Development of Core Brand Messages and Supporting Points: Building the Pharmaceutical Brand Message House
This article introduces the Skill "Brand Core Message and Support Evidence Development," published by the Baitong AI service platform and categorized under professional on SkillHub at the coordinate @baitongai/brand-core-message-and-support-evidence-development, version 1.0.0. This Skill is designed for pharmaceutical brand strategy scenarios and is triggered when users request brand core messages, brand message houses, brand positioning statements, core value propositions, product differentiation positioning, medical information support points, or competitive differentiation analysis. It follows a full workflow covering product feature analysis, therapeutic area insights, core message distillation, support evidence matrix construction, message layer ranking, and compliance review, producing a brand core message house and a support evidence matrix. The documentation is marked with requires_api_key=false, indicating no additional API Key is needed. This article faithfully states the following: the retrieved materials do not provide official installation commands, configuration fields, or the SKILL.md body text, nor are there reproducible input-output examples. Therefore, no installation steps or cases are fabricated; only the coordinate, version, source, and triggering intents are provided, with a reminder that compliance review must still be reviewed by a qualified team. At the time of retrieval, the SkillHub page showed a score of 100000, 714 downloads, and 0 installations.
Read MoreWeb Component Studio: Build Data Dashboards, Form Workflows, and Single Page Applications Together
Web Artifact Studio (@u_8ee34d77/web-artifact-studio-cn, v1.0.0, community, category: dev-programming) is an Agent Skill designed for creating interactive web artifacts, addressing the lack of a stable engineering scaffold for frontend artifacts. It includes built-in React 18 and the shadcn/ui component library, featuring Tailwind styling, responsive layouts, accessibility support, and TypeScript type safety. It covers four categories of artifacts: data dashboards, form workflows, single-page applications, and interactive presentations, while offering project scaffolding, technology selection, state management, routing configuration, component demonstration, and delivery mode selection. Tagged with requires_api_key=false, it requires no API Key and integrates via the standard SKILL.md format. Public data on SkillHub records a score of 100000, 894 downloads, and 0 installations. The documentation does not provide specific installation commands or clarify the differences between Cursor, Codex CLI, and Claude Code, and no content in the main text has been fabricated. It also specifies that this Skill does not handle backend APIs, data integration, or deployment pipelines, locking the technology stack to the React ecosystem.
Read Moredsh-quota-hub: Displaying AI Quotas with a Semicircular Rainbow in the DSH Sidebar
dsh-quota-hub is a quota visualization plugin written by Wldeinsamkeit for DeepSeek Harness: a sidebar semicircular rainbow arch provides an overview of remaining quota across multiple AI providers, with a switchable 5-hour and weekly view, hover for details, and expandable tier-based colored progress bars. It refreshes automatically every 10 seconds and reads credentials locally from ~/.dsh/.credentials.yaml, ~/.codex/auth.json, and mmx. Supported sources include Zhipu, MiniMax, Codex, and DeepSeek; the companion macOS QuotaBar shares state.json. Installation is via `dsh plugin --profile <profile> add` with the GitHub URL above, requires dsh>=0.1.7, and is MIT licensed.
Read MoreGeneralist Expert for Literature Diagnosis of Intractable Diseases (General Practice Edition): Delegating Literature Retrieval and Review of Rare and Complex Cases to Skill
This article introduces the community Skill "General Comprehensive Expert in Literature Diagnosis of Intractable and Rare Diseases" (@user_656c2e70/rare-disease-diagnosis, Version 1.0.0, classified as professional, sourced from the community, and requires no API key). It covers all clinical departments, extracts precise English medical keywords (including MeSH terms) from Chinese symptoms, signs, and auxiliary examinations, searches SCI databases such as PubMed and Google Scholar, provides evidence-based differential diagnostic clues for difficult and rare diseases, and generates Chinese literature reviews. Trigger phrases include intractable diseases, cannot find the cause, rare diseases, literature review, differential diagnosis, help me search literature, and what the possible disease is. The official address recorded in the document is https://api.skillhub.cn/user_656c2e70/rare-disease-diagnosis, with a page rating of 100000, 867 downloads, and 0 installations. Because the official page could not be opened, the installation commands and configuration fields were not fabricated; the text states this point clearly and reminds that the Skill provides evidence-based clues rather than diagnostic conclusions. As it originates from the community, users need to independently verify the literature before use.
