wf-composer

What problem does it solve?

It turns vague natural-language workflow ideas into a structured, checkpointed, user-approved execution DAG that can be saved and reused.

Core Features & Use Cases

• Natural-language workflow parsing: Extract steps and input variables from an informal description.
• Executor resolution into a node graph: Map each step to concrete skill/CLI/agent/command executors using a node catalog.
• Automatic checkpoint injection: Insert checkpoint nodes at meaningful phase and artifact boundaries, with optional user pauses.
• Confirmation and editing loop: Visualize the pipeline, confirm with the user, and allow node add/remove/edit before persisting.
• Persist reusable templates: Save the final DAG as a reusable JSON template in a local template library.

Quick Start

Use the wf-composer skill with the prompt “Design a workflow to plan, implement, and test a small Rust feature, pause for review after planning, and save the reusable template.”