dsh-memory: A Long-Term Workspace Memory Plugin for DeepSeek Harness
The dsh-memory plugin is designed to address memory management challenges in DeepSeek Harness by providing workspace-based long-term memory capabilities. Its core features include indexed briefings to reduce token consumption, relevant recall at the end of each turn, version-chain management, and automatic distillation. Users can perform operations such as saving and searching through the `memory` tool, and manage the memory store in the settings panel. Installation requires using a PowerShell script and ensuring configuration consistency. Data is stored locally by default, and a repository mirror mode can be enabled. The plugin is maintained by BangBang-03 and released under the MIT license.
Read Moredsh-client-ui-quote: DSH Passage Quoting and Chat Plugin
The DeepSeek Harness (DSH) client plugin dsh-client-ui-quote implements Kimi Code-style interactions, aiming to address scattered citations and fragmented context in long conversations. It allows users to select text in the main content, files, or terminal to generate citation chips, supports three modes—commenting, adding to the conversation, or creating a standalone side conversation—and provides code citations, quick navigation, and citation list management. Installation requires DSH >= 0.2.0-rc.2. The plugin runs in the renderer process and depends on a granular token that expires in 2027.
Read MoreFeishu Workspace Bridge: One-to-One Mapping Between DSH Workspaces and Feishu Groups
Feishu Workspace Bridge, a plugin for integrating DSH (DeepSeek Harness) with Feishu, addresses the limitations of traditional integration approaches, which require configuring public-network callbacks and handling large volumes of historical messages. By binding a Feishu bot, the plugin creates a private group on demand when a workspace receives its first message, enabling real-time, bidirectional message synchronization. Core features include Markdown support, on-demand group creation, queued message delivery, continued processing of failed messages after a process restart, and secure credential storage. Installation requires using the official command and configuring the corresponding permissions on the Feishu Open Platform. This solution eliminates the need for public-network callbacks while ensuring real-time and stable message synchronization.
Read Moredsh-work-badge: DeepSeek Harness Work Time Indicator
`dsh-work-badge` is a lightweight plugin designed specifically for DeepSeek Harness (DSH) to display in real time, in the top-left corner of the interface, whether the current period is a “peak” or “valley” (idle) time window. It is calculated locally based on Beijing time, requires no network connection, reports no data, supports both Chinese and English, and does not interfere with user operations. After installing it via the terminal, users can immediately start using it and customize the language and peak time window by editing the source code configuration file.
Read MoreIntroduction to the DeepSeek Harness Plugin: dsh-bafx (Blue Archive-Style Click Effect)
`dsh-bafx` is a Blue Archive–style mouse interaction effect plugin for DeepSeek Harness (DSH). It provides click effects, cursor trails, and ten real-time adjustable parameters (such as size, length, and blend mode) through a sidebar entry and a popup settings panel, without interfering with the main interface. The plugin is zero-build and zero-dependency, supports both Web and desktop platforms, and requires a client restart after installation to enable.
Read Moredsh-flymemory: Local Long-Term Memory for DeepSeek Harness
dsh-flymemory is a DeepSeek Harness plugin designed to address the challenges of limited context and difficult state persistence in local agent development. The plugin runs the FlyMemory service locally and provides 15 memory tools and two automatic hooks, supporting memory CRUD operations, state tracking, and conflict management. It ensures that data is stored only in the local file system and does not involve network transmission. Built with Python, the plugin relies on `torch` and `sentence-transformers` for vector retrieval. After installation, users must configure environment variables and download the model; initial startup is time-consuming, but the plugin supports offline operation.
Read Moredsh-godot-workflow: A Local Plugin for Godot Project Development Workflow
DSH's `AGENTS.md` has a 65536-byte limit, causing resident workflow rules to be easily discarded and rendered ineffective. The `dsh-godot-workflow` plugin, developed by AtlacherCoder, splits the Godot workflow into two parts: a lean resident core that does not count against the budget, and full skills loaded on demand. The plugin includes five strict rules, a workflow skeleton, and detailed specifications, and is installed via the `link:` protocol, but modifying its source code requires restarting `dsh web`. It also includes a self-check script to ensure the workflow remains available when the budget is tight.
Read Moredsh-tool-indent Plugin Overview: Adding Tool Call Indentation and Appearance Customization to DSH Web GUI
The DSH Harness UI plugin `dsh-tool-indent` improves the visual presentation of the DeepSeek Web GUI by using CSS to address the confusion between tool calls and body text. Its core features include unified left indentation for tool calls, adjusted gray tones and line spacing, and customizable appearance through settings or CSS variables. Settings are persisted in the browser locally, and nested calls are automatically indented level by level. The plugin only affects the main text view, not trajectory table views, making it suitable for users who prefer a clearer reading experience.
