dsh-mood：DeepSeek Harness 心情系统插件
dsh-mood 是 DeepSeek Harness (DSH) 的一款心情系统插件。它拥有 33 种心情数值，通过自动判定和提示词注入影响模型语气与内心独白。该插件独立于人设插件工作，支持混合心情、每日日记、时事获取、语音分析及数据迁移。安装时需在 profile 目录使用 `github:` 路径，安装后必须重启 DSH，且在配置中可单独关闭心情功能。
Read Moredsh-partition-board：AI 认为需要跟人确认「怎么分区」时自主调用的画板插件
DSH理念下开发的`dsh-partition-board`插件，旨在解决AI在空间布局中无法精确界定边界的问题。该插件通过浏览器弹出的交互式画板，提供色卡标注、绘图工具及约束书写功能，支持用户直接参与分区设计，并具备缩放、平移及AI辅助补全能力。安装需进行目录联结及配置，状态存储于本地。插件通过消息式通信与宿主交互，用户确认后会将包含图例、几何及栅格矩阵的标准化载荷回传给AI，推动项目解读与进行。
Read Moredsh-mermaid-diagram: Mermaid Diagram Rendering Plugin for DeepSeek Harness
The `dsh-mermaid-diagram` plugin for DeepSeek Harness (DSH) provides models with a `mermaid` tool that can render flowcharts, sequence diagrams, and various other charts directly in the Web UI, addressing the difficulty of rendering code as plain text. The plugin supports streaming, adaptive light/dark themes, local execution, and fault-tolerant fallback. Installation requires DSH 0.2.1-alpha.1 and a Web runtime environment; when a model invokes the tool, it can generate interactive diagram cards in the conversation stream.
Read Moredsh-media-studio: Generate Images and Videos Directly in DeepSeek Harness Chat
DeepSeek Harness provides the dsh-media-studio plugin, seamlessly integrating image and video generation into the conversation flow. Users can generate media through background tasks or slash commands (such as `/image` and `/video`), with support for text-to-image/video, image-to-image, and local file sending. The plugin uses the Agnes model by default, supports other service providers as well, and includes secure credential management and automatic saving. It requires Node.js >= 22.19.0.
Read Moredsh-agnes-studio
The Agnes Creative Workstation plugin for DeepSeek Harness (DSH) integrates Agnes AI’s free large models for image and video generation, providing DSH with visual creation tools. Its core features include text-to-image and image-to-image generation, text-to-video, short-drama pipelines, digital human talking videos, and infinite-canvas workflow orchestration. Installation requires running npm commands inside the container and configuring ffmpeg, Chinese fonts, and the MIMO_API_KEY. Using a host-client separated architecture, the plugin protects API key security while enabling end-to-end content production from generation to synthesis, making it suitable for video production and short-drama development.
Read Moredsh-lark-claw: Turning DeepSeek Harness into a Personal AI Assistant on Lark
`dsh-lark-claw` is a plugin gateway connecting DeepSeek Harness with Lark, aiming to embed agent capabilities directly into instant messaging. It supports continuous conversations, file and image processing, progress display, and scheduled task management. Once developers configure the App ID, Secret, and required permissions, they can use agents within Lark to handle tasks, analyze files, or manage schedules. Released under the MIT license, the plugin addresses development challenges in integrating DSH into the Lark ecosystem.
Read MoreIntroducing the DeepSeek Harness Plugin: dsh-retro
dsh-retro is a plugin under the DeepSeek Harness ecosystem that addresses the fragmentation of session data. It converts DSH session events into retrospective drafts, which are manually reviewed and saved to an Obsidian vault. The plugin supports accumulating experience entries, generating weekly summaries, proposing rules, and integrating with blogs. Installation requires environment prerequisites such as Node.js 22, and configuration and management are handled through CLI commands.
Read MoreDeepSeek Harness Context-Aware Web Search Plugin
`dsh-context-aware-search` is a plugin for DeepSeek Harness (DSH) designed to address the issue of irrelevant general search results. It rewrites queries by reading conversation context, uses TF-IDF to re-rank results and mark their credibility, and supports one-click source summarization. The plugin uses DuckDuckGo by default and requires no API key, significantly improving search precision in specific technical scenarios, such as programming documentation retrieval.
