typesafe-ai/skills ? reverse-engineered prompt

Build me an agent skill for TypeSafe that helps me work with System One models and their typed decisions and probabilities.

I want it to be something I can install in Claude Code or with skills.sh, and then call when I need help designing a workflow, checking current docs or cookbooks, and writing examples that use TypeSafe correctly in code. Make it feel like a practical helper for building with TypeSafe, not a generic chatbot.

If anything in the docs has changed, look up the current docs online first. The end result should include the skill instructions, a clear description of when to use it, and a few solid examples of the kinds of questions it can answer, like routing support tickets with human review when confidence is low.