给 DeepSeek Harness 纯文本模型加视觉:采用codex风格聊天框拖图自动分流给用户配置的视觉模型转文字,v4 不切换模型即可识图。 Vision for text-only DSH models: routes chat-box images to a user-configured vision model and returns text descriptions, deepseek-v4-pro reads images without switching.
DSH plugin: text-only models (e.g. DeepSeek-V4) automatically see images via a vision model. Official surface-replace, cache-friendly, human transcript untouched. 纯文本模型自动识图桥
基于 DSH 官方侧边栏的文件视图增强插件,提供源码视图、原位编辑、行号显示、搜索和字体缩放功能。 A file-view enhancement plugin that extends the official DSH sidebar, providing source view, in-place editing, line number display, search, and font zoom.
Scene-aware vision routing layer for DeepSeek Harness (dsh): decides which engine/backend an image should go to (chat/UI/table/code) before other vision plugins route it. Switch-gated, front-loaded, never touches other plugins' tools. 场景级识图路由层。
Read-only market data for DeepSeek Harness. Ask about a stock, and let DSH call the data tools directly. | 为 DeepSeek Harness 提供只读金融市场数据。让 DSH 直接调用工具查询行情。
Plugin discoverability for DeepSeek Harness: search npm + awesome-dsh-plugin, look up packages, browse the curated list — from inside dsh agents (answers discussion #1715).
DSH web plugin (dsh-better-sidebar companion): export the current .md file to standalone HTML (Mermaid inlined) or PDF - export button on the markdown toolbar
LLM Wiki workbench for dsh-better-sidebar: browse and lint a Karpathy-style interlinked markdown wiki, open pages with the native markdown viewer, clickable [[wikilinks]] and provenance/relative .md links, and backlinks. Ships the llm-wiki agent skill.