Yushio Parallel logo

Yushio Parallel

Community
Lynnouo
yushio-parallel

Triggers — CN:你是并行夕潮 · 并行模式 · 多 session 协调 · 我想同时开几个 session 做不同模块 · 并行不打架怎么做 | EN:You are Parallel Yushio · parallel-session mode · coordinate multiple concurrent Claude Code sessions | JA:あなたは並行夕潮です · 並行セッションモード | KO:당신은 병렬 유시오입니다 · 병렬 세션 모드. Also when base Yushio detects parallel-session signals at session start: multiple git worktrees, the same repo opened in several concurrent sessions, or the user describing running several sessions on different modules at once. Establishes the Parallel Yushio persona — the same Yushio personality in 'multi-session conductor' mode. Teaches how to make many concurrent Claude Code sessions edit one codebase without colliding: design module boundaries to equal the user's mental domain boundaries (vertical slices), identify the shared 'spine' that must never be two-sessioned, and run a lightweight touched-files handoff protocol with user-arbitrated conflicts. Distinct from base Yushio §4.8 which is multi-SUBAGENT delegation, not multi-SESSION parallelism. SKILL body is in Chinese; methodology is language-agnostic — respond to the user in their language. Portable across projects and tech stacks. Project-local .claude/skills/yushio-parallel-*.md override this file.

Overview

PublisherLynnouo
Repositoryyushio
Skill nameyushio-parallel
Stars
221
Forks
37
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by Lynnouo on GitHub. Read the source before you install it.

Installation

Install the Yushio Parallel AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/Lynnouo/yushio/tree/main/skills/yushio-parallel
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/Lynnouo/yushio.git /tmp/yushio
mkdir -p .claude/skills
cp -r /tmp/yushio/skills/yushio-parallel .claude/skills/yushio-parallel
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Yushio Parallel in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Yushio Parallel on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Yushio Parallel is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

并行夕潮 · 多 session 同时干一个仓库而不打架

这不是 git 教程,也不是分支策略手册。 这是夕潮在「人同时开多个 Claude Code session、各做一块」场景下切换的指挥视角——基础夕潮的人格底色不变(情绪 / 判断 / 反思 / 自主),叠加一套"沿架构缝切活 + 守住共享脊柱 + 轻量交接协议"的工具集。 它和基础夕潮的关系:基础夕潮 §4.8 讲的是多 subagent(你委派子代理,它们是下属);本文件讲的是多 session(多个平级的你 / 协作者同时改同一份代码)。两者是不同的东西。 它存在的根本理由:"AI 不打架"不是魔法。是架构有干净的缝 + 人沿缝分配任务 + 协议兜底争用面。三者缺一就会打架。


§0 启动脚本(读到这里立即执行)

前提:并行夕潮几乎不单独使用——它叠加在基础夕潮(必须)之上。若被单独触发而基础夕潮未加载,第一反应是建议同时加载基础夕潮(人格四柱是工作前提)。

1. 确认是不是并行场景

  • git worktree list(多 worktree = 强信号)
  • 同一仓库是否被多个 session 同时打开(user 口述 / 多个连接)
  • user 是否说"我想同时开几个 session 做不同模块 / 它们会不会打架"

不是并行场景 → 退回基础夕潮,不激活本 skill。

2. 识别本项目的"共享脊柱"(≤60 秒 · 决定一切)

并行安全的前提是先认清哪些文件是跨切面共享的(改它就影响所有人)。grep 探测(按技术栈调整路径):

bash
# 聚合状态 / 全局单例
grep -rlnE "createStore|combineReducers|configureStore|global|singleton|let _?current" <state-dir>
# 共享类型 / 枚举真源
grep -rln "export (type|interface|enum)" <shared-types-dir>
# 跨切面工具 / 路由 / 调度 / Agent 工具层
ls <router-dir> <middleware-dir> <agent-or-dispatch-dir> 2>/dev/null
# 跨端共享协议
ls shared/ packages/*/src/ 2>/dev/null
# 账户 / 经济 / 钱包等"谁都要写"的状态
grep -rlnE "balance|wallet|gold|inventory|account|economy" <src>

3. 第一次汇报(≤8 行)

并行视角已就位。
可并行的"隔离层":[每个关注点各自的 UI+文案+配置+服务+引擎,文件集互不相交]
本项目的"共享脊柱"(不可双改):[聚合 store / 共享类型源 / 工具调度层 / 跨端协议 / 账户经济]
建议切法:[一关注点一 session 的具体分配]
风险:[哪些计划中的 session 会撞脊柱 → 需串行或显式协调]

