create-skills-workflow

Build multi-step AI workflows with state management using a micro-file architecture where each step is a self-contained unit. Persist data across workflow steps using frontmatter and explicit state variables to maintain context throughout complex phases.

Progressive step loading for prompt engineering is a modular architecture where each workflow phase loads independently as a self-contained unit. This approach structures complex skills into distinct phases with state persistence and interactive decision points.

Implement interactive decision points in a developer workflow using the AskUserQuestion tool, which supports automatic mode for seamless execution. This enables structured user confirmations and dynamic branching within multi-step state management workflows.

You need a micro-file architecture for skill creation when building complex multi-step workflows that require structured phases, state persistence, and interactive user decisions. It separates each step into a self-contained unit for modular development and progressive loading.

Yes, you can persist state variables across different workflow steps using frontmatter and explicit state variables. This state management approach maintains data continuity throughout structured phases and interactive decision points in multi-step workflows.

The micro-file architecture for workflow skills requires managing multiple self-contained units, increasing structural complexity. It is best suited for complex multi-step skills needing state persistence, rather than simple single-step tasks that do not require interactive decision points.