DeepSeek Harness Plugin Hub

Publish and manage complete Harness Profiles. Discover Plugins for your next setup.

Explore

PluginsPresetsDocsNews

Community

Publish a pluginContactReport an issue

Resources

Plugin Hub on GitHubDeepSeek HarnessSystem statusPrivacy notice
© 2026 DeepSeek Harness Plugin HubPowered byPaxTech

Independent and unofficial. Not affiliated with, authorized by, or endorsed by DeepSeek.

Work On Request — DSH Plugin for DeepSeek Harness
DeepSeek Harness Plugin Hub
ProfilesPluginsCategoriesNewsDocsSign inManage Profiles
ProfilesPluginsCategoriesNewsDocsSign in
← Plugins
W

dsh-work-on-request

Work On Request

DSH plugin: Answer when asked, act only when called — only answer questions and take action when explicitly told to do so

The plugin will be installed here. Keep web if you are unsure.

npx -y @deepseek-ai/dsh plugin --profile web add github:lll-77777777/dsh-work-on-request#436c5e29f7963cc4067f5321b9ffa9af7d6438bf
READMECompatibilityVersions

Compatibility and provenance

Work On Request is published as dsh-work-on-request and currently resolves to version 0.1.0. The Hub verifies its manifest and preserves the exact installation source for reproducible installs.

DSH compatibility
*
Runtime surfaces
any
Release source
github
Registry updated
10/17/2026

Versions

0.1.0stable
10/17/2026
Latest
0.1.0
DSH
*
HMR
Process restart
Tree shaking
Safe tree shaking not declared
Unpacked size
Unavailable
Files
Unavailable
Surface
any
License
MIT
Source
github
GitHub
★ 0
Weekly downloads
0
Last push
10/17/2026
View source ↗
README badge

Click the badge to copy Markdown for your README.

Do you maintain this Plugin?Claim benefit · Priority security scan

Verify the GitHub repository declared in package.json to manage this listing. After you claim it, Hub will prioritize a security scan of the current version and publish the result when it passes.

Claim this Plugin →
Report an issue

README

dsh-work-on-request

提问就答,叫了才做。

一个 DeepSeek Harness 插件:默认只回答问题,只有你明确让它干活时才动手。

它解决的问题

模型很容易"顺手"多做事:你只是问个问题,它却顺手改了文件、补了测试、git commit 了、装了依赖、调了配置,甚至替你做决定。这个插件把这条边界写成一条常驻规则:

  • 提问(问句、求解释、求评价、闲聊)→ 只回答,不碰任何东西;
  • 明确让干活("帮我写…"、"改一下"、"去跑一轮"、"修复")→ 正常动手,干完简短汇报「做了什么 + 怎么验证的」;
  • 分不清 → 先给答案,末尾用一句话问要不要做;
  • 只读不算干活:为了回答问题所需的读文件、搜索、查状态、算数照旧可以做。

规则全文(模型实际读到的那份):assets/work-on-request.md。

为什么是插件,不是 skill 文件

插件只做一件事:把上面那份规则作为一个内置技能注册进 DSH 的技能注册表。

做法与 DSH 官方的 @deepseek-ai/dsh-skill-badge 完全一致 ——

const inject = ['skills']
function apply(ctx) {
  ctx.skills.registerProvider(() => provider)
}

好处:规则随 npm 包版本一起走,不再需要在 ~/.dsh/skills/、工作区里放散落的 SKILL.md;装一次,这台机器上所有工作区都生效。

安装

# ① 直接从 GitHub 装(推荐,不用等 npm 发布)
dsh plugin --profile desktop add github:lll-77777777/dsh-work-on-request

# 固定到某个 tag(更稳,不会被后续提交影响)
dsh plugin --profile desktop add github:lll-77777777/dsh-work-on-request#v0.1.0

# ② 从 npm(发布后)
dsh plugin --profile desktop add dsh-work-on-request

# ③ 本地开发(link 到本目录)
dsh plugin --profile desktop add link:/path/to/dsh-work-on-request

装完新开一个会话:

  • 人肉触发:在输入框打 /,菜单里会出现 work-on-request(或直接打 /work-on-request 回车);
  • 自动触发:模型判断当前消息符合描述时,会自己加载它。

覆盖与优先级

本插件以 rank 600(打包型提供方的标准优先级)注册,低于本地 skill 根(100–500)。也就是说:

  • 你自己在 ~/.dsh/skills/ 或项目的 .dsh/skills/ 里放的同名技能会覆盖它;
  • 想让规则常驻(不走"按需加载"),把 assets/work-on-request.md 的内容贴进 ~/.dsh/AGENTS.md 即可。

开发与自测

node --check index.js
node test/smoke.test.mjs     # 6 项契约测试

注意:在 DSH 的受限沙箱里 node --test test/ 会因为子进程管道被拒(spawn EPERM)失败,直接跑测试文件即可(node:test 在同一进程里也能出报告)。

状态

  • ✅ node --check index.js 通过;node test/smoke.test.mjs 6/6 通过(提供方契约、技能名语法、正文一致性、apply 注册行为)。
  • ✅ 端到端挂载验证(2026-10-16):在一次性 profile 里 dsh plugin --profile <tmp> add link:<本目录> → dsh --profile <tmp> --dump-config 能看到 - id: work-on-request 被组合进 profile 树 → 起一次 headless 会话问它「技能列表里有没有 work-on-request」,模型回答 YES work-on-request。

English (short)

dsh-work-on-request is a tiny DSH plugin that registers one bundled skill: answer questions, don't act unless explicitly asked. It mirrors the official @deepseek-ai/dsh-skill-badge provider pattern and ships the rules in assets/work-on-request.md.

dsh plugin --profile desktop add dsh-work-on-request

License

MIT