Claude Code 技能 + 子代理：查询、交接、看板工作流、pair（经理+实现者）以及带有合并门的无人值守编排器。麻省理工学院。
概览
Krish 的 Claude 代码技能、子代理和开发配置。 这是一个个人收藏的Claude代码技能、子代理和开发配置，分享给你（或者我在另一台机器上）也能安装和使用它们。 A 是一个目录，带有一个 SKILL.md，Claude Code 在判断该技能应用时（或你输入 / 时）加载到当前对话中。A 是一个 .md 文件，定义了一个独立的工作者，拥有自己的上下文窗口、模型、工具允许列表和预加载技能;Claude 为独立工作生成该工作。技能是“如何做 X”;代理是“谁用什么权限做 X”。 每个技能文件夹有两个文件：SKILL.md 是 Claude 遵循的提示（前置页 + 指令），README.md 是给你的：它的功能、使用时间、如何调用、逐步工作原理以及示例。代理在 agents/README.md 中被一起文档化;每个代理 .md 是 Claude 加载的定义。从 README 开始;只打开一个 SKILL.md 以查看或更改具体行为。 刚加入仓库？
概述（AI 翻译）
Krish 的 Claude 代码技能、子代理和开发配置。
这是一个个人收藏的Claude代码技能、子代理和开发配置，分享给你（或者我在另一台机器上）也能安装和使用它们。
skill 是一个目录，带有 SKILL.md]，当 Claude Code 判定该技能适用时（或你输入 / ]）时，会加载到当前对话中。subagent 是一个 .md 文件，定义了一个独立的工人，拥有自己的上下文窗口、模型、工具允许列表和预加载技能;Claude 为独立工作生成该文件。技能是“如何做 X”;代理是“谁用什么权限做 X”。
如何阅读这个仓库
每个技能文件夹有两个文件：SKILL.md 是 Claude 遵循的提示（前言 + 指令），README.md 是给你的：它的功能、使用时间、如何调用、逐步工作原理及示例。代理在 agents/README.md 中被共同记录;每个 agents/ .md 是 Claude 加载的定义。从 README 开始;打开 SKILL.md 只查看或更改具体行为。
从这里开始
刚开始使用仓库？查看先修课程，安装好，然后试试一个技能。
| 需要 | 为什么 | 检查 |
|---|
本页由 AI 翻译，技术细节请以原文为准。
README
以下为原文 README 节选，完整内容请查看原文。
skills
Krish’s Claude Code skills, subagents and dev config.
A personal collection of Claude Code skills, subagents and dev config, shared here so you (or I, on another machine) can install and use them too.
A skill is a directory with a SKILL.md that Claude Code loads into the current conversation when it decides the skill applies (or when you type /). A subagent is a .md file that defines a separate worker with its own context window, model, tool allowlist and preloaded skills; Claude spawns it for isolated work. Skills are “how to do X”; agents are “who does X, with what permissions”.
How to read this repo
Every skill folder has two files: SKILL.md is the prompt Claude follows (frontmatter + instructions), and README.md is for you: what it does, when to use it, how to invoke it, how it works step by step, and an example. Agents are documented together in agents/README.md; each agents/.md is the definition Claude loads. Start with the READMEs; open a SKILL.md only to see or change exact behaviour.
Start here
New to the repo? Check the prereqs, install, then try one skill.
| Need | Why | Check |
|---|---|---|
| Claude Code | runs every skill and agent here | claude --version |
gh CLI, authenticated, origin on GitHub |
pair, meta-orchestrator, pr-watch, gap-scan and everything under kanban/ read and write GitHub. kanban/, wayfinder and pr-watch stop with a message on a non-GitHub remote. Kanban Projects boards also need gh auth refresh -s project -s read:project once. Every gh call names origin’s repo (-R). On a fork clone, bare gh would default to the parent, so skills ask which repo to write to and never write to the parent without a yes naming it (Target repo) |
gh auth status |
| Python 3.8+ | the lint step of the install, scripts/build_site.py, and the guard-git hook that stops agents pushing to main or merging |
python --version |
| Git Bash (Windows only) | the guard-git hook is a POSIX sh command; without it the hook fails open, so the installer refuses |
sh --version |
Which skill first
grill-docs: zero setup, and it leaves a decision record (skilleddocs/decisions.mdunless the repo has its own) the other skills read. Good for seeing what a skill writes.pair: one task end to end in an isolated worktree, you push at the end. The smallest taste of the agents.meta-orchestrator: a whole backlog unattended. Try it last; it is the most autonomous and builds on the two above.
60-second quickstart
In Claude Code:
/plugin marketplace add KrishP147/skills
/plugin install krish@krish
推荐工具
换一个关键词，或者移除筛选条件。
安装
npx skillfish add krishp147/skills