is usually zenml-quick-wins for production improvements and zenml-pipeline-authoring for deeper customization.
Skill Description Install
zenml-airflow-migration Migrate Apache Airflow DAGs to idiomatic ZenML pipelines with concept mapping, code translation, severity-classified flagging, and redesign suggestions /plugin install zenml-airflow-migration@zenml
zenml-argo-migration Migrate Argo Workflows to idiomatic ZenML pipelines with concept mapping, code translation, Kubernetes-native pattern flagging, and redesign guidance /plugin install zenml-argo-migration@zenml
zenml-databricks-migration Migrate Databricks Workflows (Lakeflow Jobs) to idiomatic ZenML pipelines with notebook refactoring, concept mapping, code translation, and unsupported pattern flagging /plugin install zenml-databricks-migration@zenml
zenml-prefect-migration Migrate Prefect flows and workflows to idiomatic ZenML pipelines with concept mapping, dynamic-execution analysis, code translation, and unsupported pattern flagging /plugin install zenml-prefect-migration@zenml
zenml-vertexai-migration Migrate Vertex AI Pipelines (KFP v2 / PipelineJob workflows) to idiomatic ZenML pipelines with concept mapping, artifact-contract translation, GCPC rewrite guidance, and unsupported pattern flagging /plugin install zenml-vertexai-migration@zenml
zenml-azureml-migration Migrate Azure Machine Learning SDK v2 pipelines to idiomatic ZenML pipelines with concept mapping, Azure-aware compute translation, code patterns, and unsupported pattern flagging /plugin install zenml-azureml-migration@zenml
zenml-dagster-migration Migrate Dagster assets, ops, jobs, and asset-graph workflows to idiomatic ZenML pipelines with concept mapping, pipeline-boundary planning, code translation, and unsupported pattern flagging /plugin install zenml-dagster-migration@zenml
zenml-flyte-migration Migrate Flyte workflows, tasks, and LaunchPlans to idiomatic ZenML pipelines with concept mapping, special-type planning, code translation, and unsupported pattern flagging /plugin install zenml-flyte-migration@zenml
zenml-kedro-migration Migrate Kedro pipelines and projects to idiomatic ZenML pipelines with Data Catalog to artifact translation, concept mapping, code translation, and unsupported pattern flagging /plugin install zenml-kedro-migration@zenml
zenml-metaflow-migration Migrate Metaflow flows to idiomatic ZenML pipelines with FlowSpec mapping, artifact translation, control-flow redesign notes, and unsupported pattern flagging /plugin install zenml-metaflow-migration@zenml
zenml-sagemaker-migration Migrate Amazon SageMaker Pipelines to idiomatic ZenML pipelines with concept mapping, artifact-model rewrites, runtime-setting translation, and unsupported pattern flagging /plugin install zenml-sagemaker-migration@zenml

Coming Soon

  • Debugging — Investigate pipeline failures and performance issues

Combine with MCP Servers

For the best AI-assisted ZenML experience, pair skills with MCP servers:

# Add ZenML docs MCP (Claude Code)
claude mcp add zenmldocs --transport http https://docs.zenml.io/~gitbook/mcp

This gives your AI assistant both structured workflows (skills) and doc-grounded answers (MCP).

Learn More