Product Expert: Introduction to the Skills of the P8-level Product Assistant
"Product Expert" is a Skill designed for product managers, with coordinates @indiv-skill/product-expert, version 1.0.1, sourced from community, and categorized under business-ops. It first determines the product stage (validation/PMF battle/growth/commercialization/platformization) and then selects the appropriate methodology. It covers opportunity identification; user research (Mom Test); requirements analysis (JTBD/true-vs-fake needs/Kano/user story mapping); requirements and design; prototypes and interaction specifications; PRDs; competitive analysis; MVP definition (ranking the four major risks); roadmaps; version iteration and backlog management; prioritization (RICE/WSJF/value–cost); metrics systems (North Star/AARRR/HEART/cohorts/instrumentation); growth experiments and sample-size estimation; business models and pricing; positioning and GTM; product architecture; AI product design; B2B delivery and customer success; engineering collaboration and technical requirements; validation loops; upward reporting; journey maps; requirements reviews; iteration retrospectives; product memory management; as well as a discipline layer covering anti-hallucination for market data, decision audit trails, stakeholder alignment, compliance red lines, and delivery safeguards. The document provides the official URL, capability tiers, and typical input examples, and notes that before installation you should verify the version, metric definitions, and official commands. It is suitable for turning product judgments into structured outputs, but it cannot replace business judgment, real user research, data instrumentation, or compliance confirmation.
Read Moredsh-beta-layout-folder-file: Adding Folders, Fork Trees, and Claude-style Drag & Drop to the DSH Sidebar
dsh-beta-layout-folder-file is a DSH 0.2.x client plugin maintained by William20100221 (MIT, 0.3.1). Without replacing DSH’s sidebar UI, it provides collapsible workspace folders, forks session trees, includes Claude-style drag-and-drop and long-title scrolling, and adds ZIP export, AI-generated titles, and optional per-turn automatic titles in the session menu. The article describes the installation command, basic usage for the first folder, the configuration path ~/.dsh/dsh-beta-layout-folder-file.json, and the applicable scope, and it links to the SkillHub directory page and GitHub source repository.
Read MoreAdvanced Litigation Visualization – Chronology Table Master: Organizing Case Materials into *Case Chronology*
A community Skill designed to organize case materials into a "Case Chronology." It transcribes original text verbatim, cites sources for every entry, and abstains from legal judgment. It generates a Word deliverable, an Excel master sheet, a concise A4 summary, and a case fact draft for downstream skills. It supports PDF, Word, text, scanned documents, and photos. Ideal for pre-litigation assessment, factual sections of complaints or defenses, and pre-trial preparation.
Read MoreSupplier Quotation Comparison: Organizing Multi-Vendor Quotes into a Specification-Aligned Table
Vendor Quote Compare is a Community Edition Skill with the coordinates @user_bd3231bd/vendor-quote-compare. Version 0.1.0, it is categorized under business-ops and sourced from the community, requiring no additional API key. It addresses the problem of inconsistent formats and terminology across multiple supplier quotes, where manual alignment of specifications is time-consuming and prone to overlooking risk items. Its capabilities are limited to three functions: receiving multiple quotes, aligning specifications from various vendors into a comparison table based on consistent dimensions, and flagging price discrepancies alongside risk items that require manual verification. The official description explicitly defines its usage boundary as "for user price comparison without placing orders," meaning it is not responsible for order-related operations. The article also highlights the fact that installation commands and configuration fields are not provided in the official documentation, suggests recommended verification methods, and cautions that quote data constitutes internal business information, necessitating confirmation regarding data handling permissions before use with the Agent.
Read MorePenguin Reading Companion: A Reading Companion Skill Covering All School Stages for Children Aged 6-18
Penguin Reading Companion is a children’s reading Skill published by community developer user_c99ad0c1 on SkillHub. Version 1.0.0, categorized under education, and it does not require an API key. It is positioned as an all-in-one reading companion for children, covering all school stages from ages 6 to 18. It uses Socratic dialogue-based guidance and includes capabilities such as book recommendations, in-depth book discussions, role-playing, parent–child shared reading, reading comprehension training, memorization of classical poetry and prose, and emotional companionship. Its brief states that it comprises 12 major modules and includes an authoritative book list from China’s Ministry of Education along with data linking to real exam questions from China’s high school and college entrance exams. The article introduces its positioning, verified features, and installation information such as its coordinates and tags (the official page is currently inaccessible, and the specific installation command has not been verified), and it also provides applicable scenarios, notes that the community-sourced version is relatively early and data authority needs to be verified independently, and lists the official URL.
