Generate Argo Workflows with YAML

This skill provides expert Argo Workflows generation - the container-native workflow engine for orchestrating parallel tasks in Kubernetes - covering YAML specifications, template types, dependencies, artifacts, and resource management. It structures every workflow around the core apiVersion/kind/metadata/spec pattern with an entrypoint and parameters, then covers three template types: container templates for single containerized tasks, script templates for multi-step Python logic embedded directly in the workflow, and DAG templates for expressing task dependencies and parallel execution explicitly (test then build, then security-scan and deploy in parallel where dependencies allow). It covers artifact management for passing files between steps via S3-compatible storage (including MinIO endpoint configuration), conditional execution using step-output-based when clauses, loops and parallelism via withItems and a parallelism cap, and resource management with CPU/memory requests and limits, node selectors, and tolerations.

Use this skill when designing or reviewing Argo Workflow YAML for Kubernetes-native CI/CD or data pipelines that need explicit dependency graphs, parallel task execution, or artifact passing between steps - the DAG template pattern (test, then build, then security-scan and deploy in parallel) is built for exactly this kind of multi-stage pipeline. It also covers production concerns directly: Kubernetes secrets via secretKeyRef instead of hardcoded credentials, least-privilege RBAC through serviceAccountName, retry strategies with exponential backoff, exit handlers for cleanup, and Prometheus metrics for workflow duration. It is not a general Kubernetes deployment guide - it's scoped to Argo's own workflow, template, and artifact model rather than broader cluster configuration.

- name: ci-pipeline
  dag:
    tasks:
    - name: test
      template: run-tests
    - name: build
      template: build-image
      dependencies: [test]
    - name: security-scan
      template: security-scan
      dependencies: [build]
    - name: deploy
      template: deploy-app
      dependencies: [build, security-scan]

Given a pipeline requirement, it produces complete Argo Workflow YAML: the top-level spec with entrypoint and arguments, container or script templates for the actual work, a DAG or steps block encoding dependencies like the one above, artifact input/output blocks wired to S3 or MinIO storage, retryStrategy blocks with backoff duration/factor/maxDuration, resource requests/limits and node placement rules, and metadata labels/annotations plus Prometheus metrics blocks for observability.

Platform and DevOps engineers building Kubernetes-native CI/CD or data pipelines on Argo Workflows who need production-ready YAML covering not just the happy path but secrets handling, retries, resource limits, and monitoring from the start. It suits teams that want reusable workflow templates across the organization and that validate every workflow with argo lint before submission, per the skill's own closing guidance.