wolfjkd-trader-data：面向 A 股场景的 Agent 多源数据获取 Skill
wolfjkd-trader-data（@user_abc07f4d/wolfjkd-trader-data，v3.0.0）是 SkillHub 社区 data-analysis 类 Skill，由 wolfjkd 以 MIT 维护，面向 A 股行情、公告、新闻与 Wind 深度数据，内置 data_router.py 并行探测腾讯、Wind、ftshare 并评分选源。本文基于 SkillHub 元数据与官方安装包 SKILL.md，说明 SkillHub CLI 安装路径、Cursor 等 skills 目录、健康检查与 quote JSON、配套 ftshare/Wind 配置，以及 API Key、额度与接口容错等注意事项。
Read Morechatwin：商业诊断与流量工具箱的智能路由入口
chatwin（@user_0720cd1c/chatwin）是 SkillHub 上的 business-ops 工具箱，以 /chatwin 为主入口读取对话上下文，支持新手教程、任务前路由与任务后导航。涵盖商业诊断、名人聊天、流量与内容资产等子工具，适合个体创业与内容创作者。本文基于官方 SkillHub 资料说明能力划分、skillhub 安装路径与三种典型用法。
Read Moredsh-egress-router：DeepSeek Harness 按需出口路由插件
`dsh-egress-router` 是 DeepSeek Harness 的一款代理分流插件。它利用 sing-box 内核，仅对阻断的域名走代理，其余保持直连，支持智能选路（竞速与记忆）、节点状态管理及多种协议。插件以用户权限运行，向 Agent 注入环境变量，使 `curl`、`git` 等工具自动按策略走代理。支持通过 Web 界面导入节点、切换模式及管理内核。注意：数据本地存储，不支持自动更新，且 Harness 自带的 `web_fetch` 不走此代理。
Read Moredsh-graph-runtime: Graph Runtime Capabilities of DeepSeek Harness
`dsh-graph-runtime` is a standalone Cordis plugin that provides graph runtime capabilities for DeepSeek Harness (DSH), maintained by developer snail30xx. It aims to address the complexity of integrating LangGraph in a DSH environment, providing graph definition, automatic mounting, tool registration, and intelligent routing. Core features include: using `defineGraph` to define and automatically mount graphs; leveraging `GraphRoutingAgent` to enable LLM-driven intelligent selection of graph tools for execution; and providing `pre-execute` and `post-execute` extension points for validation and auditing. The plugin supports TypeScript and can be used by adding configuration after installation, making it suitable for scenarios requiring unified management.
Read MoreDeepSeek Harness Plugin: OpenRouter Provider Routing and Cost Analysis
xuan666-lab/dsh-openrouter-provider-advisor is a DeepSeek Harness (DSH) plugin designed to automatically evaluate and switch OpenRouter providers. It ranks providers by comprehensively considering cost, speed, context, and reliability, and offers multiple strategies (such as price-first and speed-first). The plugin supports one-click switching to the optimal route, features a built-in weighted pricing model and reliability protection mechanism, and allows users to customize weights. Installation requires supported DSH and Node.js versions and can be completed via the command line. It runs with host-level permissions to ensure API key security and supports Chinese/English interfaces as well as Agent integration.
Read Moredsh-smart-route: DeepSeek Harness Smart Routing Plugin
The DeepSeek Harness (DSH) plugin `dsh-smart-route` provides an intelligent vendor routing solution as a replacement for `polyglot`. It supports multi-channel chained routing, automatic fallback for all error codes, and an exponential backoff cooldown mechanism to ensure high availability. The plugin registers a single virtual provider with a clean interface, and supports one-click start/stop, retry policy configuration, and multi-chain management. After installation, it defaults to the `deepseek-official` chain, making it suitable for users who seek high availability and fine-grained control.
