Every DeepSeek Harness plugin

What it is — This plugin adds cat themes and an interactive companion to DeepSeek Harness Web UI.

Who it is for — If you need cat-themed interface and a draggable interactive companion in DeepSeek Harness Web UI, you can install this plugin. If you prefer stable workbench layout or other sidebar features, you can skip it.

Watch out — Restart DeepSeek Harness Web UI after installation or upgrade. It runs normally after registration in a new profile. No obvious issues found.

The verdict — I would install it because it offers three original cat themes and a local interactive companion, but only if you don't need terminal and file system features from Better Sidebar.

What it is — This plugin provides interface beautification for DeepSeek Harness Web UI, including four color presets and custom background upload.

Who it is for — When adjusting interface layout and appearance in the DeepSeek Harness Web client, this plugin is suitable. Users not using the browser need not install this plugin, as they do not access the Web UI.

Watch out — Manual approval of the build script is required when installing from GitHub source. Sandbox testing succeeded: installed in a fresh profile and registered by Harness. Compatibility is with DSH 0.1.0-rc.6; some style selectors may fail after upgrade.

The verdict — I would install it, as it implements VSCode-style card layout, but only if using DSH 0.1.0-rc.6.

What it is — After installing dsh-cluster, you can use canvas node-graph to orchestrate multi-agent collaboration in DeepSeek Harness.

Who it is for — If you use DeepSeek Harness web mode for multi-agent collaboration design, dsh-cluster plugin is suitable. If you want to manage agent messages using other methods, you don't need this plugin because it focuses on canvas node-graph.

Watch out — Installation and registration succeed in a fresh profile. No obvious issues found.

The verdict — I would install it because it provides canvas-based multi-agent collaboration orchestration, but only if official bundles are already present.

What it is — This plugin adds file upload and automatic zip extraction to DeepSeek Harness workspace, with drag-and-drop support and configurable limits.

Who it is for — Users who need to reference uploaded files in commands within DeepSeek Harness workspace can install this plugin. Users who do not manage file lists or handle uploads can skip it, as it displays numbered chips with delete options.

Watch out — Installation is from GitHub source, and sandbox testing confirmed successful registration in a new profile. It executes shell commands for decompression fallback. No obvious issues found.

The verdict — I would not install it because it executes shell commands, which may not fit strict security environments.

What it is — Remote-capable recursive UI surface container providing document projection and command routing for DeepSeek Harness browser plugins.

Who it is for — If you are building a DeepSeek Harness browser plugin that supports cross-frontend UI document and command routing, you should install it. If you are developing non-browser plugins or do not need UI container support, you do not need to install it.

Watch out — From source installation you need to manually approve its build script first, otherwise Harness cannot complete registration. No other issues found after approval.

The verdict — I would install it because it allows connecting frontends without transmitting DOM, but I must manually approve the build script first.

What it is — Supplies composable workspace primitives to support UI building for DeepSeek Harness browser plugins.

Who it is for — If you are building browser plugins for DeepSeek Harness requiring draggable pane stacks and document editors, using the DeepSeek Harness model and lacking certain UI capabilities. If you do not rely on the shared UI services provided by dsh-ui-container, you do not need this package.

Watch out — Tested result: build script is blocked requiring manual approval to complete registration. The workspace package must remain installed while the container is enabled.

The verdict — I would not install it because it must be installed together with dsh-ui-container to function.

What it is — Provides a reading pipeline for DeepSeek Harness, automating epub conversion, atomic note writing, continuation reading injection, and local cross-book retrieval.

Who it is for — Suitable for users performing in-depth reading of long books with DeepSeek Harness. Users handling only short documents without needing the pipeline can skip this plugin.

Watch out — Needs python and pandoc tools. Sandbox test passed with high understanding depth and internalization scores. No obvious issues found.

The verdict — I would install it because it solidifies the reading process into repeatable tools. The premise is that the preset is already configured.

What it is — This plugin extracts text from screenshots, clipboard images, and image files using Apple's Vision framework locally on macOS for DeepSeek Harness, with no API key or upload required.

Who it is for — Users who need to extract text from screenshots or images within DeepSeek Harness on macOS are suitable for this plugin. Users who require cloud-based vision models or cross-platform image processing should look for other solutions.

Watch out — Sandbox testing passed with successful installation and registration in a fresh profile. It requires macOS 13 or later and Xcode Command Line Tools to compile the Swift helper. From source installation, the build script needs to be manually allowed.

