Professional Skill Creator Pro with Onboarding

Problem It Solves

A prompt can finish one task, but that is not enough to become a reusable, auditable Agent Skill. Skill Creator Optimized Pro targets that gap: it turns natural-language requirements, business workflows, rough drafts, or existing skill packages into a structured SKILL.md, optional references/ and scripts/, and release-ready materials. It keeps the trigger boundary, execution flow, and delivery contract checkable instead of leaving them as vague advice.

How It Works

The workflow starts with mode routing before generation or refactoring:
- Onboarding: for fuzzy requests, it runs a six-step intake, keeps each round short, and produces a minimal L1 or L2 package.
- Standard creation and optimization: it diagnoses requirements, designs structure, edits the skill body, validates the result, and packages the final deliverable. For existing packages, it reads current files first, lists defects, then changes and re-verifies.
- Quality gates: it checks YAML front matter, description boundaries, local references, name and version consistency, temporary files, and zip layout, with optional checks via validate-skill.py, score-skill.py, and pack-skill.py.
- Depth control: when PRDs, case libraries, or research notes are provided, it expects source mapping, playbooks, phrasebooks, or example assets, and separates structural compliance from professional completeness.

Boundary

Use it when converting SOPs, methodologies, prompt patterns, or team standards into distributable skills, or before listing review. It is not needed for one-off prompt rewrites, ordinary writing, or tasks without reuse intent. Unknown or high-risk skills should be audited first rather than generated into a runnable package.