Read MoreDSH Document Mode
`docs-mode` is a technical-documentation-specific agent preset for DeepSeek Harness (DSH), designed to address issues such as LLM hallucinations, inconsistent terminology, and excessive AI-like tone when large models write documentation. The mode includes a built-in workflow that prioritizes research before drafting and covers 11+ types of technical documents, such as development notes, API documentation, and test reports. Key capabilities include template-driven and autonomous learning behavior (automatically extracting a consensus-based skeleton), a standardized toolchain to reduce the “AI tone,” quality checks with quantitative metrics and cross-validation, and flexible delivery formats (Markdown, Word, and PDF). Users can either upload a template or directly issue writing instructions, with the system automatically matching a style card. This plugin is installed from the plugin marketplace, is available under the MIT License, and is suitable for developers who need to generate standardized technical documentation at scale.
Read Moredsh-git-sync: Git Configuration Sync Center for DeepSeek Harness
`dsh-git-sync` is a Git configuration synchronization plugin for DeepSeek Harness, designed to address multi-machine configuration synchronization and unified management of multiple tools. It converts DSH configurations into a Git repository, supports adaptation for various AI tools such as OpenCode and Cline, and provides features including real-time/scheduled synchronization, conflict self-healing, and secure credential management. After users create a private repository on GitHub and configure local Git, the plugin enables automatic synchronization and version control of configurations across multiple devices.
Read Morexby-todolist: A DSH Plugin for AI Agents That Provides External Working Memory and Task Management
xby-todolist is a “memory” plugin within the DeepSeek Harness ecosystem, designed to provide AI agents with external working memory and task management capabilities to address context loss in complex workflows. The plugin offers core features such as configuring API keys, reading to-do items, and writing task lists. Users can install it via the command line and follow a defined workflow using the `todo_read` and `todo_write` tools to track the progress of multi-step tasks. The plugin is available under the MIT license and can effectively improve the reliability of agent execution for complex tasks.
Read MoreThalamus: A WeChat Companion Plugin with Its Own Personality and Life
Thalamus is a WeChat bot plugin built on DeepSeek Harness, designed to turn a WeChat account into an AI companion with an independent personality, long-term memory, and emotional states. It is a non-commercial project featuring core capabilities such as a six-dimensional personality model, a World Engine, a memory filter, and a closed-loop commitment system. Installation requires the DSH environment, and configuration involves separately setting up the dialogue interface, the independent World Engine, the WeChat account runtime, and persona parameters. The plugin supports independent model calls, vector-based memory retrieval, and a web admin panel. Please ensure compliance before use, and it is recommended to bind a secondary WeChat account.
Read Moredsh-nga-reader: NGA Forum Reading and AI Integration Plugin for DeepSeek Harness
`dsh-nga-reader` is a plugin for DeepSeek Harness (DSH) that provides a Cordis-based persistent Host/Client bundle, aiming to enhance NGA forum browsing and integrate AI dialogue. It supports forum browsing, login management (QR code/password), reading in standalone Electron windows, and AI interaction features such as text explanation and summarization. It also implements MNGA's core features, including check-in, notifications, private messages, and favorites. Users can install it through GitHub, and it supports Agent invocation to retrieve forum content or perform text analysis.
Read MoreDeepSeek Harness Plugin: dsh-embedded-browser
`dsh-embedded-browser` (formerly `dsh-browser-panel`) is a plugin designed to address the challenge of agents being unable to handle graphical interface interactions, such as login and CAPTCHA, in headless environments. It embeds a browser within the DSH host and uses the `browser_embedded_*` toolset to enable AI-driven browser control while allowing humans to take over page interactions in real time through the Web UI sidebar. The plugin provides 12 lazily loaded tools that support navigation, clicking, text input, screenshots, device and theme emulation, and JavaScript execution. Installation requires Node.js >= 22.19, and it is recommended to use it with the `browser-use` skill to ensure persistent login credentials are shared across sessions.
Read Moredsh-memory: Adding Persistent AI Memory to DeepSeek Harness
`dsh-memory` is a static profile plugin for DeepSeek Harness that addresses the stateless session problem and gives AI persistent memory capabilities. Its core features include memory tab management, automatic naming, multi-AI privacy isolation, experience archiving, and AI name display in the sidebar. To install the plugin, place the repository in the designated directory and modify the configuration, then restart to activate the changes. The plugin supports writing different types of memories (such as global, shared, and experience) via the `memory_record` tool and injects guidance into the SystemPrompt to instruct AI on memory management, making it suitable for developers who need long-term context and privacy isolation.
