personal-skill-authoringlisted

# Personal Skill Authoring Help the researcher preserve a repeatable way of working without requiring them to understand plugin structure, YAML, scripts, or terminal commands. ## User experience Work in ordinary chat. Ask only what is still unknown, one short question at a time. Offer concrete choices and always accept a free-form answer. The default route needs four decisions: 1. What repeated result should the skill produce? 2. What inputs does the researcher normally provide? 3. Which decisions must remain with the researcher? 4. What would prove that the result is good enough? Infer obvious details from the current conversation and files. Do not make the researcher repeat information already available. Before creating files, show a plain-language card with the proposed name, trigger, inputs, output, human confirmation points, and acceptance checks. Continue only after the researcher confirms or corrects that card. ## Authoring workflow ### 1. Bound one reusable job Create one skill for one recognizable job. Separate a broad request into multiple skills when it has independent triggers or outputs. Keep transient project facts, client data, credentials, and unpublished source material out of the reusable instructions. Write a discriminating description that says both what the skill does and when it should activate. Include a negative boundary when an adjacent skill could be confused with it. ### 2. Design the smallest useful package Start with `SKILL.md`. Add onl