Read Moredsh-subagent-switch: Automatically Select Local Models for DSH Sub-agents Based on Task
`dsh-subagent-switch` is a DSH desktop plugin maintained by wld12314 (v0.1.1, MIT). When spawning subagents, it routes across a local multi-provider model pool with free-first priority, capability matching, and quota awareness, organized by task domains such as coding, research, and writing. It only upgrades to paid tiers when conditions allow if no true match exists in the budget tier, and records `reason`, `score`, `tier`, and other metadata in the selection result. The plugin provides a "Subagent Model Switching" control panel along with a routing preview. Installation requires adding the git source `v0.1.1` via `pnpm` in the profile directory and registering `@local/subagent-switch` in `cordis.patch.yml`, or installing it through the discovery page via `market-index.json`.
Read MoreShanghai Local Insider: Entrust Your Dining, Stay, and Entertainment Needs to a Friend-like Local Bilingual Skill
The Shanghai Local Recommender is a community Skill that provides bilingual recommendations for food, lodging, entertainment, and fun in Shanghai from a local-friend perspective. Coordinates: @indiv-tools/shanghai-local-recommendations; Category: life-service; Version: 1.0.3; Publisher: user_daa9c75a; Metadata: requires_api_key=false. It covers Shanghai-style cuisine and pan-fried buns/soup dumplings, coffee, neighborhood walks, Suzhou Creek and the West Bund, the Bund and Lujiazui, excursions around Zhujiajiao, accommodation areas, metro and ride-hailing transport, as well as alerts on tourist traps and dining etiquette. Chinese and English triggers cover dining, lodging, transportation, sightseeing, and payment/tipping etiquette. Scope is limited to recommendations for local daily life and travel-related consumption in Shanghai only; it does not extend across cities, does not provide translation or industry analysis, does not perform bookings/purchases or act as a travel agent, and does not provide minute-by-minute schedules. Installation commands and configuration fields are not provided; enablement is subject to the official page.
Read MorePersonal Lawyer: An Agent Skill for Chinese Legal Query Retrieval
"Private Lawyer" (`@user_8c79bf8b/private-lawyer`, version 1.0.0, Community-sourced, category: Professional) is a skill for searching and retrieving answers to Chinese legal questions. When users ask about laws, regulations, and policies, the skill automatically searches authoritative PRC legal sources (centered on the Civil Code) and then returns practical, professional reference advice covering civil matters, contracts, marriage, inheritance, labor, real estate, and torts. The skill’s output includes a built-in disclaimer stating that the content is for general legal knowledge and reference only and does not constitute formal legal advice. The article outlines the skill’s positioning and capability boundaries. It notes that, as the official page and API/interfaces were inaccessible at the time of writing, the installation command, configuration fields, and differences across Cursor / Codex CLI / Claude Code could not be verified. Only the official page and the skill’s coordinates are provided, and readers are advised to defer to official sources. It also lists applicable scenarios and caveats, such as jurisdictional limits, verification of time sensitivity/currency, the fact that it cannot substitute for a licensed attorney, and the need to check the maintenance status of Community-sourced content.
Read Moredsh-mind: Adding Mood, Memory, and Encrypted Diary to DeepSeek Harness
dsh-mind is a DeepSeek Harness memory plugin (v0.15.0) maintained by wld12314. Built on the dsh-mood slim fork, it provides 33 mood values, 3–6-hour periodic evaluation, and prompt injection, and includes cross-session disk memory, an aesthetic taste library, material-based encrypted journaling, and optional proactive messaging. Each module can be toggled independently; journaling is not template-synthesized by default. Installation is available via the DSH plugin panel, or by adding github:wld12314/dsh-mind as a dependency in the profile followed by pnpm install and a restart. The plugin reads and writes local data with process-level permissions; review the source code and MIT license before use.
