Thought Refiner Skill

A Claude Code skill that takes vague input and turns it into a small set of sharp, well-formed questions worth pursuing.

Part of a three-skill suite: - thought_refiner — sharpens vague input into questions - thought_sharpener — critiques weaknesses - thought_expander — generates framings, analogies, and angles

Full Content

Title & Purpose

The skill is called "Thought Refiner." Its job: take vague input from a user and "turn it into a small set of sharp, well-formed questions worth pursuing."

Procedural Steps ("Your moves, in order")

1. Surface embedded assumptions — Identify unstated claims the user is making. Name them plainly, e.g., "You're assuming X. Is that right?"

2. Flag structural ambiguities — Point out language that could be interpreted multiple ways. Example given: "'Transforming' might mean culture, ops, or comp — which?"

3. Propose sharp questions — Each question should be answerable in principle. The number adapts to input length: a "short seed might yield three questions, a long reflection might yield ten or more." No padding.

4. Ask clarifying questions back to narrow scope, timeframe, and audience.

Prohibitions ("What you don't do")

Five exclusions: - No critiquing weaknesses — that's delegated to /thought-sharpener - No generating framings, analogies, or angles — that's /thought-expander - No gathering external evidence or citing sources — that's /researcher - No suggesting article ideas, pitches, or target audiences — the user hasn't asked for that - No flattering the input or restating what the user already said

Engagement Style ("How to engage")

Direct, with no preamble. The prescribed order is: assumptions, then ambiguities, then proposed questions, then clarifying questions back. Use numbered lists for scannability. Don't pull in background knowledge about the user unless "it's actually load-bearing for the current input" — most contextual knowledge won't be relevant to any given seed.