Read MoreDeepSeek Harness Plugin: dsh-ollama-cloud
`dsh-ollama-cloud` is a DeepSeek Harness plugin designed to simplify Ollama Cloud configuration. Once installed, it automatically registers as an inference provider, leveraging an Ollama-compatible interface to deliver one-click setup, self-updating model catalogs, and Web capability integration. The plugin supports per-model inference strength settings and provides UI-integrated API Key management. To avoid conflicts, do not use it alongside the `dsh-llm-ollama` plugin; API keys are securely stored via the credentials service.
Read Moredsh-workbuddy-websearch: Integrating WorkBuddy Cloud Search
The WebSearch plugin for DeepSeek Harness (DSH) reuses the logged-in session and cloud search capabilities of the WorkBuddy desktop client to provide DSH with zero-configuration online search. It supports returning summaries with cited sources, has no hard dependencies, and is ready to use after installation and a restart, without the need to apply for an API key. It is primarily intended for Windows and macOS developers, but supports only `web_search`, not `web_fetch`, and is subject to platform restrictions.
Read Moredsh-reedit: Rewrite and Send Your Own Message
The DSH session log is immutable. The `dsh-reedit` plugin allows users to double-click their own messages to rewrite them and send the rewritten version as a new branch. It intercepts the DOM to provide hover tooltips, double-click editing, and visual markers. The installation command is `dsh plugin --profile <profile> add github:Arbeiter-bit/dsh-reedit`. The plugin generates a light blue bubble tooltip, though the color may vary depending on the theme.
Read MoreDeepSeek Harness Plugin: dsh-plugin-preflight Preflight Tool
`dsh-plugin-preflight` is a zero-dependency DeepSeek Harness plugin designed to automate the compliance validation of DSH plugins. It checks the `dsh.bundle` configuration, `cordis.patch.yml` mounting, client exports, Peer Dependencies version compatibility, and manifest integrity, preventing plugins from being rejected due to configuration errors. Users can invoke it via the CLI or DSH tools, supporting local and CI environments. The tool has currently been tested only on Windows, and the npm release has not yet been verified.
Read MoreDeepSeek Harness (DSH) Web Search Plugin: Zero-Cost Replacement for Official Search
`dsh-web-search-ddg` is a zero-token web tool plugin for DeepSeek Harness, maintained by aooyoo. It replaces DSH's built-in high-cost search routes with a local headless browser and API, supporting Bing and DuckDuckGo. The plugin requires no token consumption or external dependencies and provides a zero-dependency design, fallback mechanisms, and compatibility, enabling one-click switching back to the official search. Installation must be performed in the DSH profile directory and requires configuring the search provider.
Read MoreDSH Custom Wallpaper Plugin
The DeepSeek Harness desktop version allows interface personalization through a custom wallpaper plugin. Developed by Ansellwan and available under the MIT license, the plugin lets users use any local image as a background and adjust opacity, blur, clarity, and darkening effects in real time. All data is stored locally, with no network requests. Users must install it via CLI commands and restart the application, then either drag an image into the settings panel or paste it with Ctrl+V to fine-tune the parameters. Note that major version updates may cause the wallpaper display to fail.
Read MoreDeepSeekHarness Session Delete: Adding Delete Functionality to Sidebar Sessions
The DeepSeek Harness sidebar lacks a session deletion feature, making management cumbersome. The dsh-session-clean plugin, developed by Ansellwan, solves this problem. It supports fully deleting sessions, including archived ones, through the sidebar menu. Active sessions are automatically grayed out, and deletion is blocked. The plugin does not access external networks and operates locally only, but deletions are irreversible. The installation command is `dsh plugin --profile desktop add dsh-session-clean`.
Read Moredsh-memory: Persistent Memory Plugin for DeepSeek Harness
The `dsh-memory` plugin for DeepSeek Harness (DSH) is designed to address the loss of context after an agent session restart. It implements persistent storage using pure Markdown files and supports user-level and project-level memory isolation, ensuring agent consistency across different sessions. Its core features include cross-session memory recall, a visual memory manager, and project memory isolation. When using it, note the ordering of the `--profile` parameter and restart the service to register the tool. The plugin has zero dependencies and ensures safety through atomic writes and boundary design, making it well suited for agent development scenarios requiring long-term context management.
Read Moredsh-trustedrouter: The TrustedRouter Plugin for DeepSeek Harness
The `dsh-trustedrouter` plugin resolves the issue where DeepSeek Harness cannot write TrustedRouter policies into the request body. It wraps the `fetch` method to enable per-model privacy and routing policies to be configured in dsh. Users can set privacy levels, preferred providers, and other options through the web interface or a configuration file. The plugin only supports dsh web, depends on pnpm, stores its configuration in `cordis.patch.yml`, and requires manual upgrades.
