dsh-notify: DeepSeek Harness Web UI Session Completion Notification Plugin
`dsh-notify` is a web client plugin for DeepSeek Harness used to detect task completion in the background. When a session transitions from running to idle and either the current page is not visible or the session is not selected, it triggers a browser system notification. The plugin includes intelligent do-not-disturb behavior (triggering only when the page is in the background), per-session de-duplication, and appropriately timed permission prompts. It has no interface and takes effect immediately after installation, using a green “completed” dot in the sidebar as an in-app signal, together with system notifications to remind users.
Read MoreEnabling the DSH Agent to Think and Respond in Simplified Chinese: Introduction to the dsh-zh Plugin
`zoahdev/dsh-zh` is a Chinese language optimization plugin for DeepSeek Harness (DSH). It uses a registry mechanism to force the Agent to think and respond in Simplified Chinese, while strictly preserving the original form of technical details such as code, commands, and URLs, addressing translation issues in Chinese development scenarios. The plugin takes effect automatically after installation, provides default instructions to balance concise explanations with content fidelity, and supports customizing behavior through configuration options.
Read MoreIntroducing the DeepSeek Harness Plugin: dsh-poison-guard
To prevent supply chain poisoning, DSH Harness (DSH) provides a preinstalled scanner plugin, `dsh-poison-guard`. This tool statically analyzes plugin code and uses a three-layer mechanism to identify threats: AST analysis detects dangerous operations and data leakage; deobfuscation decodes common encodings; and regex-based heuristics match suspicious scripts. After installation, the `scan` command can be used to scan plugin directories, with exit codes supporting CI checks. Note that it performs static analysis only, may produce false positives, relies on Harness’s sandbox configuration to restrict permissions, and cannot defend against heavily obfuscated code.
Read Moredsh-answer-sound: Agent Response Sound Plugin for the DeepSeek Harness Web Interface
dsh-answer-sound is an open-source plugin for the DeepSeek Harness Web GUI that synthesizes or plays local audio in real time via the Web Audio API, providing auditory feedback for an agent’s response lifecycle. The plugin allows users to independently adjust the volume or replace the files for three types of sounds—start, complete, and error—while storing the configuration locally in the browser. After installation, restart the service and enable it in Settings to use it, helping developers monitor agent status more intuitively and improve the debugging experience.
Read Moredsh-ivory: A Calm, Warm-Neutral Interface for DeepSeek Harness
`dsh-ivory` is a browser-based visual plugin designed specifically for the DeepSeek Harness (DSH) web interface, intended to alleviate visual fatigue caused by cold gray interfaces. It offers a warm neutral color palette, light/dark modes, and a responsive layout, supports Markdown rendering and accessibility, and features complete zero telemetry with no network requests. Users can install and enable it from the command line, making it suitable for developers who value reading experience, privacy, and accessibility.
Read Moredsh-compact: DeepSeek Harness Context Compression and Overflow Recovery Plugin
`dsh-compact` is a context management plugin for DeepSeek Harness, designed to address context overflow caused by long conversations or frequent tool calls. It automatically compresses historical messages, provides manual compression controls, and includes an overflow recovery mechanism: when context usage reaches 75%, it automatically generates summaries to replace older content, and it attempts retries when an overflow error occurs. The plugin supports basic mounting and Agent-level mounting, offering flexible configuration strategies to effectively maintain coherence in long conversations.
Read Moredsh-ui-state: DeepSeek Harness UI State Persistence Plugin
The Web GUI of DeepSeek Harness loses its state after a restart. This plugin is designed to persist the interface state. It supports restoring the sidebar's expanded or collapsed state and width, recording the details panel's on/off state, and saving the conversation scroll position. The plugin adopts event-driven and batch-write mechanisms to optimize performance, and can be installed via CLI without configuration. It is compatible with DSH 0.1.x versions and provides robust fault tolerance and mobile responsiveness.
Read Morezink-ning-lkr/dsh-chat-width: DSH Web GUI Chat Area Width Adjustment Plugin
`dsh-chat-width` is a client-side plugin developed by zink-ning-lkr designed to address the limited width of the chat area in the DeepSeek Harness (DSH) Web GUI. It allows users to customize the overall chat area width (480–1680px) and the maximum width of user message bubbles (320–1200px) via a slider or a YAML configuration file, and supports resetting to default settings with one click. The plugin requires Node.js >= 18, but automatically falls back to default values in remote browser environments.
Read Moredsh-figma-to-lottie: A Figma/SVG to Lottie Compiler in DSH
`dsh-figma-to-lottie` is a DeepSeek Harness plugin that fills the gap in compiling design data into animation assets. It provides two tools, `lottie_compile_shape` (parses SVG paths) and `lottie_compile` (generates complete Lottie JSON), supports paths, rectangles, images, and text layers, as well as keyframe animations and gradient fills. The plugin requires no build step and can directly generate Lottie files for iOS/Android/Web through natural-language instructions.