Read MoreDSH Connect AI-Toolkit: Automatically Configure the Local AI-Toolkit Proxy and Select a Default Model
The DSH Connect AI-Toolkit is a model inference plugin maintained by whyfail. It automatically reads the local AI-Toolkit proxy port and the API Key from system credentials, and discovers models such as WorkBuddy, TRAE, and Qoder via the `/v1/models` endpoint. The configuration page contains only the "Default Model for New Sessions," with the result written to the DSH default model service. It is compatible with DSH 0.2.1-rc.3/rc.5 and supports one-click installation of `tgz` files or manual addition of packages via the macOS/Windows desktop CLI. The plugin runs with the privileges of the host process and is released under the MIT license; it is consistent with the SkillHub Community Directory and the v0.2.4 package on GitHub Releases.
Read MoreDSH Git Commit: One-click Group Commit in Composer
DSH Git Commit (@eternalheart/dsh-git-commit) is a DeepSeek Harness workflow plugin. By clicking commit in Composer, it reads the workspace, groups changes by Angular type/scope, and generates messages using the current session model. After validation through an editable dialog, the plugin executes Git or SVN commits via the JSON channel, without consuming agent turns. It supports commitlint constraints, sensitive file skipping, and cordis configuration. Install using `plugin_manager install_bundle`.
Read MoreBao'an Order Recovery Daily Report: Automating Grid Order Rejection Data into Text Bulletins and Long Images
The Baoan Return Order Recovery Daily Report is a community-published Skill in the office-efficiency category, maintained by user_74b9b8c0. Currently at version 2.1.1 and built on the general SKILL.md format, it is tagged with requires_api_key=true. It automates a workflow previously performed manually each day: extracting grid return order data from Kingsoft Docs, calculating the recovery success rate, automatically filling in claimant fields, generating text bulletins and long-form images, and sending emails. It is applicable to scenarios such as setting up a return order recovery daily report, deploying to a new grid, or adjusting notification configurations. The article introduces its positioning, core features, activation prerequisites, typical usage path, and notes considerations such as its dependence on Kingsoft Docs, the requirement of an API Key, and its currently low install count.
Read MoreDeepSeek-Chan Sleep Shutdown: Automatic Shutdown After Task Completion (Whale Girl Button)
The DSH workflow plugin `deepseek-chan-sleep-shutdown` (v0.2.0, MIT), maintained by weiplwei, provides a draggable "DeepSeek-chan whale girl" button in the Web interface. When all root sessions are idle and meet success conditions, a cancellable countdown begins, triggering an automatic shutdown via `shutdown.exe` on Windows. It supports right-click toggling between "shutdown on success only" and "shutdown on success/failure," `cordis` configuration for countdown duration and margin, SPA self-healing, and runtime logging. This is effective only on Windows; refer to the official README for installation commands.
Read MoreA Guide to Reading Skincare Ingredient Lists—Understanding Products at a Glance: A Skill for Decoding Ingredient Labels
This article presents the community Skill "The Skincare Ingredient Label Reading Guide — Understand Bottles and Jars at a Glance" (located at @user_2af611f4/skincare-ingredient-literacy, posted by user_2af611f4, source: community, category: life-service, current version: 1.2.0, no API key required). The Skill deconstructs an ingredient list according to fixed rules: ordering rules and locating the 1% line, identifying five major ingredient groups—solvents, functional actives, preservatives, thickeners, and fragrance—the functions and effective concentrations of 10 common active ingredients, an ingredient incompatibility table, and how to verify product filings using the "Cosmetics Regulation" app. Version v1.1 introduces an ingredient analyzer; pasting an ingredient list automatically locates the 1% line, identifies functional actives, and outputs incompatibility warnings along with a 0–100 score. Version v1.2 adds 12 real product structure examples (cleanser/serum/sunscreen/moisturizer cream/AHA products/toner), a complete six-step walkthrough demonstration, and four beginner-oriented Q&A entries. The article covers core capabilities, trigger keywords, installation and the official address, a typical walkthrough workflow, applicable scenarios, and version timeliness reminders.
