dsh-foxbell-pet: DSH Desktop Bottom-Right Multi-Pet Status Monitoring and Productivity Dashboard
`dsh-foxbell-pet` is a visual desktop pet plugin for DeepSeek Harness (DSH), designed to address multi-project status monitoring and efficiency tracking through an intuitive interface. Its core features include: multi-Agent status cards (red/yellow/green/dark red) for session switching and progress tracking; hot-swappable pets with local/online import and integrity verification; and a timeline-based efficiency display alongside a three-tier Token usage dashboard. The plugin offers natural drag interactions, voice feedback (announcing completion, approval, and errors), and extensive right-click customization options. Users must install it using version v2.x and a DSH environment of 0.1.2-rc.1 or later.
Read Moredsh-mindmap: Transform Markdown Files into a Real-Time Mind Map
The dsh-mindmap plugin deeply integrates Markdown files with mind maps, supporting “file as map” and “chat as editing.” It uses mindmap_* tools to guide the AI in making step-by-step changes to the file, with real-time rendering in the right-side panel. Core features include synchronization between the file and the root node, MarkGrove-style node mapping, canvas interactions (zoom, search, and collapse), draft protection, and safe export as PNG or Markdown. It runs in the DSH environment and uses an optional sidebar, making it suitable for converting outlines into visual mind maps and for AI-assisted refinement of structure. It is currently in the pre-1.0 stage.
Read Moredsh-workspace-presets
dsh-workspace-presets is a plugin for DeepSeek Harness (DSH) designed to simplify agent preset configuration across multiple workspaces. It allows users to bind a specific preset to a workspace so that new sessions automatically apply that preset. The plugin manages bindings through the native settings page, provides a fallback mechanism, and does not affect already started sessions. It can be installed from npm, GitHub, or locally; note that local installation requires running `npm install` first, and it only takes effect in blank sessions.
Read Moredsh-cmdwatch: Monitor dsh Command Execution Progress in Real Time
`dsh-cmdwatch` is a plugin for DeepSeek Shell (DSH) designed to address the need for viewing command execution progress during development and debugging. It uses a real-time streaming mechanism to scroll-display the execution output of foreground commands and background tasks in a panel. The plugin supports auto-scrolling, abbreviating long commands, session isolation, and command sanitization (such as injecting `Tee` and `PYTHONUNBUFFERED`). After installation, the web profile must be restarted, and parameters can be adjusted in the configuration file. Note that enabling real-time streaming may interfere with reading by the `job_output` tool, and logs are stored only in memory.
Read MoreDeepSeek Harness Plugin: coolgech/dsh-siyuan
`coolgech/dsh-siyuan` is a SiYuan Note plugin designed specifically for DeepSeek Harness, aiming to enable reading, writing, and management of local knowledge bases through a unified interface. Built on the DSH protocol with zero dependencies and instant configuration, it offers 30+ tools, including document CRUD, full-text search, SQL queries, backlink analysis, and batch import/export. Users only need to install SiYuan Note and enable the kernel HTTP API to invoke it securely within an agent, making it well suited for building private, self-hosted agents.
Read MoreDeepSeek Harness Plugin: dsh-evidence (Enabling Models to Actually Read Files)
dsh-evidence is a local file parsing plugin for DeepSeek Harness. It parses and indexes files locally to transform them into structured, versioned evidence, rather than feeding them directly into context. The plugin supports multiple formats and provides two tools, `search_documents` and `read_document`, enabling models to retrieve content precisely. Installation requires Node.js 22.13.0+ and Harness 0.1.7+, and it is currently in beta.
Read More01men/rq-dsh: Rongqi Custom Track (Gate 01) — AI Agent and Human Collaboration Frontend
Rongqi Custom Rail (01men) is a DeepSeek Harness (DSH) plugin maintained by 01men, designed to build a unified enterprise-grade AI resource management and collaboration front end. The plugin integrates core capabilities such as IAM authentication, DingTalk login, MCP deployment, a skill marketplace, and a model gateway. It enables inter-component collaboration through an event bus and supports a zero cross-origin browser proxy mode. The platform is open source under the MIT License, supports offline deployment on intranets, has removed wallet and revenue-sharing features, and focuses on team and asset pricing, making it suitable for building a private AI governance system.
Read MoreDeepSeek Harness Plugin: dsh-postmortem Post-Mortem Review
dsh-postmortem is a local-first, read-only plugin for DeepSeek Harness that quickly analyzes Agent task failures and provides sanitized recovery plans. It intelligently detects faults, distinguishes terminal states from causal evidence, and generates structured postmortem recommendations without relying on a model. It supports viewing failure details, retrieving next-action commands, creating new recovery sessions, and exporting JSON recommendations. The plugin is compatible with DSH 0.1.1-rc.2, does not store keys by default, and can enable an optional model-based postmortem layer through configuration.
