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.

Typesafe Ai Skill — DSH Plugin for DeepSeek Harness
DeepSeek Harness Plugin Hub
ProfilesPluginsCategoriesNewsDocsSign inManage Profiles
ProfilesPluginsCategoriesNewsDocsSign in
← Plugins

dsh-typesafe-ai-skill

Typesafe Ai Skill

Register the TypeSafe / Jev (System One judgment model) building skills in DeepSeek Harness’s global skill layer. Tasks such as classification, routing, scoring, extraction, and confidence gating—those that have AI make judgments—can be called directly in every session. No keys are included; the API

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

npx -y @deepseek-ai/dsh plugin --profile web add dsh-typesafe-ai-skill@1.0.1
READMECompatibilityVersions

Description

Register the TypeSafe / Jev (System One judgment model) building skills in DeepSeek Harness’s global skill layer. Tasks such as classification, routing, scoring, extraction, and confidence gating—those that have AI make judgments—can be called directly in every session. No keys are included; the API key is provided by the user through the OPENROUTER_API_KEY environment variable.

Compatibility and provenance

Typesafe Ai Skill is published as dsh-typesafe-ai-skill and currently resolves to version 1.0.1. The Hub verifies its manifest and preserves the exact installation source for reproducible installs.

DSH compatibility
*
Runtime surfaces
any
Release source
npm
Registry updated
10/2/2026

Versions

1.0.1stable
10/2/2026
1.0.0stable
10/2/2026

Related plugins

Loading related plugins…

Latest
1.0.1
DSH
*
HMR
Process restart
Tree shaking
Safe tree shaking not declared
Unpacked size
28.2 kB
Files
6
Surface
any
License
MIT
Source
npm
GitHub
★ 1
Weekly downloads
0
Last push
10/2/2026
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 agents-orchestration.

Headless@deepseek-ai/dsh-headlessThe dsh one-shot bundle: a direct core Agent/Session runner over dsh-base with no Host, HTTP, or browser layerExperimental Ralph Bundle@deepseek-ai/dsh-experimental-ralph-bundleRun explicitly requested repeated work with fresh agents.Hindsight Coding Agents@vectorize-io/hindsight-coding-agentsReflect-only Hindsight long-term memory for coding agents (harness-pluggable: opencode, …), with automatic background ingestion (no setup CLI).Experimental Agent Team Web Profile@deepseek-ai/dsh-experimental-agent-team-web-profileExperimental Web profile layer for Agent Teams Remote and UI plugins

README

dsh-typesafe-ai-skill

把 TypeSafe / Jev 的构建技能装进 DeepSeek Harness,每个会话自动可用。

这个插件做什么

它在 Harness 的全局技能层注册一个名为 typesafe-ai 的技能。因为注册来自主机行而不是某个 agent preset,所有会话——包括以后新建的会话——都能看到它,不需要每个 preset 各配一遍。

技能本身教模型怎么用 TypeSafe 的 System One 判断模型(旗舰模型是 Jev):把自然语言和应用状态转成带类型的判断和概率,代码可以直接 if 它,不用再「调大模型 → 解析回复」。

三种原语:

原语回答什么返回
Noul某个条件成不成立0–1 的概率
Choice从定义的选项里选一个选中项 + 各项概率 + 置信度
Score沿有序等级打一个分概率加权位置 + 等级图例

适用场景:给消息/工单/请求分类或路由、按某个维度打分排序、判断意图与紧急度、从自由文本抽取值、用置信度决定能否自动处理、拿不准转人工。

安装

dsh plugin --profile web add github:chen112233usa/dsh-typesafe-ai-skill

装完技能即刻可用。不需要重启:技能注册在挂载时完成,目录会自己刷新。

配置 API key

本插件不含任何密钥,密钥由使用者自己提供。先去 https://openrouter.ai/keys 建一把,然后在启动 dsh 之前导出:

export OPENROUTER_API_KEY=sk-or-v1-...        # macOS / Linux
$env:OPENROUTER_API_KEY = "sk-or-v1-..."      # Windows PowerShell

技能里的所有示例代码和 curl 命令都从这个环境变量读密钥,所以密钥不会被写进任何文件、提示词或提交记录。忘记配置时调用会返回 401,技能文档的排错表里有这一条。

技能内容

SKILL.md 是技能本体,也是这个仓库真正的主体。前半部分是给模型的行为指令(怎么从用户目标倒推judgment、怎么设计问题、怎么用置信度),后半部分是实测过的部署配置:端点、模型 ID、三种原语的 schema、curl / Node / PHP 调用示例,以及踩过的坑——比如 Choice.criteria 收对象而 Score.criteria 收对象数组,传反就是 400。

技能每次加载都会提示模型去读实时文档(https://docs.typesafe.ai),所以版本相关的细节以官方文档为准,而不是以本仓库为准。

插件本身的行为边界

  • 只读取与自己同目录的 SKILL.md,然后调用 Harness 的技能注册接口。
  • 自己不发起任何网络请求。 联网调用是技能教给模型做的事,用的是使用者自己的 key。
  • 卸载后注册随之撤销,技能目录立即恢复原状。

不装插件也能用

SKILL.md 是自包含的,不依赖 Harness 的任何特性。把它全文贴进任意平台的系统提示或技能目录即可:Claude Code 放到 ~/.claude/skills/typesafe-ai/SKILL.md,ChatGPT Projects / Gemini Gems 直接粘进指令栏。装成插件只是省掉手工步骤,并让它在每个会话里自动出现。

卸载

dsh plugin --profile web remove dsh-typesafe-ai-skill

许可

MIT。技能内容源自 TypeSafe 官方插件,部署配置由实际调用线上 API 验证后整理。

插件由 陈先生Cool 研发,允许全网全球免费使用。