The verdict — I would install it because it enables local text extraction from images into the DeepSeek Harness conversation on macOS, assuming the necessary development tools are installed.

What it is — This plugin provides grouped workflow boards for the DeepSeek Harness Web GUI.

Who it is for — It applies when using DeepSeek Harness to manage multiple grouped workflow tasks requiring real-time status monitoring. It is suitable when only single workflow views are used.

Watch out — Sandbox testing confirms successful installation and registration in a fresh profile. Requires the workflow engine plugin; approval strategy set to never needs cordis.patch.yml configuration override.

The verdict — I will install it because it fills the gap in real-time grouped workflow monitoring under the workflow engine, which is directly relevant for multi-task scenarios.

What it is — DeepSeek Harness file resource management plugin providing sidebar file tree and multi-tab preview.

Who it is for — When using the web profile in DeepSeek Harness to handle local project files and needing sidebar file tree and multi-tab preview, this plugin is suitable. Users who rely only on command line tools for file management without needing graphical previews can skip it.

Watch out — Sandbox test passed: installed and registered by Harness in a new profile. No issues found in static checks. Installing from source requires manually allowing the build script.

The verdict — I would install it because it provides local file browsing in DeepSeek Harness web profile with no modifications to official packages.

What it is — DeepSeek Harness plugin routes LLM requests to local Qoder CLI to enable model and credential reuse.

Who it is for — When you integrate Qoder models into DeepSeek Harness for tool calling with local login reuse, this plugin provides MCP bridging. Users who do not care about context window and reasoning effort switching can use other plugins since they do not need to report these metadata.

Watch out — Sandbox testing confirms successful registration in a fresh profile. No obvious issues found. It supports long-lived sessions, tool bridging, and real-time model directory updates.

The verdict — I would install it because it enables seamless connection of DeepSeek Harness's compression and tool cycles to Qoder CLI warm sessions without manual credential management.

What it is — Provides cross-session memory library, convention auto-injection, session summaries, agent memory tools, and force-directed graph panel for DeepSeek Harness.

Who it is for — Users who need to manage multi-turn conversations and use agent tools for memory management in DeepSeek Harness are suitable for installing this plugin. Users who do not use the model for memory save or recall can skip this plugin.

Watch out — Sandbox test passed, successfully installed in new profile and registered to harness. No obvious issues found.

The verdict — I would install it because it supports cross-session sharing and non-destructive memory management.

What it is — This plugin integrates DeepSeek Harness agent with Feishu, enabling chat interaction from Feishu, persistent sessions, Markdown replies, and proactive push.

Who it is for — Users who need to converse with a DeepSeek Harness agent directly from Feishu messages and leverage its tools for multi-turn tasks should install this plugin. People who do not want to use Feishu as the input interface do not need to install it.

Watch out — Sandbox testing shows it can be installed but the build script is blocked by pnpm; manual permission for github:kriskwok/dsh-feishu-gateway build is required to complete registration. No other obvious issues found.

The verdict — I would not install it because sandbox testing requires manual build permission, adding unnecessary steps.

What it is — DeepSeek Harness plugin that lets you review Markdown documents in a local browser, with support for comments, replacements, and deletions. Annotations are stored centrally in annotations.jsonl for agent tool calls.

Who it is for — When your DSH agent needs to review Markdown documents word by word, such as architecture docs or ADRs, this plugin provides a local browser interface for annotations.

If your workflow does not involve reviewing Markdown content, you do not need this plugin.

Watch out — Installation is done via npm. Tested successfully in a fresh profile where it registers with Harness. Annotations use a central JSONL store visible across files; no obvious issues found with shell commands or configuration.

The verdict — I would install it because it fills a gap in DeepSeek Harness for document review, letting agents read central annotations and apply changes via normal editing tools.

What it is — dsh-im-bridge integrates DeepSeek Harness with instant messaging, enabling remote control.

Who it is for — When you need to dispatch long tasks remotely in Feishu and receive result notifications, this plugin is suitable for you. If you prefer to interact with DeepSeek Harness directly in the web interface, you don't need to integrate it with instant messaging platforms.

Watch out — Sandbox test passed, installed in a fresh profile and registered. Needs to provide API key or token. No obvious pitfalls found.

The verdict — I would install it because it allows configuring permissions on multiple platforms for explicit approval of dangerous operations.