Read MoreStory: Webnovel Toolbox Router — Unified Gateway for Webnovel Writing Tools
Story: Web Novel Toolkit Router serves as the main entry skill for the web novel toolkit, published by RealQ on SkillHub (version 1.0.0, category: content-creation, source: enterprise, tag: requires_api_key=false). Rather than performing text deconstruction or writing itself, it automatically routes to the corresponding skill within the toolkit based on user requirements, and supports launching a local Dashboard to view the text deconstruction library, writing projects, and edited content. Trigger methods include command format (/story, $story, /story dashboard, $story dashboard) and natural language (/网文, "我想写小说" [I want to write a novel], "打开工作台" [open the workbench], "检查更新" [check for updates]). The article covers its positioning, two verified core capabilities, installation and activation methods as provided on the official page (since no specific installation commands were included in the source materials, the official address should be consulted), reproducible trigger examples, as well as applicable scenarios and considerations. Notably, discrepancies exist between the official page's displayed rating of 100,000, downloads of 1,180, and installs of 0 versus the 99,157 and 1,059 figures indicated in the task clues; real-time re-verification is recommended prior to adoption. The recorded installation volume for this skill is relatively low; the entry point is solely responsible for routing, and the capabilities of individual sub-skills must be verified separately.
Read MoreEnterprise Lead Generation Skill: Targeting Business Leads by Keywords and Filter Criteria
"Enterprise Customer Acquisition" is a Skill published by Lingxi Technology Co., Ltd. on SkillHub, located at @org-w5mcz0hs/lx-enterprise-precision-customer-acquisition. Classified under business-ops, the current version is 1.0.3. Built upon the Lingxi Business Database, it supports querying via four entry points: enterprise name, enterprise abbreviation, name keyword, and legal representative name. It combines these with multiple filtering conditions—such as region, registration time, registered capital, enterprise scale, enterprise type, listing status, financing information, leading enterprises, high-quality enterprises, key qualifications, and key awards—to define an enterprise list for customer acquisition. Trigger scenarios include inquiring about enterprises in a specific location, listing specific types of enterprises (e.g., Unicorns, Gazelles, "Little Giants", listed companies, central SOEs, National High-Tech Enterprises), checking company lists by region and type, and finding customers or suppliers. Results are intended solely for lead discovery and organization and do not represent a complete matching set or verified business facts. This Skill is marked with requires_api_key=true; an API Key must be prepared beforehand. Specific configuration details should refer to the SKILL.md file included in the package.
Read MoreXiaobai-Research: DSH Deep Research Report Engine
Xiaobai-Research (Wattson-Law) is a DeepSeek Harness networked tool plugin that generates in-depth research Markdown reports based on the research-report skill specification. These reports include a plain-language timeline, five-dimensional technical bottlenecks, self-development/procurement trade-offs, and six major comparison tables. It performs gate scanning and three-level directory DOCX formatting through check_research_report_threads and build_research_report_docx (or check_threads.py, build_docx.py). Installation command: dsh plugin add github:Wattson-Law/Xiaobai-Reasearch. Please review the MIT source code and process permissions before use.
Read MoreWorkBuddy Automated Task Sharing Tool: Export Automated Tasks as JSON, Collaborators Can Paste to Create
In a team, when one person sets up an automation task, another often has to manually recreate it based on screenshots, with scheduled tasks being especially prone to omissions. The WorkBuddy automation task sharing tool serializes existing automation tasks into a JSON text. After sending it to a colleague, they can simply paste it back to create the task with one click, and it supports batch operations. Released by community publisher user_97e58a7c, version 1.0.0 is categorized under ai-agent and tagged with requires_api_key=false, indicating it does not depend on an additional API key. Its capabilities revolve around a single sharing workflow: export, cross-channel transmission, and paste-to-create. It is designed solely to recreate identical tasks from scratch, without addressing task merging or incremental updates. At the time of drafting, fetching the two official URLs failed, and the provided materials lacked installation commands and configuration fields. Therefore, this text refrains from providing unverified commands and instead directs readers to the SkillHub page and SKILL.md for instructions. Use cases include internal team task distribution, scheduled task migration and backup, and task template distribution. Please note that this is a first-version community skill with an installation count of 0, and its rating field on the page displays 100000 rather than a five-point scale. The JSON structure is custom-defined by the skill, and cross-version compatibility has not been documented. Furthermore, users must independently assess whether any tasks contain sensitive configurations before sharing them.
Read MoreWeChat Channels Viral Content Analyzer: Breaking Down Videos into Reusable Structures
The WeChat Channels Viral Content Analyzer is an AI skill designed for content creators and short video operators. It leverages AIJJ's capabilities, developed and maintained by community author user_da878c5a on SkillHub under the "data-analysis" category. By inputting a WeChat Channels video link, the AI generates a structured analysis report across 8 major sections, eliminating manual, repetitive tasks of logging rhythm, topics, and angles. This skill requires an AIJJ API Key configuration and operates as a paid API capability, with the current version at 1.1.11. The public page does not specify installation commands or detailed fields of the 8 sections—actual output should be referenced. The article outlines core functions, activation methods, typical usage, and applicable scenarios.
