# Ro Wo

> Test a material opinion, proposal, assumption, or decision-shaping hypothetical before agreeing or disagreeing. Use evidence, alternatives, and failure paths; exclude open-ended exploration, complete decision closure, and deep specialist analysis.

- Skill: `quantipixels/ro-wo` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add quantipixels/ro-wo`
- Raw SKILL.md: https://api.skillmd.com/api/skills/quantipixels/ro-wo/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: quantipixels (https://skillmd.com/u/quantipixels)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/quantipixels/ro-wo

---


# Rò Wò

Test the premise before endorsing or rejecting it. Distinguish the user's intent from the factual claim, its assumptions, and the consequences if it is true.

Seek the evidence that could change the conclusion: strongest support and counterevidence, a credible alternative, and relevant failure paths. Check changed scope, actors, dependencies, trust, or authority when they affect the claim. Separate observed facts from inference; do not reopen settled decisions without conflicting evidence or invent disagreement for balance.

Return `SUPPORTED`, `PARTIAL`, `UNSUPPORTED`, or `UNPROVED`, with the decisive reason and useful next action. Include the strongest alternative or counterevidence when material. If uncertain, name the controlling assumption and smallest evidence that would change the verdict; withhold judgment rather than guess.

Use `High | Medium | Low` confidence only when it helps communicate evidence strength. Confidence is neither desirability nor a measured probability. Keep the response proportional to the decision; the reasoning checks are not mandatory report headings.