Read MoreDeepSeek Harness Plugin: vikasranax/dsh-plugin-cats
`vikasranax/dsh-plugin-cats` is a plugin for DeepSeek Harness (DSH) that provides entertainment content for agents. The plugin relies on free public APIs and offers two tools: `get_cat_fact` (retrieve a cat fact) and `get_cat_image` (retrieve and display a cat image), with support for limiting the fact length via the `maxLength` parameter. Developers can enable the plugin locally using `git clone`, `pnpm install`, and `pnpm dsh web` commands. The project is licensed under the MIT License and is suitable as a reference for developers who want to integrate fun features into DSH agents.
Read Moredsh-ppt-master: An AI-Driven PPT Workflow Plugin for DeepSeek Harness
`dsh-ppt-master` is a native plugin for DeepSeek Harness (DSH), designed to generate, populate, and enhance presentations through the PPT Master skill. It supports exporting editable PPTX files and SVG previews, and provides template population capabilities. Users need to install dependencies in a Python 3.10+ environment, configure environment variables, and add the plugin to DSH to use it. This tool provides developers with a complete solution for presentation-related workflows.
Read Moredsh-nightshift: Night Off-Peak Scheduling Plugin for DeepSeek Harness
dsh-nightshift is a plugin for DeepSeek Harness designed to take advantage of the model's lower nighttime pricing. It allows users to queue tasks during peak hours and automatically release them for processing during off-peak hours, while generating save reports automatically. The plugin provides a sidebar panel to manage the queue and supports retry policies, persistent queues, and idempotent task handling. Users need to configure peak/off-peak windows and pricing; after installation, they can queue tasks via the 🌙 button to optimize costs.
Read MoreIntroducing the DSH with ChatGPT Plugin
`dsh-with-chatgpt` is a DSH plugin maintained by BeforeWave, designed to address issues such as the inability of web-based ChatGPT to directly access the local filesystem and the excessive token consumption when DSH reads large project contexts. By bridging the two, it allows ChatGPT to read and write local files directly, run commands, and seamlessly hand off complex tasks to a native DSH session. It supports ChatGPT in performing code reviews and minor edits, while routing tasks such as building and testing to local DSH execution, enabling a separation and coordination between understanding and execution capabilities.
Read Morezysftd/dsh-quote: Quoting AI Replies in DSH Web
The `dsh-quote` plugin aims to simplify quoting AI responses back into the input field in DSH Web. It provides three quoting methods through a UI card: text selection, quoting an entire message, and pasting large text (>300 characters). Implemented entirely on the client side, it requires no network requests and automatically inserts the quoted content at the beginning of the message in Markdown format. The installation command is `dsh plugin --profile web add -w git+https://github.com/zysftd/dsh-quote.git`, and the plugin is open-sourced under the MIT license.
Read Moredsh-pdf-reader: AI-Powered PDF Reader Plugin for DeepSeek Harness
dsh-pdf-reader is an external plugin for DeepSeek Harness that deeply integrates PDF reading with AI analysis to reduce fragmentation between reading and conversation. It supports native text, image region analysis, multi-PDF joint inquiries, and on-demand OCR. Core capabilities include annotations, an image/text asset basket, and source backlinks, while allowing reading state to be saved per conversation. It is compatible with modern browsers, stores data locally, and requires a vision model for image analysis. A beta version has been released.
Read Moredsh-seedance2: Seedance 2 Image and Video Generation on DeepSeek Harness
`dsh-seedance2` is a plugin for DeepSeek Harness (DSH) that integrates Seedance 2 AI's image and video generation capabilities. After installation, the environment variable `SEEDANCE2_PUBLIC_API_KEY` must be configured. The plugin supports image and video generation as well as task queries, and includes built-in retry, backoff, timeout, and idempotency handling. When using it, prompts should include elements such as subject, action, and scene, and parameters can be adjusted via a configuration file. It leverages the Harness credential system to secure keys, reducing the risk and cost associated with integrating external models.
