dsh-plugins/dsh-auxiliary 预览 preview

dsh-plugins/dsh-auxiliary

DeepSeek Harness的辅助模型:通过专用模型路由实现视觉理解与上下文压缩。

Project Overview项目介绍

This is a native plugin built exclusively for DeepSeek Harness that adds auxiliary model routing capabilities without altering or replacing the main conversation model. Each feature acts as an independent, optional route that only activates for its specific task category, letting users assign specialized or lower-cost models to niche tasks like vision processing, context compaction, approval reviews, and subagent delegation. It ships with a dedicated settings panel in the DSH web UI that lets users configure all routes in place, with changes taking effect immediately after saving without requiring a restart or conversation rebuild.

To install the plugin, users simply copy the provided installation block and paste it into their DSH assistant, which handles all installation and verification steps automatically. No manual npm installation or profile configuration editing is required, and the plugin manager automatically adds the bundle to the user's profile, so no manual edits to config files like cordis.patch.yml are needed. After installation completes, users just need to open the Auxiliary Models settings page and mark which models they want to use for each route before they can start using the new capabilities.

The plugin is released under the open-source LGPL-3.0 license, and the development workflow includes standard npm steps for installing dependencies, running type checks, executing regression tests, and building the output library. One notable caveat for users is that if you enable the optional compaction engine replacement feature, you should not load the @deepseek-ai/dsh-compaction-basic plugin at the same time. The plugin automatically detects this configuration conflict, skips loading the conflicting engine, and emits a clear warning to let the user know about the issue.

这是一个专为DeepSeek Harness开发的原生插件,用于在不替换主对话模型的前提下,为各类专项任务分配专用的辅助模型。它提供独立的专用模型路由功能,涵盖视觉理解、上下文压缩、审批审查、子代理调用、会话标题生成和图像生成等多个常用任务,每个功能都是独立可选的路由模块,仅在对应任务调用时触发。

安装时只需将指定安装指令粘贴到DSH助手,即可自动完成安装和验证,无需手动编辑npm或配置文件。安装完成后,所有路由功能都可在DSH的「设置→辅助模型」面板中配置,保存后立即生效,无需重启或重建当前对话,用户可根据自身模型的成本和能力特性灵活分配不同任务对应的模型。

插件以LGPL-3.0协议开源,开发依赖TypeScript和DeepSeek AI官方DSH相关包,提供完整的类型检查、测试和构建脚本。使用时需注意,如果启用了可选的上下文压缩后端,不要同时加载@deepseek-ai/dsh-compaction-basic,插件会检测冲突并跳过加载,同时发出警告提示。

Pre-install check安装前体检Compatibility · Security兼容性 · 安全性 1 note1 项提示
  • 12 stars - an early-stage project星标 12,属于早期项目
DSH walks through these 9 checksDSH 会逐条核对这 9 项

Compatibility兼容性

  • DSH, Node, OS and profile requirementsDSH 版本 / Node 版本 / 操作系统 / profile 是否满足要求
  • External dependencies and runtimes (Electron / Python / Docker, ...)外部依赖与运行时(Electron / Python / Docker 等)是否齐备
  • Conflicts with installed plugins: command names, skill / tool names, ports, duplicate MCP registration与已装插件是否冲突:命令名、skill / tool 重名、端口占用、重复 MCP 注册

Security安全性

  • Repo matches the facts registered here; archived or abandoned?仓库是否与页面登记一致,是否归档或长期停更
  • Safety of preinstall / install / postinstall and install.sh / setup.ps1preinstall / install / postinstall 与 install.sh、setup.ps1 是否安全
  • curl|bash, download-then-execute, obfuscation, unrelated domains → stop immediatelycurl|bash、下载即执行、混淆代码、无关域名 → 立刻停止
  • Typosquatting or unmaintained packages among the new dependencies新增依赖里有没有 typosquatting 或无人维护的包
  • Requested permissions vs. what the feature actually needs申请了哪些权限、是否超出功能所需(filesystem / network / shell / clipboard)
  • Any sudo / admin requirement, plus uninstall and rollback是否要求 sudo / 管理员权限,以及卸载与回滚方式

Anything uncertain must be marked unknown with a note on how to confirm it. This site's signal screen is a static snapshot, not a security audit.拿不准的必须标「未知」并说明要我怎么确认。本站的信号筛查是静态快照,不能替代安全审计。

Or use CLI install (for developers)或使用命令行安装(适合开发者)

CLI Install命令行安装

dsh plugin --profile web add github:dsh-plugins/dsh-auxiliary

把 dsh-plugins/dsh-auxiliary 加入你的 DSH 配置(web profile)即可启用。

READMEREADME

Banner

dsh-auxiliary

Auxiliary models for DeepSeek Harness — dedicated model routes, tools, and system guidance for vision, compaction, reviews, subagents, titles, and image generation, without touching the main conversation model.

English | 简体中文

DSH Plugin Build Status Version License

dsh-auxiliary is a DeepSeek Harness plugin that layers auxiliary-model capabilities on the harness LLM seam (ctx.llm). It never replaces the main conversation model: each feature is an independent, optional route that kicks in only for its own narrow call category, so you can give expensive or specialized work (vision, compaction summaries, approval reviews, delegated subagents, session titles, image generation) its own cheap or capable model.

Feature overview

Feature What it does Where to configure
Vision understanding inspect_image tool: read a local image and ask a vision model about it Models page → Auxiliary Models → Vision understanding
Image handoff Chat images survive a text-only main model via describe_image Vision understanding card → Image handoff
Context compaction Summaries (purpose: 'compaction') use a dedicated model Auxiliary Models → Context compaction
Compression engine (optional) Replaces the compaction backend with an explicit compression prompt engine.enabled in the config
Approval model Reviews from @dsh-plugin/dsh-approve-for-me use a dedicated model Auxiliary Models → Approval model
Subagent model Delegated child agents use a dedicated model Auxiliary Models → Subagent model
Title model Session titles (purpose: 'session-title') use a dedicated model Auxiliary Models → Title model
Image-generation model generate_image tool: create images through an OpenAI-compatible images API Auxiliary Models → Image-generation model
Model capability marks Checkboxes on your own llm-pi-ai models: Allow image input / Allow image generation Models page → provider → customized settings → model

Showing the opening section of the README — the full document lives in the repository以上为 README 开头摘要,完整文档在仓库内 · View the full README on GitHub →在 GitHub 查看完整 README →

← 上一个 Prev dsh-annotate 下一个 Next dsh-opencode-free-models →