Read MoreDeepSeek Harness Plugin: dsh-pet-panel (Desktop Pet and Conversation Overview)
dsh-pet-panel is a self-contained client plugin specifically designed for the DeepSeek Harness (DSH) Web UI and developed and maintained by zw11591-sketch. It has no dependencies on external services and aims to provide developers with intuitive interface feedback and context management support through desktop pet visuals and a conversation overview panel. Users can run it locally and enable it via plugin configuration. Before installation, ensure the source is trustworthy and review the license information. For the plugin's specific interactions, refer to the source code implementation.
Read Moredsh-agentenv-sandbox: AgentENV E2B Sandbox
The DeepSeek Harness plugin `dsh-agentenv-sandbox` uses an E2B adapter and the AgentENV Firecracker microVM backend to migrate an Agent's local filesystem and subprocess operations to a remote isolated environment, ensuring secure execution. It supports Bash, file tools, and LSP processes, and adopts a fail-closed policy to guarantee isolation without falling back to the host machine. Deployment requires a Node.js 22.19+ environment, a prebuilt AgentENV template, and specific environment variables. This solution addresses security risks in Agent development and enables secure, persistent sandboxed operations.
Read MoreDeepSeek Harness (DSH) Real-Time Output Repetition Guard
`dsh-dupguard` is a DeepSeek Harness (DSH) plugin that detects and halts consecutive repeated output in real time during the model inference stream. By default, it triggers an interruption after 10 or more consecutive repetitions, with relaxed handling for code blocks and similar contexts. It allows configuring thresholds, allowlists, and global notifications via the settings page. Installation requires DSH ≥ 0.1.2-rc.1 and Node.js ≥ 20, and DSH must be restarted for the plugin to take effect.
Read MoreDeepSeek Harness LAN Access Plugin
DSH's Web profile allows local access only by default. The `dsh-lan-access` plugin uses a Bundle Patch to override the configuration, resolving API errors and permission issues when accessing via a LAN IP through a mobile browser. **Core features** include: overriding `webserver.host` to `0.0.0.0` to listen on all network interfaces; injecting a `randomUUID` shim for non-secure context compatibility; automatically adding the LAN IPv4 address to the `/api` trusted list. **Installation** requires running `dsh plugin --profile web add ...` in the `web` profile, then restarting the service after installation. To **use** it, access it via `http://<LAN IP`
Read Moredsh-agnes-media: Agnes AI Image and Video Generation Plugin
The `dsh-agnes-media` plugin integrates Agnes's media generation capabilities into DeepSeek Harness (DSH) through the `agnes_image` and `agnes_video` tools. It supports text-to-image, image-to-image, text-to-video, and keyframe generation, with generated files saved by default to the local workspace. Configuration requires managing the API key via environment variables and setting the model name. It provides robust error handling (such as retry mechanisms) and asynchronous video task polling, addressing DSH's inability to handle non-conversational media requests and offering reliable support for multimodal agents.
Read Moredsh-terminal-plugin: A VS Code-style Terminal Panel for DeepSeek Harness
`dsh-terminal-plugin` is a plugin for the DeepSeek Harness (DSH) Web GUI, providing developers with a VS Code-style terminal panel at the bottom. It supports multiple tabs, a real PTY, color output, and interactive programs (such as vim and python), preserving sessions after a page refresh. It supports click-to-navigate for URLs/paths, right-click menu actions, and configurable parameters such as shell and size. The plugin runs in the DSH host process with system permissions, but has concurrent input limitations.
Read MoreDSH Desktop Operator: Windows Desktop Automation and Security Computer Use Plugin
`dsh-desktop-operator` is a Windows desktop automation plugin designed specifically for DeepSeek Harness. It packages an adapted Open Computer Use native runtime to support operations such as clicks, text input, and window management, and it exposes an MCP interface. The plugin prioritizes accessibility elements and supports confirmation for high-risk operations to ensure safety. Users must install the `.tgz` package using PowerShell and configure permission policies in the Agent Preset (for example, `highRiskActionPolicy: confirm`). It currently supports Windows only and depends on specific versions of Harness and DSH.
Read Moredsh-call-shrink: DeepSeek Harness Call Compression Plugin
`dsh-call-shrink` is DeepSeek Harness's ultimate call compression plugin, designed to merge N LLM calls into a single call. It significantly reduces costs through caching, deduplication, and batch consolidation within a 10-second window. The plugin features stable fingerprinting, protocolized merging, a failure circuit breaker, and a real-time monitoring dashboard. It is easy to install, supports RPC operations, and automatically falls back when parsing fails without affecting the final output.
