WIT v0.1.0 — Initial Public Release

WIT v0.1.0 — Initial Public Release

WIT (Writing Is Thinking) is a human–LLM collaborative skill for
scientific thinking and writing.

Advance the research. Grow the researcher.

Core ideas

• Writing is thinking.
• WIT is a question generator, not a checklist completer.
• Finding → New Question → Answerability → Next Experiment / Interpretation.
• Competing hypotheses and falsification should guide scientific reasoning.
• Prioritize experiments by information gain.
• Match claim strength to evidence strength.
• Automate labor; augment judgment.

Skill structure

The installable Agent Skill is located in:

wit/

and contains:

• SKILL.md — executable collaboration protocol
• references/ — full WIT framework
• tests/ — external assessments of WIT
• case-studies/ — real-world applications of WIT

Included examples

• AlphaGo: external assessment / stress test of WIT
• MSFold: case study of applying WIT to real scientific research and writing

Status

WIT is an evolving framework. This release establishes the initial
public structure and core reasoning workflow. Further releases will
focus on cross-domain stress testing, human–LLM collaboration, and
researcher growth.