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.

DeepSeek Harness Plugin Hub
ProfilesPluginsCategoriesNewsDocsSign inManage Profiles
ProfilesPluginsCategoriesNewsDocsSign in
← Plugins

dsh-haoba-mode

Haoba Mode

豪霸模式 (Haoba Mode) for DeepSeek Harness: one gold switch beside the composer that pushes every model step to its maximum effort and token budget, with an exhaustive no-shortcuts working discipline.

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

npx -y @deepseek-ai/dsh plugin --profile web add dsh-haoba-mode@0.1.0
READMECompatibilityVersions

Compatibility and provenance

Haoba Mode is published as dsh-haoba-mode 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
web
Release source
npm
Registry updated
10/23/2026

Versions

0.1.0stable
10/23/2026

Related plugins

Loading related plugins…

Latest
0.1.0
DSH
*
HMR
Process restart
Tree shaking
Safe tree shaking not declared
Unpacked size
31.9 kB
Files
11
Surface
web
License
MIT
Source
npm
GitHub
★ 1
Weekly downloads
0
Last push
10/23/2026
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

Related plugins

More verified plugins in models-usage.

Experimental Cot Translation Bundle@deepseek-ai/dsh-experimental-cot-translation-bundleExperimental machine translation of expanded reasoningOur Free Model@ebony-vinyl/dsh-our-free-modelOur Free Model — Install this plugin in dsh to use cutting-edge models including DeepSeek V4.1 Flash and Kimi K3, with no login, registration, or API Key required. The desktop version also offers an EAC co-payment channel; starting with 2.0, it supports ten account channels: LobsterAI, Qoder, TRAE, Usage@linxin666/dsh-usageUsage statistics plugin for the dsh web GUI: per-provider balance and coding-plan quota detection plus a live token usage ledger, with a compact price glance card above the sidebar's Settings rowClient Ui Model Capabilities@linxin666/dsh-client-ui-model-capabilitiesPer-model capability declarations (image input and reasoning efforts) for custom pi-ai providers, edited in place on the Models settings cards; writes the official llm-pi-ai settings namespace, no DSH source changes.

README

豪霸模式 · dsh-haoba-mode

性能最好、最费 token 的一档。聊天框旁边一枚土豪金开关,打开就别想省钱。

一枚开关,两种世界:

  • 关:一切照旧,插件只在后台待命。
  • 开:从下一条消息开始,每一步模型调用都被推到该模型支持的上限,并在系统提示里追加一段「豪霸纪律」。

安装

在 DeepSeek Harness 里打开 插件 → 添加插件,在输入框里填包名,点 安装,装好后点 立即启用:

dsh-haoba-mode

也可以填 GitHub 仓库地址 https://github.com/FelixJiashengFu/dsh-haoba-mode。连不上 npm 官方源时,在 安装源 里选中国大陆镜像源再填包名。

输入框旁随即出现金色的「豪霸」开关。升级:在插件页卸载后,用同样的方式重新安装。

它到底做了什么

杠杆实现效果
推理强度agent/request 瀑布里读取 llm.resolveModelInfo(),把 reasoningEffort 换成该模型对外声明的最强一档思考更久、更不容易跳步
输出预算同一处把 maxTokens 抬到模型的 defaultMaxTokens不因默认输出上限被截断
工作纪律systemPrompt.section 注册一段长提示(order: 900,靠后)穷尽检索、完整交付、逐条自检、失败即诊断
视觉关闭时中性描边,开启时金色渐变 + 流光 + 全局品牌色转金一眼看得出花了钱

三点说明:

  1. 不换模型。它只在你当前选中的模型上把预算拉满,不改 provider/model,不会打破会话的模型粘性或安全设置。想更贵,自己在模型选择器里挑更贵的。
  2. 只喊模型真有的档位。reasoningEffort 取自适配器声明的候选,绝不硬塞一个会被拒的值;模型没有推理档位时只抬 maxTokens。
  3. 不越过安全边界。更大的预算不等于更多权限:破坏性操作仍然要确认。

结构

package.json        dsh.bundle.patch + dsh.client(把 lib/client.js 送进 Web 客户端)
cordis.patch.yml    Host 半边挂载行
lib/protocol.js     路由、section 名与排序、模式名(浏览器半边复制一份常量)
lib/prompt.js       豪霸纪律提示词
lib/index.js        Host:状态持久化 + 回路 HTTP 路由 + agent/request + systemPrompt
lib/client.js       浏览器:金色开关(仅依赖 react,无 JSX、无 import)
locale/zh.json      插件市场里的中文名与简介
locale/en.json      英文名与简介
icon.svg            金色皇冠

状态与接口

开关状态持久化在 ~/.dsh/haoba-mode.json({ "enabled": true }),重启后自动恢复。

回环 HTTP 路由 GET|POST /plugins/dsh-haoba-mode/state:

{ "ok": true, "enabled": true, "lastError": null,
  "applied": { "provider": "…", "model": "…", "reasoningEffort": "high", "maxTokens": 65536, "at": 0 } }

applied 是最近一次真正改过的调用参数,开关的悬浮提示会把它显示出来。

本地开发与卸载

本地开发:克隆本仓库后,在 添加插件 里填本目录的绝对路径,或让 agent 调用插件管理器(不要手改 profile 的 package.json):

plugin_manager action=install_bundle target=<本目录绝对路径>

卸载:在插件页卸载,或 plugin_manager action=remove_bundle target=dsh-haoba-mode。

代价

每步多出的 token 主要来自两处:更长的推理、以及每次请求都携带的纪律段落(约 600 tokens/步)。这是这个模式的设计目的,不是缺陷——它用钱换确定性和完整度。真在意账单时把开关关掉即可,状态立即生效,无需重启。