dsh-jev
DeepSeek Harness plugin that classifies each user message's emotion, urgency and department with Jev and injects it as context.
Overview
This dsh-jev plugin for DeepSeek Harness sends the latest user message each round to Jev and asks a configurable set of questions, by default department, emotion and urgency. The results are injected into that step's request as a plugin-sourced context message so the model can adjust its tone and priority.
Everything is configured in the GUI settings page: enable or disable calls, toggle injection while still logging, API key (stored as a secret and never returned to the browser, falling back to TYPESAFE_API_KEY), endpoint, model, timeout (skipping injection rather than blocking on timeout), maximum input length, a behaviour hint, and the questions themselves in noul, score or choice form. A test button and a recent-calls view aid debugging. It is a community plugin documented in Chinese.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Which department should this message go to? | — |
| Choice | What emotion does the user express? | — |
| Score | How urgent is this message? | — |
How it works
- Before each step, take the latest user message
- Ask Jev the configured questions
- Inject the results as plugin-sourced context
- Model adapts tone and priority; timeouts skip injection
Strengths
- Configurable questions in a GUI
- Non-blocking on timeout
- Secret key never sent to the browser
Caveats
- Messages are sent to TypeSafe
- Requires a restart after install
- Chinese documentation
Getting started
dsh plugin --profile web add github:Excalibur9527/dsh-jev, restart dsh web, then configure under Settings → JEV 情绪分析.At a glance
- Type
- Plugin
- Status
- beta
- Language
- TypeScript
- Runtime
- DeepSeek Harness plugin
- Has code
- yes
- License
- MIT
Use cases and integrations
How curated lists describe it
This DeepSeek Harness plugin sends each round's latest user message to the systemone (Jev) API for emotion and intent classification and injects the result as plugin-sourced runtime context, with API Key and all parameters configured in the GUI settings page.
logicrw/awesome-jev-projects
Frequently asked questions
What does Jev classify?
Each latest user message's department, emotion and urgency by default.
How are results used?
They are injected as context so the model adjusts tone and priority.
What if Jev times out?
Injection is skipped for that round and the conversation continues.
Related projects
Zero-dependency Android app that judges the visible WeChat message's intent, emotion, urgency and reply posture with Jev.
AppbetaJavaAndroid WeChat assistant where Jev judges the chat and scores the relationship, and a chat model drafts three replies.
AppbetaKotlinWeChat plugin that classifies a long-pressed message's intent, emotion, urgency and suggested reply posture with Jev.
PluginbetaJavaLocal chat analyser that labels each WeChat message's emotion and intent with Jev and grades how your replies landed.
ProjectprototypeJavaScript- jev-chat-windows★ 802
Windows WeChat reply assistant: local offline OCR reads the chat, Jev judges intent and tension and ranks three drafted replies to fill in.
ProjectproductionPython Android chat co-pilot that reads the active chat via Accessibility and returns typed verdicts on intent, emotion and risk; never replies.
AppbetaKotlin