Yunomi banner
kazuph kazuph

Yunomi

AI community

Description

Human-in-the-loop review interface for AI coding workflows

Installation

This entry records only its repository, not the path inside it, so there is no exact command to give. Open the source below and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

README

どうぞ 🍵

a voxel robot quietly serving tea while an engineer reads a report

Reviews, served like tea.
yunomi(湯のみ)— a human-in-the-loop approval gate for AI coding workflows

npm version license 日本語

yunomi demo: the agent serves its report, you scroll the evidence, leave a comment, and approve


**Installation: none.** Just tell your AI — *"run `npx yunomi`"* — and it will know what to do.

A **yunomi**(湯のみ)is the everyday Japanese teacup: no handle, no saucer, warm in both hands. Tea served in one always arrives the same way — set down gently in front of you, with a slight bow and a single word: *douzo* — "here you go."

The story

yunomi was not built so humans can demand *"show me what you did."* It is the opposite gesture.

When an AI agent finishes a piece of work, it shouldn't just say "done." It should brew a report — what changed, why, the evidence, the test results — pour it into a cup, and set it down in front of you:

npx yunomi REPORT.md

*Douzo.* 🍵

Your browser opens. You read at your own pace. You leave comments; you approve — or you hand the cup back. The agent waits the way good service waits: quietly, until you decide. When you submit, yunomi exits and hands your verdict back to the agent as structured YAML, and the loop continues until you say it's good.

In the vibe coding era, humans no longer read every