Read Moredsh-vditor: Replacing the DSH Web Input Card with Vditor 3.11.3
DSH plugin `dsh-vditor` rebuilds the web chat input box based on Vditor 3.11.3 (IR mode), providing a Markdown WYSIWYG experience. It supports code highlighting, KaTeX formulas, dark themes, and dynamic placeholders. For file interaction, it supports path pills, image pasting, and `@` search references. The plugin also integrates toolbar features such as slash commands, model switching, permission control, and context usage monitoring. It requires DSH ≥ 0.1.2-rc.1 and depends on an internet connection to load Vditor.
Read Moredsh-kit: A Page Capability Suite Plugin for DeepSeek Harness
dsh-kit is a zero-dependency plugin package for DeepSeek Harness (DSH) that enhances interface capabilities through modular extensions. After installation and restart, it provides rich features such as a terminal, file tree, source code management, schedules, and a knowledge base, along with practical tools including a built-in browser, web search, session monitoring, and notifications. Users can enable features via toolbar toggles or shortcuts and configure parameters in the "Plugins" sidebar panel. The plugin is compatible with dsh ≥ 0.1.5, requires Node.js ≥ 22, maintains a zero-build architecture, and does not affect the existing experience.
Read MoreIntroducing the dsh-collapsible-models Plugin
`dsh-collapsible-models` is a DeepSeek Harness plugin designed to address the issue of excessively long model provider lists. It takes over the model selector by registering shadow entries and supports collapsing and expanding model groups. Users can interact by clicking provider names or using keyboard navigation, with accessibility and multi-language environments supported. Collapsed states are only retained for the current session. The plugin is open source and requires no build step.
Read Moredsh-deepseek-usage: DeepSeek Harness Real-Time Usage Monitoring
`dsh-deepseek-usage` is a DeepSeek Harness plugin for monitoring API account balances and token consumption. It supports multi-key balance aggregation, real-time token statistics, and cost estimation, and provides a floating Web UI panel to visualize data. The plugin supports low-balance blocking, off-peak pricing strategies, and historical chart analysis. Installation requires a compatible pnpm version, and the configuration supports overriding default thresholds and pricing tables. It stores keys locally only, making it secure and reliable, and is suitable for developers who need fine-grained cost control.
Read MoreDSH Basic Capabilities Panel
The "Basic Capability Panel" of the DeepSeek Harness Web plugin provides a visual configuration management interface. Designed with a modular architecture and maintained by developer `yxsj245`, it allows users to intuitively manage MCP servers (supporting hot reload and masked display), skills (supporting search, filtering, and priority management), global and project-level rules, and archived sessions (supporting restore and secure deletion) in the browser. Installation is performed through the official plugin channel; during development, run `pnpm run sync` to synchronize the code and restart the process. The plugin places strong emphasis on data security, masking all sensitive information and protecting active sessions.
Read Moredsh-env-profile: Automatically Detect Environment and Inject System Prompt
dsh-env-profile is a global memory plugin for DeepSeek Harness, designed to address the token consumption and workflow slowdown caused by repeatedly confirming local environment details (tools, paths, and network) in every new session. It automatically detects the local environment, consolidates the results into a global profile, and injects a low-overhead summary into the system prompt to achieve “instant readiness.” The plugin supports on-demand tool queries and leverages prefix caching to further reduce token costs. It is currently in the v0.1 stage and only supports Windows.
Read Moredsh-shortcut-creator: One-Click Windows Desktop Shortcut Creation from the DSH Settings Page
`dsh-shortcut-creator` is a client plugin for DSH, designed to simplify the local service startup process. It generates a desktop shortcut with a single click from the settings page and supports automatic browser launch, port checking, and taskbar pinning. The plugin uses PowerShell to create a local startup script, ensuring that the window opens only after the service is running. After installation, the web process must be restarted for the changes to take effect, and it currently supports only the Windows platform.
Read Moredsh-cost-estimate: Estimating Token Usage and Cost Before Answering
dsh-cost-estimate is a web plugin for DeepSeek Harness designed to provide upfront cost estimation and post-run billing reconciliation for large requests. It displays estimated input/output token ranges and costs before a response begins, then updates these values with actual data after the response completes. The plugin supports peak/off-peak pricing, threshold triggering, and cumulative cost tracking across multi-step tool calls. Its flexible configuration is well suited for developers who need fine-grained control over invocation costs.
