agentleFS
PC2005-cloudPC2005-cloud/dsh-pet/.agents/skills/dsh-pet-dev/SKILL.md

dsh-pet-dev

Skill5.9k starsChanged 15 days agoMITView on GitHub

本仓库(dsh-pet 宠物插件)的开发总则:把任务边界、插件代码规范、注释规范与提交规范串成一条执行链,一次调用覆盖从改代码到提交的全过程。

Adopted by

Steady: 0 new projects took it in the last 4 weeks, none the 4 weeks before.thin data

Week of Sep 13: 0 new projectsWeek of Sep 20: 0 new projectsWeek of Sep 27: 0 new projectsWeek of Oct 4: 0 new projectsWeek of Oct 11: 0 new projectsWeek of Oct 18: 0 new projectsWeek of Oct 25: 0 new projectsWeek of Nov 1: 0 new projects
Sep 2026Oct 2026Nov 2026
New projects taking a version of this file each week, over the 8 weeks our latest run read: only 0 in all, so this is a hint, not a trend.

Who's talking about PC2005-cloud/dsh-pet

The biggest 5 of 14 mentions linking to the repository that our crawl found. Each line is their own headline.

See all 14

Talked about on GitHub and Bluesky (14 times).

On GitHub · 13

8 more on GitHub

The file

Files
SKILL.md

dsh-pet 开发总则(dsh-pet-dev)

目标

本仓库唯一的入口 skill。它不重复四条规范的内容,只回答一件事:这次请求该调用哪几个子 skill、按什么顺序执行。

四个子 skill 是权威正文;本文件是索引与顺序表。二者冲突时以子 skill 为准。

四个子 skill

子 skill什么时候必须调用
task-guardrails每个请求的第一步,任何工具调用之前
dsh-plugin-standards要动 host/client 源码、加删文件、改接口、补测试
dsh-plugin-comment-spec和 dsh-plugin-standards 一起调:写代码就要写注释
git-commit要提交、要写或改写提交信息

四个子 skill 都已注册,直接按名字交给 skill 工具,不需要去找文件,也不需要读文件。门禁命令在 dsh-pet/ 子目录下执行。

执行链

按顺序走,不要跳步,也不要一次全部调用。

  1. 判类型(task-guardrails):纯对话 / 只读调查 / 需要变更。判定不清时往更保守的一类靠。
  2. 先开口再动工具:说出理解与打算看什么;只读也要先开口。
  3. 写操作前停住:说清「改什么、达到什么效果、怎么验证」,然后用弹窗等回复。预告不等于获批。
  4. 动代码前一起调 dsh-plugin-standards 与 dsh-plugin-comment-spec:前者定插件类型、文件规模、目录归属、导出面、测试位置,后者定模块头分档、导出符号一行摘要、只写 why。写代码必然带注释,不拆成两步。
  5. 收尾跑门禁(在 dsh-pet/):npm run typecheck、npm run lint、npm run format:check、npm run check:cycles、npm run check:runtime、npm run check:openapi、npm test、npm run build。没跑过的不要说「已验证」。
  6. 提交时调 git-commit:从实际 diff 推断 type(scope),正文写实现方式、行为变化与涉及文件;推送必须有明确指令。

调用方式

  • 按需调用,别一次全调:纯问答一个都不调;只读调查只调 task-guardrails;改代码把 task-guardrails、dsh-plugin-standards、dsh-plugin-comment-spec 三个一起调;提交调 git-commit。
  • 同一子 skill 在一次任务里调一次即可,不要重复调用,也不要凭记忆复述它的规则。
  • 本 skill 只由用户点名触发;四个子 skill 由 AI 按需调用,用户也可以直接点名。

冲突与优先级

  • 边界优先:task-guardrails 决定「能不能动」,dsh-plugin-standards / dsh-plugin-comment-spec 决定「怎么写」,git-commit 决定「怎么记录」。规范与边界冲突时,先满足边界。
  • 规范与代码不一致时报告,不擅自改:发现规范里的说法与仓库现状冲突,先说明再问,不直接改规范。

守则

  1. 没有明确授权,不产生任何副作用:不写文件、不装依赖、不动 Git。
  2. 不把举例当任务:用户描述问题或吐槽时,先确认理解再动手。
  3. 删而不留:确认无引用后删除,不留兼容层、不留旧实现、不留注释掉的代码(除非用户要求)。
  4. 一处事实只写一遍:定位算术、常量、类型不复制第二份;规范内容不搬进本文件。
  5. 提交信息来自本次 diff,不照抄本 skill 的示例文字。

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.