Read MoreDSH Plugin: satan9394/dsh-model-gateway
satan9394/dsh-model-gateway is a model inference routing plugin within the DeepSeek Harness (DSH) ecosystem. It aims to address multi-model management, cost control, and intelligent routing. The plugin supports over 290 providers, includes 19 built-in routing strategies with fault-tolerance mechanisms, and provides RTK+Caveman compression to save 15-95% of inference resources, along with a free-tier budget feature. Since there is currently no official installation command, users need to obtain it through the ecosystem catalog or from the source code. It is suitable for DSH users who need to integrate multiple providers and control costs, but they should review the source code and license before installing.
Read Moredsh-model-pro: Full Lifecycle Management for the llm-pi-ai Provider
`dsh-model-pro` is a dynamic Cordis plugin for DeepSeek Harness (DSH) that provides a visual full-lifecycle management UI for the `llm-pi-ai` provider. It supports provider CRUD, encrypted API key storage, remote model discovery, connectivity testing, intelligent routing (with seamless fallback), and provider composition. After installation, users can create, configure, and test models through the Web GUI and use the observability dashboard to monitor health status and logs, addressing the cumbersome management of multi-model invocation chains.
Read Moredsh-smart-router: DeepSeek Harness Smart Model Routing Plugin
The `dsh-smart-router` plugin implements intelligent routing via a virtual model, automatically switching to configured models based on request complexity and visual requirements without manual intervention. Core features include difficulty-based routing (hard, normal, and easy tiers), automatic image detection to invoke a vision model, and two classification modes: heuristic or LLM-based. Configuration requires specifying the model for each tier and the visual mode (replacement or pass-through), and it is recommended to use models from the same provider to optimize caching. The plugin supports a fail-open fallback mechanism and can coexist with existing vision plugins.
Read Moredsh-agentic-router: A Learning-Based Intelligent Routing Plugin for DeepSeek Harness
dsh-agentic-router is an intelligent routing plugin for DeepSeek Harness (DSH), designed to automatically select the optimal model tier based on task complexity. It employs a four-expert architecture (Rule-based, UCB, LinUCB, and k-NN) with an EXP3 meta-selector, and continuously optimizes its routing policy in real time through a data flywheel mechanism driven by quality signals such as failures, cost, and latency. The new version supports actual model switching and step-level routing while providing various control tools. It defaults to shadow mode and can be switched to active mode via configuration or commands. All decision and reward data are persisted to disk, enabling replay to restore the learning state, and human feedback correction is supported through the Web UI.
Read Moredsh-auto-router: A Configurable Auto-Routing Layer on the DeepSeek Harness
dsh-auto-router is a plugin for DeepSeek Harness (DSH) designed to enable intelligent model routing through local text analysis. It registers a synthetic Provider that offers routing options such as intelligent, fast, balanced, deep, code, and vision, and supports configurable tier thresholds and session stickiness mode to balance cost and performance. The plugin does not require key transmission, depends on a specific version of DSH, and requires specifying the target Provider and its tier-based model mappings in the configuration file.
Read Moredsh-auto-preset-router: DeepSeek Harness Auto-Routing Plugin
`dsh-auto-preset-router` is a DeepSeek Harness plugin designed to solve the pain point of manual intervention when switching presets. It uses the DeepSeek V4 Flash model to automatically detect the intent of a user’s first prompt and route it to a preset such as Standard, Code, Minimal, or Cordis. The plugin uses one classification call to determine the preset and a second call to select the required tools. If classification fails, the system falls back to Standard. It only processes prompt text and attachment metadata, does not send image content, and does not modify the underlying configuration, improving the user experience through intelligent routing.
Read Moredsh-routing-suite: Adding Selectable Intelligent Routing Modes for DeepSeek Harness
dsh-routing-suite is a DSH community plugin maintained by dragonbaba (v0.1.2, MIT). It adds an "Intelligent Routing Mode" to the mode menu while preserving the Standard preset persona, context, and full tool directory. It supports three strategies: auto, inspect-first, and direct, which guide tasks by checking before modifying, or implementing directly and then verifying. It does not make additional model requests, nor does it read/write files or execute commands. The guide covers desktop and CLI installation, manual preset copying, configuration options, applicable scenarios, and security boundaries.
Read More