What problem does it solve?
Creating well-structured, outcome-driven AI skills and workflows is difficult. Most skills are either too prescriptive, wasting tokens on instructions the LLM already knows, or too vague, leaving the AI without enough context to make good judgments. This skill solves that by guiding you through a conversational discovery process to build lean, effective skills that trust the LLM's reasoning abilities while providing just enough structure to ensure quality outcomes.
Core Features & Use Cases
- Build New Skills: Guide users through a six-phase conversational discovery process to create new BMad-compliant workflows and skills from scratch, ensuring they are outcome-driven and properly structured.
- Convert Existing Skills: Take any existing skill—whether bloated, poorly structured, or non-conformant—and rebuild it into a lean, BMad-compliant equivalent with a before/after HTML comparison report.
- Quality Analysis: Run comprehensive quality checks on existing skills using deterministic lint scripts and LLM scanners to identify over-specification, structural issues, and enhancement opportunities.
- Use Case: A team lead can use this skill to convert a verbose, 2000-line team onboarding workflow into a lean 300-line skill that produces better results while cutting token costs by 70%.
Quick Start
Use the bmad-workflow-builder skill to create a new workflow for processing customer support tickets by describing the desired outcomes and letting the skill guide you through discovery and build.
