What problem does it solve? Creating a well-formed agent skill is error-prone: descriptions undertrigger, instructions are vague, and structural rules (kebab-case names, frontmatter format, no README) are easy to miss. This Skill walks you through a disciplined Discovery → Architecture → Craft → Validate → Deliver workflow so the resulting skill triggers correctly and behaves consistently. ## Core Features & Use Cases - Structured Discovery & Architecture: Defines use cases, trigger phrases, success criteria, and selects from five proven workflow patterns (sequential, multi-MCP, iterative refinement, context-aware selection, domain intelligence) before writing any code. - Craft Guidance with Anti-Patterns: Enforces frontmatter hard rules, progressive disclosure across SKILL.md and reference files, and provides good/bad description and instruction examples. - Automated Validation: Runs scripts/validate_skill.py to check folder naming, frontmatter fields, description triggers, and body quality, producing human or JSON reports. - Use Case: You say "turn this deployment checklist into a skill" — the Skill extracts the workflow from context, drafts a description with trigger phrases, writes the SKILL.md, and validates it before delivery. ## Quick Start Ask the agent to help you create a new skill for a workflow you want to automate, and answer its discovery questions.