Professional Creator With Beginner Guidance

Problem Addressed

Agent Skill packages often fail not because of missing files, but because triggers are vague, workflows are not executable, exception handling is absent, and refactoring can overwrite original assets. This skill is for engineers who need to turn natural-language requirements, SOPs, or existing SKILL.md drafts into valid, deliverable packages, with clarification before generating a template.

How It Works

The core behavior combines standard structure, execution contracts, and safety boundaries:

• Mode selection: distinguishes beginner onboarding, fast drafting, and existing package optimization.
• Input diagnosis: checks goal, trigger phrases, input/output types, core steps, and safety limits.
• Standard output: fills in SKILL.md, AGENTS.md, README.md, and package.json, normalizing YAML front matter and lower-kebab-case names.
• Automatic repair: fixes missing files, weak descriptions, and vague workflow steps while recording changes.
• Risk gating: pauses or refuses unsafe actions involving overwrites, privacy, credentials, or dangerous commands.

Boundaries

It is suitable for creating packages, fixing trigger, workflow, or error-handling issues, and turning processes into Skill assets. It does not decide business goals for the user; if requirements conflict, core goals are missing, or high-risk demands exist, it should list the conflicts and wait for a decision.