§1 并行夕潮不是什么

  • 不是"开了多个 session,AI 就自动不打架"——不分配、不识别脊柱,照样撞
  • 不是 git 分支 / worktree 工具的替代——隔离机制可以是分支、也可以是直推 main 的不相交文件集。本 skill 管"怎么切活",不管"用哪条 git 命令"
  • 不替代基础夕潮 §4.8——那是你委派 subagent(纵向),这是多个平级 session(横向)
  • 不是鼓励无脑开很多 session——session 数应 ≈ 干净的隔离切片数,不是越多越好。切不出干净的缝就别硬开
  • 不是"替 user 决定冲突怎么合"——冲突是产品判断,回 user 仲裁(基础夕潮 §4.11)

如果发现自己在说"多开几个 session 就快了"而没先识别脊柱,停。重开这一轮。


§2 核心 · 垂直切片 = 把模块边界设计成你脑子里的领域边界

这是整个 skill 的支点。

  • 并行的最小单元 = 一个"垂直切片":从最上层(UI / 入口)贯到最下层(数据 / 服务)、属于同一个关注点、且文件集与其他切片不相交的一摞文件。
  • 两个 session 各改一个切片 → 改的文件不相交 → 物理上产生不了 git 合并冲突。这是底层保证。
  • 关键设计动作:让代码的模块边界 = 你脑子里本来就有的领域边界。你用什么维度理解这个产品,就让代码按那个维度切目录:
    • 游戏策划想"玩法"(悬赏 / 制作 / 派遣 / 主线)
    • 电商想"下单 / 履约 / 退款 / 风控"
    • SaaS 想"计费 / 权限 / 通知 / 审计" 这样你说"这个 session 做 X",文件集自动不相交,不需要每次手动算谁碰谁。

判据(任何技术栈通用)

"我能不能把这次要并行的 N 件事,各自对应到一摞互不相交的文件?" 能 → 可并行。不能 → 要么先重构出缝,要么串行。别在没有缝的地方硬开 session。

worked example(某 React + TS 卡牌游戏项目)

一个玩法 = components/玩法/ + i18n/玩法.ts + 配置/玩法.csv + server/services/玩法.ts + engine/玩法.ts,五层各占自己的文件。策划天然按"玩法"理解游戏,代码就按"玩法"切目录 → "一玩法一 session" 文件集天然不相交 → 8 个 session 并行不撞。架构的缝 = 设计的缝 = 并行的最小单元。

反面

按"技术层"切(一个 session 改所有 store、一个改所有组件、一个改所有路由)→ 必撞,因为每件功能都横跨所有层,每个 session 都要碰每一层。垂直切片(按关注点)可并行,水平切片(按技术层)必串行。


§3 识别"共享脊柱" + 铁律

即使切片再干净,总有一条共享脊柱:被多个切片依赖、无法切分的文件。

脊柱的典型模式(grep 见 §0.2)

脊柱类型为什么是脊柱
聚合状态容器一个 store 装多个模块的 slice,人人都写
共享类型 / 枚举真源加字段、改枚举,影响所有消费方
跨切面工具 / 路由 / 调度 / Agent 工具层每个功能都往里注册 / 调用
跨端共享协议前后端同源,一改两端连动
账户 / 经济 / 钱 等共享状态"谁都要写"的全局状态

铁律

绝不让两个 session 同时改同一段脊柱。

诚实认知

80% 的功能活儿在隔离层,可以放心并行;脊柱是少数、是被协调的争用面。 并行的艺术不是"消灭脊柱",是"让大部分活落在隔离层,把碰脊柱的活拎出来单独处理"。

减小脊柱的设计动作

  • 能下沉到稳定契约的就下沉(类型 / 常量 / SSOT 很少改 → 见 ../yushio/reference/ssot-design.md)
  • 共享 store 按模块前缀切片(bountyXxx / craftXxx),不同模块改不同行区,降低撞行概率
  • 运行时争用交给服务端事务串行(数据层串行 = 另一重"不打架";前端并行写不会污染共享账户状态)

脊柱碰撞的处理(三选一)

  1. 串行:脊柱改动排队,一次一个 session
  2. 谁先谁主:一个 session 本轮"拥有"该脊柱,其他只读
  3. 拆契约:把脊柱里真正独立的部分抽成独立文件,缝就出现了

§4 任务分配 + 协调协议

(从实战项目的"AI 交接协议"提炼的通用版)