Read MoreDSH Plugin dsh-study-mode: A Learning Mode Prioritizing Topic Knowledge Bases and Retrieval
dsh-study-mode is a DSH memory plugin maintained by waspqj, bundling the "study mode" agent preset and a sidebar learning knowledge base panel. Responses first query the local topic library (BM25), with hits citing sources, content outside the library marked as "unverified", and online verification timestamps. It supports "note this down" ingestion into the library, completion suggestions, bi-directional card linking, and a secondary confirmation verification workflow. Installation command: dsh plugin --profile web add dsh-study-mode, with the library defaulting to ~/.dsh/learn. MIT licensed, suitable for learning and review scenarios requiring traceable sources and controlled hallucination boundaries.
Read MoreAI Image Generation Frontend - Web: Web-based Image Generation Skill for AI Agents
AI Image Generation Frontend - Web (User_3c6cb52e/imagegen-frontend-web-rdxj) is a community skill published on SkillHub. It is categorized under design-media and tagged with requires_api_key=false, indicating that an API Key is not mandatory. The official description states: "A dedicated skill for AI image generation frontends on the Web, helping AI Agents efficiently complete relevant tasks." As of writing, both official URLs failed to load. The retrieved data did not provide installation commands, details on internally called image generation interfaces, differences in enabling the three client types, or reproducible usage examples; these missing elements are explicitly noted in this article without fabrication. The target audience consists of developers who require Agents to generate images on the Web without configuring additional Keys. Note that this is a community-sourced skill with zero installations, and official URLs may change; therefore, users must verify the validity of the addresses and conduct independent testing prior to integration.
Read Moredsh-auto-summary: Summarize and handoff to new Agent when session exceeds token threshold
`dsh-auto-summary` is a DSH workflow bundle maintained by Warsk-80486. It meters session tokens after each turn or end. When `thresholdTokens` is exceeded, it generates a handoff summary via an independent LLM call and uses `spawn` to launch a sub-agent to continue the task. The first user prompt is read from `titleInput` and forced into the handoff template. The original session generates no new events and does not compact, and the sub-agent session does not trigger handoffs by default. It is installed via `git+https://github.com/Warsk-80486/dsh-auto-summary.git` and requires DSH ^0.2.1-rc.5 along with mounts for agents, llm, and subagents.
Read MoreTopic Selection and Evidence List for WeCom Moments User Needs: Turn Comments and Search Queries into Priority Topics
An introduction to the community Skill "WeCom Moments User Demand Topic Selection & Evidence Table" (Publisher: user_fccf428b, Version: 1.0.0, Category: content-creation, No API Key required). Within the WeCom Moments scenario, it converts real comments and search queries into prioritized topics. It delivers a topic evidence table and conducts post-publication reviews based on the effective reply rate on Moments. The article outlines a four-step workflow: input preparation, topic generation, publication execution, and data review. It also specifies its target audience and limitations: input quality dictates the strength of the evidence, while the effective reply rate is influenced by timing and copywriting techniques. Currently, its installation count is 0, and community verification remains insufficient. The SkillHub official URL is provided at the end of the article.
Read Moredsh-roundtable: Multi-role Collaborative IM Panel in the DSH Sidebar
dsh-roundtable（侧边栏「专家团队」）是wangshouyong的DSH工作流插件，支持多角色群聊/私聊、发言权仲裁、分层上下文与事实共识、显式工具授权及原生会话。模型通过宿主ctx.llm运行，数据存储在工作区.dsh-im/目录下。版本0.2.0，采用MIT协议，处于开发阶段；可通过源码构建、发行包或git地址安装，npm包@dabuliu_wsy/dsh-im尚未发布。
Read MoreInvestment Banking Due Diligence Assistant: Designing Verification Procedures and Working Papers for Specific Matters
The Investment Banking Due Diligence Verification Assistant is a "Professional" category Skill, version 2.0.10, published by SkillHub community user user_44747f29. Based on the IMA "IPO Practice Case Library (Continuously Updated)," it focuses on research into due diligence and verification methodologies for A-share IPOs, M&A and restructuring, and refinancing projects. Users input specific verification matters, and the tool outputs actionable verification procedures, sample design, and working paper requirements. Trigger words cover due diligence, verification plan, verification procedures, verification methods, working papers, evidence, confirmation letters, site visits, walkthrough tests, verification ratios, abnormal matters, and alternative procedures. The tag requires_api_key=false indicates no API Key is needed. This article addresses developer pain points to introduce the tool's positioning, core functions, installation and activation (including official addresses and coordinates, but excluding unverified installation commands), typical usage examples based on trigger words, as well as its target audience and limitations (it is a research assistant rather than an automatic working paper filling tool; regulatory standards must still be based on the latest regulations when they change; sources are maintained by the community).