What it is — A DeepSeek Harness plugin that adds a whale pet to the web UI reacting to agent actions.

Who it is for — If you are using DeepSeek Harness web UI for agent reasoning tasks and want a whale reacting to agent actions in the bottom-right corner, you can install this plugin. If you do not use the Petdex standard desktop pet system, you can skip installing it.

Watch out — Sandbox testing passed: installed in a fresh profile and registered by Harness into the profile. Live captures show zero console errors. No obvious issues found.

The verdict — I would install it because it adds UI interactions to agent work, but whether it's worth it depends on whether you need desktop pet features.

What it is — Adds a right-edge prompt navigator rail to the DeepSeek Harness Web UI conversation, with scroll-spy highlighting and hover preview.

Who it is for — Users handling long conversations in the DeepSeek Harness Web UI who need to quickly navigate through multiple prompts are suitable for installing it. Users who do not interact with the DeepSeek Harness Web UI do not need to install it.

Watch out — Sandbox testing passed successfully in a fresh profile with the plugin registered into the profile. No obvious issues were found.

The verdict — I would install it because it is prebuilt and requires no build step regardless of the installation method.

What it is — Adds a token usage statistics entry to the DeepSeek Harness Web UI sidebar.

Who it is for — Users who perform multi-model LLM calls in the DeepSeek Harness Web UI and need to track token consumption are suitable for this plugin. Users who use only command-line tools without the Web UI do not need this plugin.

Watch out — Sandbox testing passes with installation in a new profile and successful harness registration. No obvious runtime or configuration issues found.

The verdict — If token cost estimation is needed, I would install it because it supports manual price override and data export.

What it is — This plugin provides mobile access setup for DeepSeek Harness.

Who it is for — If I need to use Tailscale network for QR code pairing to access DeepSeek Harness from my phone, this plugin fits my scenario. If I do not use a Tailscale account for device detection, I can skip installing this plugin.

Watch out — Sandbox test passed in a fresh profile with harness registration. No obvious issues found. The build script needs to be manually allowed when installing from source.

The verdict — I would install it because it provides a compatibility patch for old harness versions on plain HTTP phone origins.

What it is — DeepSeek Harness session-bound file attachments plugin, supporting file upload via drag, paste, or button with bounded readers for PDF, Office, and archive formats.

Who it is for — When using the DeepSeek Harness web interface and needing the model to read user-uploaded PDF or Office files in the next step. If you only use CLI or headless mode without web file upload, you can skip this plugin.

Watch out — Sandbox test passed in a fresh profile with successful registration. Static checks found no issues. Installation is from GitHub source code.

The verdict — I would install it because it provides tools for all profiles via shared profile directory and the sandbox test passed.

What it is — ncm-player is a floating-window player for NetEase Cloud Music running in the DeepSeek Harness Web interface.

Who it is for — If you use the web profile in DeepSeek Harness for webpage operations and need NetEase Cloud Music playback support in the interface, you can install the plugin. If your music needs are primarily met by standalone music applications or services on different platforms, you do not need to install this plugin.

Watch out — Sandbox test passed: installed in a new profile and registered by harness. It depends on the local Node.js runtime in PATH and network access to music.163.com. Login status and playback queue are saved in browser localStorage, which may not be available in privacy mode.

The verdict — I would install it because it provides NetEase Cloud Music floating window playback support in the DeepSeek Harness web interface.

What it is — DSH Web GUI plugin that supports @ referencing workspace files.

Who it is for — When using DeepSeek Harness Web GUI for project file referencing, input @ to pop up the workspace file menu. If you do not use DeepSeek Harness's inputTriggers mechanism, the plugin cannot provide the @ trigger function.

Watch out — Sandbox test passed. Installed in a fresh profile and registered into profile by harness. No obvious issues found.

The verdict — I would install it because it supports processing ignored files and git tracking changes, but the premise is using DSH Web GUI plugin system.

What it is — Registers a Tavily search tool with full parameter mapping support for the official web_search tool and rc.6 compatibility patches.

Who it is for — When using the DSH web_search tool in a profile that requires support for depth, topic, recency and other model-side controls. If using a different backend or not relying on the DSH provider registry, you can skip it because it does not replace the official registry.

Watch out — Requires a Tavily API key. Installation requires a restart to apply patches. Sandbox test passed.

The verdict — I would install it because it supports adaptive patches even if the version differs from rc.6.