dsh-reply-language
Run the following command in DeepSeek Harness:
dsh plugin install CAI-MH/dsh-reply-language
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install CAI-MH/dsh-reply-language in your terminal to add the plugin; the source is hosted at https://github.com/CAI-MH/dsh-reply-language , and you only need to restart dsh web after installation for it to take effect.
About this plugin
When working through multi-turn conversations in DeepSeek Harness, the model's reply language tends to drift—mixing English terms into a Chinese explanation, or slipping back into Chinese when you need a clean English output. Inconsistent reply language is especially painful during technical writing, documentation generation, and debugging.
dsh-reply-language injects a high-privilege system-prompt section that mandates every reply—explanations, summaries, body text, and error messages—be written in the language you configure, while keeping code, shell commands, file paths, API names, and proper nouns in their original form. Two built-in tools make it fully interactive: reply_language_set lets you switch the target language mid-conversation (Simplified Chinese, English, Japanese, French, and more), and reply_language_get shows the current setting. No configuration file is needed; it works out of the box.
Whether you are a developer who needs pure Chinese technical output, an engineer collaborating in English, or a researcher who flips between languages across a session, this lightweight host-only bundle keeps DSH's reply language consistently aligned with your intent—no client UI, no extra setup.
Use Cases
- Requiring the model to explain technical concepts entirely in Chinese across multi-turn chats
- Switching the reply language to English on the fly in a collaborative workspace
- Keeping error messages, summaries, and diagnostics in one consistent language during debugging
Best For
- Developers who need consistent Chinese technical output from the model
- Engineers who switch working languages in bilingual teams
- Researchers who want full control over reply language in DSH sessions
Related Plugins
A unified suite combining hot runtime injection, task-aware thinking-mode routing, and a graded session protocol with red-team gates to sustain model diligence across long-horizon inference.
ModLens is a vision plugin for DeepSeek Harness that gives text-only models sight by reading images pasted directly into chat, with zero-config setup and multiple vision engines.
On-demand vision for text-only DeepSeek Harness agents: built-in free keyless vision chain and 14 vision tools, routing image turns as tool calls to vision models with pixel fidelity, no Python needed, one-command install.
Give text-only models in DeepSeek Harness eyes, enabling image Q&A, long-screenshot OCR, UI restoration, and GUI visual tasks.