COMMUNITY REGISTRY
Discover verified DSH plugins
Inspect versions, compatibility, HMR behavior, source code, and exact install commands.
All plugins
14880 pluginsv1.2.1
dsh-trae-relayTrae Relay
Managed TRAE CN relay and account console for DeepSeek Harness
v0.2.0
dsh-curatorCurator
DSH Console: Manage all coding agent skills and DSH plugins on your local machine in one place, with built-in skill and plugin marketplaces, and AI-generated Chinese descriptions for every skill and plugin.
v0.3.0
@was35/dsh-reclaimReclaim
Audit DeepSeek Harness disk usage (sessions, attachments, rebuildable caches), classify it by risk, and move anything you choose to delete into the plugin's own recoverable trash. The model can query, never delete.
v0.1.0
dsh-valley-meter-sidebarValley Meter Sidebar
DeepSeek peak/valley timeline plus account balance, pinned in the DSH left sidebar above the workspace region. Port of dsh-valley-meter for DSH >= 0.2.0-rc.2.
v0.2.0
@yukieiji/dsh-command-permissionCommand Permission
DSH plugin: one-click session/workspace command allow patterns on approval prompts, plus per-workspace allow/deny lists edited from the sidebar right-click menu.
v0.1.35
dsh-plugin-galaxyPlugin Galaxy
Plugin Galaxy — visual plugin marketplace inside DeepSeek Harness: browse, search, and one-click install community plugins. · DSH 可视化插件星系：逛一逛，点一下，装好。
v1.0.0
dsh-handdrawn-skinHanddrawn Skin
Hand-drawn sketch-style skin: a monochrome paper-and-ink look — ink-black accents, kraft paper grain, light marker outlines, fine pencil lines, and Kai-style headings.
v0.1.0
dsh-tavily-cxykywTavily Cxykyw
Tavily-backed WebSearchProvider for the DeepSeek Harness web capability seam (ctx.web)
v0.3.2
tydoraTydora
DSH plugin tydora
v0.1.4
dsh-folder-attachFolder Attach
Drag folders into the DeepSeek Harness web composer instead of hitting 'only the desktop app supports folders'. Adds a host folder dialog, /folder, and drag/paste into @path mentions. Plain browser, no desktop app.
v0.1.0
@rakutek/dsh-intelligent-uiIntelligent Ui
Intelligent UI for DeepSeek Harness: the model writes markdown with component tags, expressions, and mermaid diagrams, and the chat renders them in place as live, interactive UI.
v0.1.1
dsh-video-pipelineVideo Pipeline
短视频流水线 skill 包：把一条「编排 skill + 原子 skill」的 AI 短剧生产流水线（TTS 读音校正、文生图/文生视频、图片运镜、成片合成、封面、成片核验、多平台发布）作为内置 skill 装进 DeepSeek Harness，装完即出现在 agent 的技能目录里。Bundled skill pack for DSH: one orchestration skill plus its atomic production skills, shipped inside the plugin and served through a built-in skill provid
v1.0.0
dsh-inju-bootInju Boot
DeepSeek Harness startup animation plugin: plays a full-screen INJU 3D particle startup animation every time Harness launches (WebGL2 · 60,000 particles · flower → moon → ghost text → solid LOGO · geek HUD), with automatic background music playback, top bar integration into the animation, and a full
v0.1.0
@j62268781-alt/dsh-hindsight-bank-routerHindsight Bank Router
把 Hindsight 编码插件的 bank 身份锚定到远程仓库名，而不是本地目录名 —— 让同一仓库在不同克隆目录名下共享同一个记忆库。Anchors Hindsight coding-agent memory banks to the remote repository name instead of the local directory name, so every clone of a repo shares one bank.
v0.2.4
dsh-session-evalSession Eval
Retrospective agent-session evaluation for DeepSeek Harness: deterministic grade cards from persisted session logs, cross-session regression diffs, era comparison — no benchmark runs, no LLM judges.
v0.1.0
dsh-cost-breakdownCost Breakdown
DSH cost and context insights: account-wide token and cost statistics + context composition, official RMB peak/off-peak pricing (including statutory holidays), and a persistent cost readout below the input box.
v0.1.0
dsh-mastery-loopMastery Loop
Opt-in DeepSeek Harness skill bundle: mastery-loop — an exam-revision loop that drills until every knowledge point reaches your target accuracy. / 通用应试复习闭环。
v0.1.1
dsh-codespace-workspaceCodespace Workspace
Use a GitHub Codespace directly as a DSH workspace: the agent reads, writes and runs commands inside the cloud Codespace.
v0.0.1
@dsh-external/dsh-git-panelGit Panel
DSH Web 左栏 Git 面板：分支/upstream/ahead-behind、变更文件列表、按文件 unified diff
v0.2.35
visionVision
Vision DSH bundle: give the agent screen/window vision — see tool captures via the bundled Python cvision and returns the image natively
v1.1.1
dsh-plugin-qqbridge-plusPlugin Qqbridge Plus
QQbridge plus — A plugin based on DeepSeek Harness and qq-bridge, primarily for Shuiyu (maimai DX) score lookups (use #查分 in QQ to check B50 and generate score images with song artwork; OAuth authorization, no admin restriction, and no bridge dependency). It also includes chat extensions (balance an
v0.2.1
dsh-auto-review-defaultAuto Review Default
Put DeepSeek Harness sessions into the built-in Auto review permission preset: new sessions at creation, existing root sessions when they are restored.
v1.1.1
dsh-knowledge-bridgeKnowledge Bridge
Give your DeepSeek Harness agent a brain: host-level knowledge vault access with zero-config discovery, four retrieval tools, and a guard against destructive shell commands.
v0.2.0
dsh-trelloTrello
Trello boards, lists, cards and comments as native DeepSeek Harness tools. API key + token authentication; no OAuth, no MCP, no public hostname.
v0.2.1
dsh-priceflowPriceflow
Official DeepSeek pricing and off-peak message scheduling for DeepSeek Harness.
v0.1.0
dsh-hard-deleteHard Delete
DSH session management: One-click [permanently delete] sessions from the sidebar session “…” menu (irreversible, bypasses the recycle bin)
v0.1.5
dsh-openai-responses-web-searchOpenai Responses Web Search
Configurable OpenAI Responses API provider for DSH web search.
v0.2.3
@asun-labs/dsh-plugin-agent-deckPlugin Agent Deck
Codex delegation cards and docked CLI output for DeepSeek Harness
v1.0.0
dsh-gitnexus-pluginGitnexus Plugin
DeepSeek Harness profile bundle: GitNexus MCP integration — code-intelligence knowledge graph (query, impact, context, explain) over indexed repositories
v0.5.8
dsh-better-reasoning-effort-fishBetter Reasoning Effort Fish
Configure reasoning intensity and input modalities for third-party models (DeepSeek Harness plugin, Fish supplement): declare reasoning levels and image input per model, with automatic adaptation based on the built-in model knowledge base and wire-protocol inference. Edit directly in the cards on th