dsh-model-collapse: Group and Collapse the Model Selector by Provider
The DSH plugin `dsh-model-collapse` groups and collapses the model selector by provider to reduce the space taken up by long lists. It includes a top quick bar and keyboard shortcuts (such as Alt+E to expand and Alt+F to focus), and supports remembering expanded states and filtering. The plugin is purely front-end, does not modify the official source code, and is released under the MIT license. The repository has been archived; see the official collection for details.
Read MoreDeepSeek Harness Plugin: dsh-updater
The dsh-updater plugin is designed to streamline the maintenance workflow for plugins and integration repositories in the DeepSeek Harness (DSH) ecosystem by providing an automated, secure, and low-risk update mechanism. It supports two forms: git checkout and global npm. Core features include status detection, background manual updates, task cancellation, and automatic update application during idle periods. The plugin features zero privilege and zero dependencies, uses a fast-forward strategy by default, and automatically rolls back on failure to ensure system stability and security. Developers can manage updates easily through the command line or the web panel.
Read Moredsh-capability-facade: Providing a Narrow Tool Surface for DSH Plugin Authors
`dsh-capability-facade` is a reference plugin implementation for DeepSeek Harness (DSH), aimed at addressing the problem of overly broad plugin tool surfaces. It allows plugin authors to declare semantic capabilities and compose multiple underlying tools into a deterministic execution pipeline. A single model call can complete tasks that would otherwise require multiple round trips, reducing cost and improving determinism. The plugin requires underlying tools to be registered before the facade and does not hide already exposed tools. It is primarily used to stably combine multiple underlying tools into a high-level semantic operation.
Read Moredsh-model-jury: Structured Cross-Model Peer Review for DeepSeek Harness
`dsh-model-jury` is a plugin for DeepSeek Harness designed to resolve inconsistent outputs and the lack of objective review in multi-model collaboration through a structured cross-model peer review mechanism. Featuring blind review, anonymous peer evaluation, and a revision stage, it uses code logic—rather than a specific model acting as judge—to make deterministic rulings, thereby enabling decentralized review. Users can configure providers such as GLM and DeepSeek via a configuration file and use the `/jury` command to initiate reviews in either standard or adversarial styles, making it suitable for scenarios such as cross-validating technical solutions.
Read Moredsh-mermaid-smooth: DeepSeek Harness Renders Mermaid Fenced Blocks by Default
`dsh-mermaid-smooth` is a plugin for the DeepSeek Harness Web UI, maintained by user gitByteFree. It renders Mermaid code fences in conversations as SVG diagrams by default, without relying on external CDNs. The plugin supports smooth zooming, panning, and double-click-to-fit-view, with independent “diagram/text” toggling, state memory, theme following, and security isolation. Users can install it via npm or GitHub; after installation, restarting the service makes it available, providing a smooth chart interaction experience.
Read MoreDeepSeek Harness Background Image Plugin (dsh-bg-changer)
The `dsh-bg-changer` plugin enables DeepSeek Harness (DSH) users to customize the UI background. By adding configuration entries in the settings, it supports uploading local images and adjusting layout, opacity, blur, and the UI’s own transparency. The plugin has zero dependencies and requires no build, stores data locally, supports multiple image formats and preview adjustments, and aims to provide a low-intrusive skinning solution. Installation and uninstallation are handled via scripts, and it has been verified in the DSH v0.1.6-alpha.1 environment.
Read Moredsh-plugin-tui: A Claude Code-Style Terminal UI for DeepSeek Harness
`dsh-plugin-tui` is a client-side plugin for DeepSeek Harness, designed to provide a persistent, interactive terminal environment similar to Claude Code. The plugin supports streaming output, session management, model switching, and permission approval, addressing the limitation of traditional single-turn plugins in enabling multi-step collaboration. Users must install it by specifying the `tui` configuration file; the plugin requires only two bundle layers and does not support `dsh-headless`. Its core features include internal REPL commands (such as `/sessions` and `/model`) and real-time status feedback, though known considerations include one-time approval limitations and the need for token rendering optimization.
Read Moresage-guikit: DSH Desktop Control Toolkit
`sage-guikit` is a Windows desktop control toolkit designed specifically for DeepSeek Harness (DSH), aiming to provide an Agent with the abilities to "see" and "act." This plugin packages Windows desktop control capabilities into a profile bundle, offering screen monitoring, screenshots, mouse/keyboard interaction, window management, and structured querying and validation based on UI Automation. To install it, add the plugin as a bundle in `package.json` and restart the service. The plugin defines three strict spatial contracts: virtual desktop physical coordinates, image pixel coordinates, and window-local coordinates, and provides 13 model tools, including click, typing, key presses, dragging, and semantic validation. It supports Windows only.