Read MoreWeb Crawler: Automating Formatting and Validation of Academic Dissertations Compliant with University Norms
The Web Collector (@indiv-qf/qf-wycj) is an Agent Skill published by community user user_eb2234bc on SkillHub. It originates from the community, is classified as professional, and is at version 1.7.0. It has a rating of 100000, 1161 downloads, 0 installations, and uses pay-per-use billing at ¥0.20. The configuration requires_api_key is set to false. Its core capability is applying university-standard formatting specifications item by item, including page layout, fonts, hierarchy, headers/page numbers, and bibliography formatting, while outputting a compliance revision checklist. During writing, the official page retrieval failed, and claims regarding the correspondence between university checklists, GB/T 7713.1—2006, and GB/T 7714—2015, as well as generating layout templates, exporting PDFs, and verifying existing Word documents, could not be verified from the retrieved materials and were therefore not included in the main text. Installation commands and reproducible invocation examples were also not verified; only coordinates and official addresses are provided in the text, and installation methods should be based on the official page.
Read Moredsh-umbrella-startup: Customize DSH Cold Start with Umbrella Corporation Theme Boot Animation
**dsh-umbrella-startup** delivers a six-act boot animation for DeepSeek Harness styled after the Umbrella Corporation from *Resident Evil*. Featuring a black curtain with the umbrella emblem, dossier background panels, a five-step authorization sequence, and a chromatic aberration reveal, it includes built-in authentic voices and UI sound effects. Identity, themes, and playback frequency can be adjusted on the settings page. Architecture is ported from dsh-startup-screen, utilizing a separate namespace to coexist with the original plugin. Installation requires Node.js 20+. Enable by running `npm pack` in the plugin directory and adding the local `.tgz` via `dsh plugin --profile web/desktop add`. Configuration is stored in `~/.dsh/dsh-umbrella.json`. Code is MIT licensed; see repository disclaimer for audio and trademark details.
Read MoreDSH Huawei Connect: Integrating Codeway and Guoban Models into DeepSeek Harness
DSH Huawei Connect (dsh-huawei-connect), maintained by wang111928, integrates logged-in models from Huawei CodeArts and OfficeAce into DeepSeek Harness. The plugin provides Huawei dual-platform cards in settings to synchronize login states, display credit balances per platform, select models, and write back to the host configuration. Requests are constructed in the format of the official clients. The installation example is dsh-huawei-connect@1.3.4, requiring DSH >=0.1.7-rc.1 and Node ^22.19.0 or >=24. OfficeAce depends on Windows credentials and sessions; for CodeArts, note the SDK-HMAC signatures and forwarding layer allowance. MIT License.
Read MoreJiandaoyun Assistant: Encapsulating Product Capabilities, Build Blueprints, and API Boundaries into a Skill
The Jiandaoyun Assistant is a knowledge-oriented Skill for the FanRuan Jiandaoyun (jiandaoyun.com) no-code platform. Published by community developer user_b6ce042c, it is currently at version 1.0.0, categorized under business-ops, and distributed through the SkillHub catalog. It follows the standard SKILL.md format and carries the tag requires_api_key=true, so users must prepare an API Key from the Jiandaoyun Open Platform before enabling it. The Skill embeds four types of reference material: a product capability map, a hands-on implementation blueprint, API documentation, and version boundaries. It is triggered when users need help with four areas: building enterprise applications with Jiandaoyun forms, workflows, dashboards, and permissions (such as CRM, OA approvals, inventory management, project management, and manufacturing inspections); consulting on Jiandaoyun product capabilities, functional modules, version pricing, or selection comparisons; integrating and developing with Jiandaoyun Open Platform API v5, Webhook, and single sign-on; and AI capabilities, including AI application building, AI nodes, and the Agent platform. The article begins with the pain points caused by gaps in developer retrieval, then explains the Skill’s positioning, the capabilities it covers, how to install and enable it, and four typical use cases. It also notes limitations: as a knowledge carrier it does not directly operate Jiandaoyun data, the currency of its reference material depends on maintenance, and an API Key/credential is required for access.
Read More