Read MoreDeepSeek Harness Image Generation Plugin
The `dsh-image-gen` plugin, maintained by GooDAnDReaDY, is designed to deeply integrate image generation into DeepSeek Harness (DSH), replacing the need to switch to traditional external tools. Built on the FAL queue API, it supports multiple providers and allows the model to output images directly within conversations. Its core features include image variant tuning, smart aspect-ratio cropping, background removal, and style transfer. In addition, it offers project asset bundle export (e.g., PWA icons), a native sidebar gallery, and interactive diagnostic tools, providing front-end developers and designers with a one-stop solution for image processing and asset management.
Read Moredsh-plugin-clawrouters: A plugin package and a key to connect DeepSeek Harness with Chat, Images, Videos, and Search.
# Introduction When developing agents for DeepSeek Harness (hereinafter referred to as DSH), a common pain point is scattered capabilities: chat requires a provider, while image generation, video generation, and web search each require configuring keys and writing adapters. After ClawRouters consolidates these capabilities into an OpenAI-compatible API and a single key, `dsh-plugin-clawrouters` packages this key and set of capabilities into a DSH plugin package. Once installed and configured, it can be selected directly within DSH. The following introduces the plugin's positioning, features, installation, and configuration methods. # What is this? `dsh-plugin-clawrouters` is created by rop...
Read Moredsh-arc-context: DeepSeek Harness Reversible Context Compression Plugin
# Introduction For developers of agents running long sessions, a critical issue arises when the context approaches the window limit: compression is mandatory. The Basic compression built into dsh relies on an irreversible model summarization approach, which is lossy and irreversible. In scenarios such as code engineering that depend on precise facts, losing a single constraint can cause subsequent changes to diverge. dsh's philosophy is that "everything is a plugin," meaning compression strategies are replaceable. dsh-arc-context takes a different approach: before compression, it first stores the original text completely in a local append-only log, allowing compressed blocks within the context to be restored character-by-character at any time. Below is an introduction to its implementation, installation, and real-world test data. # What is it dsh-arc-context (ARC = Adaptive Revers...
Read Moredsh-agent-sync: Sync local AI agents' MCP servers and Skills to DeepSeek Harness
`dsh-agent-sync` is an MIT-licensed plugin for DeepSeek Harness (DSH) designed to eliminate redundancy for pre-configured MCP servers and Skills from other local AI agents within DSH. Supporting over 20 AI agents, its core functions include automatically scanning local configurations, writing MCP servers to DSH profiles with a single click, and copying or linking Skills to global or workspace directories. The plugin provides comprehensive MCP and Skills management capabilities, such as enabling/disabling, deleting, grouping, migrating, and manual addition (supporting multiple file formats). Through its settings, the plugin allows users to configure the synchronization method (copy/link) and target profiles.
Read Moredsh-llm-agy: An AGY Agent Enhancement Engine for DeepSeek Harness
flg1217/dsh-llm-agy is an MIT-licensed plugin for DSH designed to extend DSH's capabilities by integrating AGY (Antigravity CLI), enabling seamless upgrades without modifying the dsh source code. The plugin requires Node.js ≥ 20 and dsh >= 0.1.0-rc.6, with the AGY CLI pre-installed. Its core features include: 1. Independent reasoning sub-agents, allowing specific tasks (such as frontend/UI) to be delegated to AGY by setting `agentOptions.provider: 'agy'`, with a `list_agy_models` tool for querying available models; 2. Deep web search, after configuring `searchProvider: agy`, `web
Read Moredsh-model-router: A Thin Routing Strategy Plugin for DeepSeek Harness (DSH)
dsh-model-router is a thin routing policy plugin for DeepSeek Harness (DSH), version v0.2.2, licensed under the MIT License. It adds rule execution on top of DSH's existing model management, with core features including deterministic keyword routing, the whitelist-based model switching tool `model_route` that is disabled by default, and an isolated Beta image generation channel `image_gen`. Routing rules are executed in configuration order, stopping at the first match, while preserving the current state if no rule matches; model switching requires precise validation of provider, model, and other fields. `image_gen` uses an OpenAI-compatible endpoint, enforces HTTPS and strictly restricts result downloads, and supports integration with third-party services such as Seedream; however, video, audio, and 3D features are not included.
Read More