← Back to list

dsh-zh-thinking

DeepSeek Harness Models & Providers Low risk

DeepSeek Harness (dsh) 插件:注入系统提示词,强制 Agent 的思维链 / 规划 / 工具推理全程使用简体中文,防止中文思考漂移为英文。零外部依赖,任意本地目录 link 即装。

How to install

DeepSeek Harness dsh plugin add github:yingjian666/dsh-zh-thinking

About

dsh-zh-thinking 让 DeepSeek Harness 里 Agent 的**内部思考过程**——思维链(chain-of-thought)、逐步规划、工具调用前后的推理、自我审查——始终使用**简体中文**的宿主插件(host plugin)。 **声明**:本项目参考 / 灵感来自 wodongx123/dsh-language-control(其 npm 包为 @deepseek-ai/dsh-language-control),在其「让 Agent 用中文思考」的思路基础上**独立重写**,并解决了本地目录挂载时的依赖加载问题。两者的关系与差异见下文「与原插件的关系」。 --- 为什么会有这个插件 DeepSeek 这类推理模型,其「思考」由同一份系统提示词(system prompt)驱动。当工具返回英文内容(例如网页抓取、文档、报错信息)时,Agent 的思维链容易发生**语言漂移(language drift)**——思考逐步切到英文,导致后续推理与最终回答的语言不稳定。 本插件在启动时向系统提示词注册一段「必须用简体中文思考」的规则,让模型的内部思考保持中文,而最终回答是否默认中文则交给模型自身判断(可保留英文专有名词、代码标识符等)。 与原插件的关系 | 维度 | 原插件 @deepseek-ai/dsh-language-control | 本插件…

Recommendation signals

42 Tool quality · Based on stars, downloads, maintenance, security and docs
User interest · Adjusted by in-site views, install copies and download clicks
42 Overall
0views
0unique visitors
0install copies
0download clicks
0outbound clicks

Meta

License
MIT
Language
JavaScript
GitHub stars
0
mo. downloads
Last push
2026-09-08
Created
2026-09-06

Links

GitHub ↗ Report issue ↗

Basic safety check

Findings
curated 收录但无 npm 包/安装命令
Sources
curated:awesome-dsh-plugin/awesome-dsh-plugin
Topics
agent, chain-of-thought, chinese, deepseek-harness, deepseek-harness-plugin, dsh, dsh-plugin, llm, simplified-chinese, thinking, zh

Related plugins

DeepSeek HarnessClaude CodeCodex
Score81

dsh-plugin-subscriptions

V1ki/dsh-plugin-subscriptions

Use ChatGPT (Codex), Claude, and Grok subscriptions as DeepSeek Harness LLM providers, with Settings login, model catalogs, usage, plus image_generate, video_generate, and x_search tools.

☆ 335 ↓ 9.2K Models & Providers
Claude Code
Score74

claude-code-router

musistudio/claude-code-router

One local control plane for every AI agent: route across models, fuse new capabilities, orchestrate tools, and stay fully in control.

☆ 37.2K ↓ – Models & Providers
DeepSeek Harness
Score73

dsh-thinking-effort

hytime/dsh-thinking-effort

Configure reasoning-effort levels and sub-agent defaults for custom DSH models.

☆ 28 ↓ 5.3K Models & Providers
DeepSeek Harness
Score72

dsh-llm-fallbacks

btspoony/dsh-llm-fallbacks

Role-based LLM retry & fallback strategies.

☆ 20 ↓ 5.7K Models & Providers