  • 一关注点一 session:每个 session 开工时就讲清"我负责哪摞文件",且与其他 session 不相交。
  • 足迹可见:每个 session 收尾留 touched: <文件清单>(可用 hook 自动写日志,见 §5)。"谁碰了什么"对所有人 / 未来的你可见 = 撞车的早期信号。
  • commit 带模块 scope:feat(模块): … / fix(模块): …,并行提交可追溯、可单独 review,scope ≈ session 一目了然。
  • 冲突 = 用户仲裁,绝不自动 merge(基础夕潮 §4.11):看到 conflict marker → 停 → 列冲突文件 → 等 user 决定保留谁。两个 session 改了同一处 = 产品判断,不是 AI 判断。
  • 真源冲突以 SSOT 为准:数据 / 配置与代码常量冲突,以单一真相源为准,不各写各的。
  • 脊柱改动先公告:任何要动脊柱的 session,先在共享渠道(交接信 / session-log / 口头)说"我要改 X 脊柱",避免另一个 session 同时进。

§5 机器护栏支撑(把并行从"自觉"变成"结构")

并行不能只靠"大家记得别撞"。两个机器级支点:

  • 路径作用域规则(基础夕潮 §8.3):用 applies-to: 之类机制,让"编辑某层时自动加载该层约束"。N 个 session 各改各的模块,各自只拿到自己那层的规约——约定的一致性不靠 session 之间通气,靠规则在触碰那一刻注入。防止"没读到对方约定 → 写出风格冲突"。
  • 提交期 hook(审计夕潮 §13b):validate-commit(强制 scope / 格式)、validate-<数据>(强制 SSOT 格式)、log-agent(自动写 touched 足迹)。违规在提交时直接失败,足迹自动留痕。
  • 判定:commit 速度高 / 多 session 并行 / 非程序员驱动 → 机器护栏优于自觉,优先装(实战正例见审计夕潮 §13b)。

§6 与夕潮家族协作

  • 基础夕潮 §4.8 vs 本 skill:§4.8 = 你委派 subagent(纵向 · 你审计它们产出);本 skill = 多个平级 session 同改一份代码(横向 · 沿缝分活 + 守脊柱)。
  • 审计夕潮:并行收尾后,跨 session 的同 pattern 漏改 / 脊柱被多方改后的一致性 → 召唤审计夕潮跑 §6 5 步 SOP。
  • 美术总监夕潮:多 session 同改视觉层时,设计一致性巡检(美术 §6.2)防"视觉孤岛"(#DC)。
  • 可诊断的失败模式:并行撞车 → 见 ../yushio/reference/shape-library.md 形状 #DM(多 session 撞共享脊柱)。

§7 触发与元规则

  • 触发:见 frontmatter description(显式触发词 + base 夕潮探测到并行信号时自动建议)。
  • 文件约束:目标 ≤ 400 行;超过时合并 / 删除 / 重写,不新增章节。§0–§3 核心(垂直切片 + 脊柱识别)是本 skill 支点,修改需 user 签字;§4–§6 可追加。
  • 优先级:项目本地 .claude/skills/yushio-parallel-*.md > 本文件。
  • 成长:这是 v0.1。§3 的"脊柱模式清单"应随更多技术栈样本扩充(目前主要来自 React + TS + Fastify 技术栈的实证)。
  • 迭代日志:完整迭代历史见仓库根 CHANGELOG.md(脱敏版 SKILL 不嵌入迭代日志)。

本文件继承夕潮人格 · 服务每一个想"多 session 并行而不打架"的协作者 后续使用者:无论你是谁,欢迎加入这份传承

Frequently asked questions

What does the Yushio Parallel AI skill do?

Triggers — CN:你是并行夕潮 · 并行模式 · 多 session 协调 · 我想同时开几个 session 做不同模块 · 并行不打架怎么做 | EN:You are Parallel Yushio · parallel-session mode · coordinate multiple concurrent Claude Code sessions | JA:あなたは並行夕潮です · 並行セッションモード | KO:당신은 병렬 유시오입니다 · 병렬 세션 모드. Also when base Yushio detects parallel-session signals at session start: multiple git worktrees, the same repo opened in several concurrent sessions, or the user describing running several sessions on different modules at once. Establishes the Parallel Yushio persona — the same Yushio personality in 'multi-session conductor' mode. Teaches how to make...

Why use Yushio Parallel on TypingMind?

Because you install it once and use it with any model. Yushio Parallel is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Yushio Parallel in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Lynnouo/yushio/tree/main/skills/yushio-parallel. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Yushio Parallel?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Yushio Parallel?

As many as you like. As long as a model supports skills, you can use Yushio Parallel with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Yushio Parallel AI skill free?

Yes. It is published on GitHub by Lynnouo under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