Read Moredsh-webui-launcher: Web UI Launcher for DeepSeek Harness
`dsh-webui-launcher` is a plugin for DeepSeek Harness (DSH) designed to simplify launching and managing the Web UI. Through Slash commands, Model tools, and setting cards, it enables background control of the Web UI lifecycle, and supports automatically opening a browser and creating desktop shortcuts. It offers flexible configuration, defaults to binding to the local loopback address for security, and can run silently in headless environments.
Read Moredsh-kb-rag: A Local-First RAG Knowledge Base Toolkit
The DSH plugin `yumu247/dsh-kb-rag` provides a local-first RAG knowledge base toolset for DSH, designed for zero-cost, local deployment. The plugin wraps four core commands: `kb_query` for semantic retrieval, `kb_ingest` for incremental ingestion, `kb_crawl` for web scraping using Scrapling, and `kb_related` for graph-based related recommendations. Before installation, ensure the host is configured with a Python environment, an Ollama model (such as `bge-m3`), and the `kb-rag` pipeline. During installation, specify the `kbRagDir` path, which the plugin relies on to run. When in use, `kb_query` returns results in JSON format, and `kb_inge
Read Moredsh-vision-bridge: Enabling Text-Only Models to Receive Images
`dsh-vision-bridge` is a DeepSeek Harness plugin designed to address the inability of text-only models to process images. It leverages the `qwen2.5vl` model run locally via Ollama, converting images into text descriptions sent to the DeepSeek API to enable localized visual processing. The plugin provides a native image experience, transparent request-layer adaptation, structured output, and coordinate support. It includes nine vision tools, such as description, OCR, and region queries, and supports caching of duplicate images. Once installed, the plugin registers a “twin” route and injects a tool set, with all operations completed locally to ensure privacy and security.
Read Moredsh-statecore: Native Memory Plugin for DeepSeek Harness
`dsh-statecore` is a native memory plugin for DeepSeek Harness, designed to address the fragmentation and limitations of traditional MCP servers in session memory management. By mounting the StateCore memory engine into the DSH plugin context, it enables automatic memory ingestion, injection, and persistence. The plugin supports automatic identification of real conversations, summary compression protection, a web-based management interface, and native tool calling, while sharing storage with Claude Code. Built on local SQLite storage and released under the MIT License, it provides agents with an auditable, tightly integrated long-term memory solution.
Read Moredsh-ezprot-plugin: A Proteomics Analysis Plugin for DeepSeek Harness
dsh-ezprot-plugin is a plug-and-play plugin for DeepSeek Harness (DSH), designed to simplify complex proteomics analysis workflows. It encapsulates steps such as normalization, differential analysis, and enrichment analysis within a conversational interface, supports automatic environment management and Docker compatibility, and requires only data file paths from users to automatically generate analysis reports. This plugin lowers the barrier to entry and is suitable for users without programming skills to perform scientific analysis.
Read Moredsh-engine-switch: Route DSH Sessions to Pluggable Engines Based on Presets
`dsh-engine-switch` is a workflow plugin for DeepSeek Harness (DSH), designed to address the need for different AI engines (such as Claude Code) across different sessions within the same client. It does not provide concrete engine implementations; instead, it uses a proxy pattern to dynamically switch the engine driving top-level sessions based on presets, falling back to DeepSeek by default. Its core features include selecting an engine by preset, supporting instant switching at runtime, automatically loading presets, and automatically restoring the engine when a session is resumed. It decouples engine selection from session logic, while subagents always use the DeepSeek loop. After installation, it must be enabled and registered in the configuration, and it supports specifying engines via preset files or configuration mappings, requiring Node.js ^22.19.0 or higher.
Read MoreMing Capability Pack: Invoke DeepSeek Harness Native Capabilities with Natural Language in One Click
Ming Capability Pack is a thin adapter layer designed for DeepSeek Harness (DSH), addressing the problem that users cannot convert natural-language instructions into executable tasks. By using tools such as `ming_auto`, it enables one-click invocation of native capabilities, automatic assembly of missing plugins, and closed-loop management, ensuring tasks are concretely completed and evidence cards are generated. It supports CLI wrappers and Headless mode, with recursive protection and output validation mechanisms, allowing users to efficiently complete tasks such as file creation and statistics without needing to deal with underlying logic.
Read MoreDeepSeek Harness Plugin: dsh-plugin-novelai
`dsh-plugin-novelai` is a plugin for DeepSeek Harness that directly calls NovelAI V4.5 Full for image generation. It addresses the pain point of manually writing prompts by enabling automatic image generation from conversational descriptions, tag queries, and prompt inference. Users must configure a NovelAI token, which can be provided via chat, a configuration file, or environment variables. The plugin includes a built-in decoder for data processing, and the token is kept secure without being written to disk.
Read More