Replace Claude Code's "thinking" filler with interesting, true facts shown on the "Tip:" line beneath the spinner — from subjects the user picks (history, space, biology, math, art, anything), in short or in-depth form. Use this whenever the user wants to customize, change, or personalize what shows while Claude is thinking/working, wants trivia or facts while they wait, wants to learn something during pauses, or says things like "change the thinking words," "show me facts while you think," "put facts in the spinner," or "I want history facts when you're working." It writes the facts into the spinnerTipsOverride setting and leaves the glowing spinner word at its defaults. This is a one-time SETUP task, not an ongoing behavior — run it, and the facts then show on their own in every new session.
Use whenever the user is hands-on building or changing software with you: writing or generating code, debugging an error or failure, refactoring, designing a data model, schema, API, or architecture, reviewing a diff or PR, or figuring out what a specific code error or message means. This is the DEFAULT mode for any such task, in any language, from a one-off script to a whole project — whether or not the user mentions wanting to learn or asks for an explanation. Rather than handing over finished code to copy-paste, explain the reasoning and tradeoffs while building and pause at genuine decision points to ask them a question, so their engineering judgment stays sharp. Do NOT use for: purely conceptual questions answered without touching their code (how an algorithm works, process vs thread), hardware/tool/product recommendations, or plain shell/git/devops operations. Also skip it when the user signals speed — "just give me the code," "that's all I need," "it's urgent," or prod is down.