Read MoreDeepSeek Harness Contract Drafting Agent Plugin
DeepSeek Harness’s Professional Contract Drafting Agent mode is designed to address standardization and multi-dimensional review challenges in contract drafting. The plugin includes an 11-stage lawyer workflow, five parallel AI review passes, and decision gating, and supports general contract, employment, and equity investment scenarios. Users must obtain installation instructions through official channels, run it with the current process permissions, and verify the license before use.
Read MoreDSH Launcher: A macOS Menu Bar Development Server Management Tool
DSH Launcher is a macOS menu bar tool for DeepSeek Harness, designed to simplify the cumbersome terminal operations required for local development servers. The tool encapsulates functions such as starting, stopping, restarting, and log extraction into a native application, enabling one-click management via the menu bar icon. It includes a companion DSH plugin that automatically retrieves the port, PID, and tokenized URL. Users need to install the plugin and application on macOS 13 or later, and customize the working directory, commands, and browser through a configuration file. By remaining resident in the background and automatically polling service status, it significantly improves development efficiency.
Read Moredsh-livedocs: DeepSeek Harness Version-Pinned Live Library Docs
`dsh-livedocs` is an open-source plugin for DeepSeek Harness (DSH) designed to address “hallucinated API” issues caused by models relying on outdated data. It ensures code accuracy by injecting up-to-date, authoritative official documentation context before code generation. The plugin supports version pinning and automatic triggering, providing three usage modes: embedded skills, slash commands, and global automatic rules. Core features include mapping library names to documentation sources, retrieving version-specific snippets, token budget control, and local caching. Installation requires prerequisites such as DSH >=0.1.5-rc.1 and Node.js >=18, and it can effectively reduce model misuse of framework APIs.
Read Moredsh-code-ui: A Cursor-Style AI Code Editor Workbench Plugin
`dsh-code-ui` is a client plugin for DeepSeek Harness (DSH), designed to provide developers with a workbench for viewing code and conversing with AI simultaneously. The plugin integrates a file tree, multiple tabs, syntax highlighting, and inline reference/translation features, and supports both a right-side docked mode and a floating window mode. Users can select code to reference, translate, or annotate it, which the AI will read once sent. The plugin is hidden by default and must be invoked via a top button; all data will be cleared after a restart.
Read Moredsh-auto-vision: Visual Bridging for Models Without Image Input Support
The DeepSeek Harness (DSH) plugin `dsh-auto-vision` provides visual capabilities to models that do not support image input (such as `deepseek-v4-pro`). It intercepts images, uses a configured vision model to recognize them and generate text descriptions, which are then written into the history as collapsed context to enable persistent memory. The plugin features automatic completion of input declarations, request-level stripping, and recursive recognition of tool results. It can be enabled through configuration alone, without modifying source code, allowing non-vision models to handle screenshots.
Read Moredsh-token-optimizer: A Deterministic Compression and Recovery Tool for DeepSeek Harness
`dsh-token-optimizer` is a plugin for DeepSeek Harness (DSH) designed to address context window overflow caused by tools returning long text in agent development. It provides deterministic, recoverable tool-result compression with cache-aware features, and supports PTC log compression and a web dashboard. Installation requires DSH 0.2.0-rc.2, with a 62.5% compression ratio enabled by default. Users can configure the compression threshold or disable early compression, and the plugin is licensed under the MIT License.
Read Moredsh-flyout-sidebar
The DeepSeek Harness (DSH) client plugin dsh-flyout-sidebar provides a draggable, pop-out sidebar that integrates a file tree, Git changes list, and code preview. The plugin supports dual-view switching and an independent tab mode, enabling real-time synchronization with the session area. The installation command is `dsh plugin --profile web add dsh-flyout-sidebar`; it can be used after restarting. Note that the workspace must be a Git repository, and local builds require installing dependencies.
Read More@ottttto/dsh-scheduled-send: Server-Side Scheduled Send Plugin for DSH Web
`@ottttto/dsh-scheduled-send` is a scheduled messaging plugin that runs on the DeepSeek Harness (DSH) host side. It moves scheduling logic from the browser to the server, preventing task interruption when the browser is closed. The plugin supports adding a scheduling button to the input box, managing tasks in the sidebar, session isolation, resending after power failures, and mobile compatibility, making it suitable for scenarios such as simulating delayed conversations.
Read More@openllmsh/dsh: Routing the DeepSeek Harness Through the OpenLLM Gateway
The DeepSeek Harness plugin `@openllmsh/dsh` is a pure configuration-layer plugin designed to route traffic to a local OpenLLM gateway. It is implemented through `cordis.patch.yml` and requires no runtime code. Core features include: 1. Adding an OpenLLM default model provider to the Harness, pointing to the local gateway; 2. Registering the OpenLLM MCP server. Before use, OpenLLM must be installed and the `openllmd` daemon must be running, and the host must meet the Node.js version requirements.
Read More