Read Moredsh-subagent-delegation: DSH Subagent Delegation and Acceptance Plugin
`dsh-subagent-delegation` is a DeepSeek Harness (DSH) plugin designed to address the clutter and resource waste that can occur when a main Agent handles complex tasks. The plugin uses built-in decision logic to determine when to delegate tasks, such as multi-file work, multiple searches, or independent items, and supports parallel execution and background running of sub-Agents. It also emphasizes verifying results by checking actual file changes and chains of evidence. Developers need to craft task-specific prompts to define sub-Agent behavior, while the plugin integrates through the `cordis.patch.yml` mechanism, effectively improving workflow efficiency and controllability.
Read Moredsh-skill-authoring: DSH Skill Authoring and Validation Plugin
The DSH plugin `dsh-skill-authoring` is designed to address the core pain points of authoring agent skills. By packaging standardized skills into the plugin system, it provides a complete solution covering frontmatter specifications, skill-root and priority management, description writing, length control, and static validation. In addition to standardizing skill structure, the plugin offers a zero-dependency static checking feature through the `scripts/check-skill.js` script, supporting validation of fields, references, anchors, and length. After installation, the `skills/` directory in the package is registered as the skill root, enabling developers to directly use its tools to author, review, and validate skill behavior.
Read MoreIntroduction to the DeepSeek Harness Fix Plugin
`dsh-repair` is a diagnostic and repair plugin for DeepSeek Harness (DSH), designed to address common operational pain points such as installation failures, silent service outages, broken dependency bridges, and upgrade issues. The plugin provides a variety of diagnostic scripts that can fix issues such as boot crashes, corrupted filesystem symbolic links, silent runtime failures, and dropped adapter model IDs. It supports dependency bridge health checks and reconstruction, covers end-to-end validation from startup to upgrade, and optimizes path and process handling for Windows remote job scenarios. The plugin is released under the MIT license, supports cross-machine operations, and must be installed using the official command.
Read Moredsh-plugin-development: Auxiliary Skills for DSH Plugin Development
The DSH plugin development assistant `dsh-plugin-development` is an open-source tool designed to help with developing, debugging, and publishing DSH plugins. Its core features include providing dual half-package scaffolding, supporting development against a fork and three-level validation, supporting publishing and remote deployment, and including a troubleshooting checklist for common pitfalls. After installation, the plugin registers the `skills/` directory in the package as the skill root. It works by leveraging the `cordis.patch.yml` extension mechanism and depends on `@deepseek-ai/dsh-skill-filesystem`. Developers should note that the plugin runs with the permissions of the current DSH process, and it provides a specific solution for scenarios in which the host plugin runs in a PowerShell subprocess.
Read MoreDeepSeek Harness Plugin: Perplexity Consumer Account Web Search
DeepSeek Harness (DSH) plugin `genson1808/dsh-web-search-perplexity-account` allows users to perform web searches using an already logged-in Perplexity Consumer Pro account, without configuring an API key or paying per request. The plugin authenticates via browser SSO or email OTP, forwards DSH search requests to Perplexity, and returns the results. Currently, the official does not provide an installation command, requiring manual setup, and the plugin runs with the permissions of the current DSH process. This solution is suitable for users who already have a Pro account and want to integrate it with DSH, but they should inspect the source code and license before installing.
Read Moreqp-exa-dynamic: Exa Dynamic Highlighting Support and Web Search Runtime Configuration
`qp-exa-dynamic` is a plugin developed for DeepSeek Harness (DSH) that enables web search via the Exa backend and resolves the inability of the native Exa provider to use dynamic highlights. It enables Exa Dynamic Highlights by default and supports multiple retrieval types, including keyword and neural. Users can use the `/exa` command to adjust highlights, retrieval type, and result count in real time. The `exa_search` tool registered by the plugin is not subject to the default limit, allowing up to 50 results. During installation, you need to override the `web` provider in the configuration and set the API Key.
Read MoreDSH Loomy Connect: Zero-Configuration Integration with Loomy's Built-in Models
DSH Loomy Connect is a DeepSeek Harness plugin designed to solve the complexity of configuring local model integrations for developers. By reading the login state and model list of the Loomy desktop application in read-only mode, it registers them as a DSH model provider to enable zero-configuration access. It supports automatic model discovery, displays account and credit information, and provides bilingual support in Chinese and English. Installation requires the Loomy desktop application and DSH version >= 0.1.5-rc.1. The plugin only performs local file reading and authentication, with clear security boundaries, and is suitable for personal learning and research.
Read MoreIntroduction to the DeepSeek Harness Plugin: gabrip780/dsh-hidden-paths
`gabrip780/dsh-hidden-paths` is a security policy plugin for DeepSeek Harness (DSH), designed to prevent AI agents from accessing sensitive files. It blocks access to predefined hidden paths through multi-channel interception—including file tools, shell commands, and others—and automatically masks credentials in the resulting output. The plugin employs a dual-layer mechanism: it rejects sensitive paths before execution and filters outputs after execution, with `.env` and `.ssh` files protected by default. Installation requires using the `dsh plugin` command together with an `id-targeted override` configuration. Note that this plugin acts as an application-layer policy guard and cannot fully prevent code execution or bypass attempts through environment variables or other techniques.