Read Morechangestab: A Change History View Plugin for DeepSeek Harness
The DeepSeek Harness plugin `changestab` provides a changes view in the web interface sidebar, integrating code diff viewing with file browsing to reduce context switching. It supports jj and git, offering change tree navigation and adaptive side-by-side or stacked diff display. The plugin is currently in read-only mode and open-sourced under the MIT license, allowing users to install it via the CLI and browse change history directly in DSH.
Read Moredsh-study-coach: DSH Study Coach Plugin
The DeepSeek Harness (DSH) plugin `dsh-study-coach` is designed specifically for self-study scenarios, supporting course breakdown, mastery tracking, and progress management. Developed by Ame-Choo, it provides a conversational tool and a web panel. Core features include a three-layer knowledge map, a six-level mastery assessment, a logging mechanism that requires evidence to be attached, and auxiliary tools such as Ebbinghaus memory cards. Users install it via npm and use it in “Study Coach” mode. Data is stored independently locally, and it is released under the MIT license.
Read Moredsh-sketch-reference: DeepSeek Harness Native Collaborative Visualization Plugin
The `dsh-sketch-reference` plugin integrates Excalidraw into the DeepSeek Harness workspace for visual collaboration. Users can create sketches and add annotations; the agent can read them on demand, identify relevant areas, and propose changes. Revisions are applied only after user confirmation. Installation requires DSH Web version 0.2.1-alpha and adding the plugin via the plugin manager. Data is stored locally, and automatic two-way binding is not supported.
Read Moredsh-inapp-browser: In-App Browser Panel for DeepSeek Harness
dsh-inapp-browser is a client-side plugin specifically designed for DeepSeek Harness, intended to provide AI with convenient web interaction capabilities. It uses the Chrome DevTools Protocol to render an independent browser instance in the Better Sidebar, supporting end-to-end operations such as tab management, navigation, clicking and typing, screenshots, and snapshot retrieval. It addresses the low efficiency and fragmented context of traditional approaches, while providing a rich toolset for multi-tab management, popup handling, dragging, and keyboard simulation. Before use, specific Node.js and plugin version requirements must be met, and the usage rules for coordinate systems and key combinations must be observed.
Read Morealkaid4521/dsh-pixel-art: Teaching an Agent to Hand-Code a Pixel Art Script
`alkaid4521/dsh-pixel-art` is a drawing plugin for DeepSeek Harness (DSH) centered on the core idea of “everything is a plugin.” It aims to solve the problem that fixed-vocabulary models cannot generate out-of-vocabulary content by teaching the Agent to write scripts that compute pixel colors one by one, enabling programmatic and reproducible pixel-art generation. The plugin supports scripted computation, style switching, and a self-evaluated delivery mechanism. Installation requires executing a command in the DSH directory and depends on Python + Pillow (preferred) or Node.js (fallback). It does not provide an asset library; it only provides criteria and tools, and runs entirely offline.
Read Moredsh-webui-m3e: DeepSeek Harness Material 3 Expressive Mobile Web UI
dsh-webui-m3e is a mobile Material 3 interface plugin for DeepSeek Harness, designed to address the inconvenience of using the native interface on smartphones. The plugin supports viewing AI interactions, tool execution results, and thought processes, and provides features such as session management, input enhancement, and utility tools. Users can install it via CLI, switch to it from the standard interface, or use it directly by accessing the `/m3e` path. Please note that this unofficial project only supports DSH 0.2.0-rc.2, and the interface supports Japanese only.
Read Moredsh-chrome-operation: Turning Local Chrome into a DSH Toolkit
`dsh-chrome-operation` is a client plugin for DeepSeek Harness (DSH) that connects to your local Chrome browser and converts browser operations into tools that DSH can recognize. It communicates with the installed `chrome-operation` extension over WebSocket. Installation requires two steps: first load the `chrome-operation` extension in `chrome://extensions`, then run `dsh plugin add dsh-chrome-operation` in your terminal and restart DSH. The plugin supports various browser operations, including snapshots, navigation, clicks, input, scrolling, tab management, and history. It only listens on the local loopback address.
Read Moredsh-memory-lite: Giving Your dsh Agent Cross-Session Long-Term Memory
The DSH plugin `alanzhao0128/dsh-memory-lite` provides dsh agents with zero-dependency, Markdown-based long-term memory across sessions. It persists preferences and decisions using a resident memory directory, five explicit memory tools (read, write, update, and forget), and a background implicit extraction mechanism. The plugin supports peer isolation and cross-peer sharing, with permission-based security controls and runtime status indicators. After installation, restart to use it; it is suitable for development scenarios requiring context retention and multi-session collaboration.
Read More