Integrations · DeepSeek Harness plugins — Page 72

dsh plugin: register codex-web-search-mcp (model-independent OpenAI Codex / Grok web search & deep research MCP server) as a native dsh MCP tool set (mcp__codex-web-search__*).

Jupyter notebook support for the dsh-better-sidebar editor: .ipynb opens inline as a runnable notebook viewer with IDEA/PyCharm-style run mode (Ctrl+Enter / Shift+Enter / run-all shortcuts, IDEA toolbar, gutter run icons, In/Out labels, execution duration, collapsible tracebacks), a lazy-start Python kernel (jupyter_client + ipykernel via python/bridge.py), background execution and save-back. Registers the ipynb file viewer through the ctx.betterSidebar service.

MCPHub 桥接插件（DSH Web GUI）：把本地 MCPHub（http://localhost:3000，@samanhappy/mcphub）聚合暴露的全部 MCP server 工具桥接进 DSH。宿主侧以 MCP Streamable HTTP 客户端连接 MCPHub 聚合端点 /mcp（或 /mcp/<group> 分组端点），把发现到的工具以 mcp__mcphub__<server>-<tool> 形式注册为 agent 原生工具；提供 mcphub_status / mcphub_resync 两个 agent 工具，/api/dsh-mcphub/status 环回路由与侧边栏「MCPHub」面板（连接状态、server/工具清单、一键打开 MCPHub Dashboard）。断线自动指数退避重连、工具集热同步（tools/list_changed），无需改 dsh 源码。

API capture visualizer for the dsh web GUI: a sidebar entry plus a live panel showing API-call records the agent captures from the client process (method / url / status / duration / headers / bodies / notes). Records are appended to a JSONL store via the api_capture_append agent tool or the /api/dsh-api-visualizer/ingest route; the panel polls and renders them. No dsh source changes.

Hang-evidence viewer for the DSH Web GUI: browse the ui-drive hang-loop evidence packs (frozen screenshot, summary timeline, process info, net-trace tail, probe and procdump logs, dump file info) inside the harness.

Postman-style HTTP client for the dsh web GUI: a sidebar entry plus a panel to compose method / url / headers / body and send. The request runs on the host (server-side, so no browser CORS), and status / headers / body / timing come back with pretty-printed JSON and a request history. Also exposes an http_request agent tool. No dsh source changes.

Loopback Codex delegation bridge for official DeepSeek Harness Desktop 0.2.0-rc.2

Stdio MCP server that delegates Codex tasks to official DeepSeek Harness Desktop

First-class DonSeTch web access for DeepSeek Harness: fetch, search and crawl with browser-grade stealth, keyless engines, BYOK providers and an auto-updating native binary.

Tailscale Serve authentication adapter for dsh Connection

Auth-capable drop-in replacement for the DeepSeek Harness web connection

Mobile companion layout for the dsh web GUI: phone-width drawer sidebar, overlay details pane, touch-sized composer and safe-area handling

FeyaGate 网关 × DSH —— 在 DSH 里托管本机 miloco-mcp-server，并提供服务/授权/账号的可视化设置界面

Eyes for text-only DeepSeek: view_image tool (any OpenAI-compatible VLM, local Ollama or cloud) + chat image-attachment bridge that rewrites pasted/dropped images into view_image path markers

MCP server studio for DeepSeek Harness: live-mounted servers with real connection status, a per-server tool browser, JSON import, and a dashboard-grade settings page.

尽可能的使dsh适配竖屏等移动端设备

DSH 插件榜单：每天整理 dsh-plugin 生态 Top 10 / Top 100，按热度（GitHub Star）或下载量（npm 周下载）降序排序，默认按热度；数据聚合自 GitHub dsh-plugin 主题、Oh-My-DSH 聚合目录、awesome-dsh 精选清单与 npm 下载统计，入口集成在设置页

远程站点与远程工作区（SSH/SFTP）：连接管理、统一添加工作区、远程文件读写

Unified SSH control center for DeepSeek Harness with stateless multi-host dispatch, direct SSH pipes, and zero-intrusive local development.

DeepSeek Harness vision bridge: session screenshots route by model capability. Image-capable models receive attachments inline; text-only models see the standard placeholder, and the bridge reads the image through a vision-language model — on demand via the vision_bridge_read tool or automatically via agent/pre-step context injection. 会话截图按模型能力自动分流：多模态模型原样直读；纯文本模型经 VLM 代读（tool/auto 双模式）。

DSH profile bundle: auto-bridge images for text-only models (DeepSeek etc.) — instead of the native "model does not support images" rejection, the image is materialized into the workspace and the model is instructed to run the configured vision tool (vision_glance, fuzzy-selectable) and answer from its analysis. Works on WSL/Ubuntu and native Windows; sessions on excluded presets (default minimal) keep the original blocking behavior.

Lightweight DSH upload bridge: images and common documents become workspace paths so a text-only DeepSeek model can read them with Qwen-MM-Plugins vision tools.

DeepSeek Harness-native integration for agent-vision-toolkit: image Q&A, OCR, grounding, UI restoration, pixel diff, Artifacts, and Web UI.

DeepSeek Harness-native integration for agent-vision-toolkit: image Q&A, OCR, grounding, UI restoration, pixel diff, Artifacts, and Web UI.