Read Moredsh-selection-tools: Explain/Translate Selected Text in Any App
DeepSeek Harness ecosystem plugin that enables system-level text selection in any Windows application. It prioritizes UI Automation for text extraction and streams explanations and translations from the DeepSeek Agent in a native GDI+ floating window. The plugin has zero dependencies, does not touch the clipboard, and supports gesture detection, draggable and resizable windows, and API calls, with freely configurable translation direction and model.
Read Moredsh-stock-assistant: A-Share Market, Stock Selection, and Backtesting Assistant
dsh-stock-assistant is a plugin for DeepSeek Harness (DSH), designed to provide users with local quantitative analysis capabilities for A-shares. It integrates three data sources—Eastmoney, Tencent Finance, and akshare—and offers 21 model tools for market data queries, quantitative stock selection, strategy backtesting, stock behavior identification, and watchlist monitoring. Users can manage watchlists and backtesting parameters in DSH settings, and monitor unusual market movements using Windows desktop notifications. Installation requires Node.js 18, Python 3, and the akshare library. The project data is intended for learning and research purposes only.
Read MoreIntroduction to the dsh-ocgo-quota Plugin
The DeepSeek Harness (DSH) ecosystem adopts an “everything is a plugin” architecture; `dsh-ocgo-quota` is a community plugin maintained by user `fuzz1og`. Given that detailed source code and documentation are unavailable, be sure to check the repository information yourself before installation. Since the plugin runs with the current DSH process permissions, handle it with particular caution.
Read Moredsh-font-settings: A Font Settings Plugin for DeepSeek Harness Web UI
`dsh-font-settings` is a font configuration plugin for the DeepSeek Harness (DSH) Web UI, designed to improve reading and development experience. It supports customizing UI and code fonts via CSS stacks or the built-in list, and uses `@font-face` aliases and the `size-adjust` property to adjust the sidebar terminal font and font size. In desktop Chromium, it also provides local font enumeration. Installation requires DSH version `>=0.1.7-alpha.1 <0.3.0`, and terminal fonts must use `local()` to reference the full name or PostScript name.
Read Moredsh-goal-auto: DeepSeek Harness Long-Session Auto-Recovery Plugin
`dsh-goal-auto` is a session-level automatic recovery plugin for DeepSeek Harness (DSH), designed to address the issue of long-running goal sessions failing to automatically recover after a Harness restart. The plugin provides a toggle button in the session header and an HTTP API, allowing it to automatically reload the agent on restart, mount presets, re-arm the goal, and keep the driver running. It handles session recovery only and must be used together with an external process watchdog to monitor the health of the Harness process.
Read MoreDeepSeek Harness Community Plugin Manager: Centrally Manage Third-Party Plugins
`dsh-community-plugins` is the native community plugin manager for DeepSeek Harness, designed to address the challenges of fragmented plugin management, updates, and compatibility. It integrates plugin discovery, installation, updates, removal, and compatibility checks; supports filtering by category, sorting, and batch management; and provides a bilingual interface in Chinese and English. To install it, ensure that `dsh` and `pnpm` are available in the environment, then run `dsh plugin --profile web add dsh-community-plugins`. Once enabled, developers can easily search, browse, and install plugins from the settings interface, simplifying the development workflow and improving toolchain maintenance efficiency.
Read Moredsh-search-hub: A Persistent Multi-Channel Search Toolkit for DeepSeek Harness
dsh-search-hub is a plugin for DeepSeek Harness designed to address unstable agent search capabilities. Through persistent configuration and a multi-channel design, it provides Exa semantic search, DuckDuckGo fallback, and xAI Grok integration. Users must manually copy the preset template and restart DSH, then execute searches concurrently using the `multi_search` tool. The plugin supports secure key storage and channel management, but note that network routing depends on the local environment and xAI calls incur costs.
Read Moredsh-token: DSH Local Token Usage Statistics Plugin
dsh-token is a local token usage dashboard plugin for DeepSeek Harness, designed to track costs and consumption. By reading local session logs, it provides four-part token composition analysis (including cache savings), supports global filtering, session drill-down, and budget monitoring. The interface offers multi-dimensional charts, and the plugin runs entirely locally without relying on online services, helping users precisely manage inference costs.
Read Moredsh-agent-evolution: Controlled Agent Composition Experiment Plugin
dsh-agent-evolution is a DeepSeek Harness (DSH) experimental plugin maintained by FriendsHL, designed to help developers validate different Agent preset combinations in a controlled environment. The plugin is in developer preview and is distributed under the MIT License. It provides three core tools: listing presets, running a single experiment, and comparing a baseline with a candidate. Users can configure the maximum depth, token limits, and a preset whitelist. The plugin does not perform automatic scoring or selection; it only provides primitives and runtime evidence, with users responsible for evaluation. Installation requires pinning a specific commit and loading it with `--profile web`.
Read More