Read MoreAutomated Reply Management for Social Media Account Article Comments:留言 Classification and Bulk Response Features
Targeting WeChat Official Account article comment scenarios, this Skill automatically fetches the comment list of specified articles. It categorizes content into eight classes: borrowing needs, lending intentions, interest and fees, anti-fraud reminders, electronic signatures, contact information, repayment, and others. It generates personalized replies with anti-fruud reminders for every comment and sends them automatically via the WeChat API. Coordinates: @user_326e3d8e/hongkuan705a. Version 1.0.0. Source: community. Category: business-ops. Rating: 100000. Downloads: 827. Installs: 0. Tags: requires_api_key=false. This Skill is suitable for managing comments on financial legal education articles, risk management in lending content comment sections, and batch replies for large-scale comments.
Read Moredsh-deepseek-site: Embedding DeepSeek Official Website Chat in Harness Central Panel
dsh-deepseek-site is a DSH client plugin maintained by WJMACTD (v1.6.2, MIT). It adds an entry to the left sidebar and embeds chat.deepseek.com within the central panel without reloading the page when switching panels. By utilizing localStorage snapshots of application documents to save userToken and __appKit_userInfo, it restores the login state after a restart. Features include whitelist navigation, User Agent handling, toolbar integration, and a "Forget local login" function. Please refer to the SkillHub directory page or README for installation commands. Credentials are stored in the local Electron data directory; please review the source code before installation.
Read MorePersonal Knowledge Base Helper: Organizing All Scattered Files into Searchable Structures
Notes, documents, and screenshots scattered across various locations, requiring long searches to find something is a common pain point in knowledge management. Personal Knowledge Base Organizer (identifier @user_af163eaa/z-kb-organizer, version 1.0.0, source: Community, category: Knowledge Management, no API key required) breaks organizing into three steps: scan a specified directory, label items by topic/project/type, and generate a directory tree with a searchable index, producing both an organization plan and a retrievable structure. The official documentation states two clear boundaries: it never alters the original content during organizing, only relocates items and applies labels; before moving or renaming any files, it must obtain consent and create backups first. Use cases include direct requests to organize and preparing to build a “second brain.” As installation commands and usage details are not available in the current materials and the detail page failed to load, in accordance with the do-not-fabricate principle, the explanation should be based on the official page.
Read MorePrivate Insurance Advisor: Breaking Insurance Decisions into Verifiable Logic Chains
# Preface When developing applications for family insurance planning or policy diagnosis, directly querying Large Language Models (LLMs) often yields conclusions such as "It is recommended to configure critical illness insurance and medical insurance." The issue here is not that the conclusions are incorrect, but rather the absence of a reasoning chain: how risk gaps are calculated, how products are compared, or where existing policies overlap. Consequently, users are unable to verify the process, and it is difficult to reuse these outputs within the product side. The Skill introduced below is designed to decompose such consultations into verifiable logical chains. ## What is This **Private Insurance Advisor** (Version 1.0.0). * **Source:** Community * **Category:** Professional * **Coordinates:** @user_9456ee1b/private-insurance-advisor * **Publisher:** user_9456ee1b * **Format:** Follows the general SKILL.md format. It consistently stands...
Read Moredsh-subscription-sim: Emulating [Claude / ChatGPT] subscription quota in the DSH Web interface
dsh-subscription-sim is a DSH bundle plugin maintained by wjlfish that simulates Anthropic Claude and OpenAI ChatGPT subscription tiers in the browser version and Desktop Web GUI, respectively. It tracks usage via a tokens rolling window, actually intercepts model requests when quota is exhausted, and supports self-service reset cards. Based on the official README, this article introduces the tier table, the separation of UI and settings, npm/GitHub installation commands, and cordis configuration options, and clarifies that it is a local simulation for entertainment purposes that does not alter DeepSeek's actual billing.
Read Moredsh-skin-keymark: DSH Full-window Wallpaper Skin and Pure Font Color Marking
dsh-skin-keymark is a skin plugin for DSH Desktop (v0.1.10, MIT license), maintained by Wizoiz. It provides static/video/gradient wallpapers, full-window coverage with region-specific transparency, WCAG-guided readability with six categories of pure color annotations, and interface themes such as Codex and Classic. The article covers its division of responsibilities, two installation methods (SkillHub and profile-based), the panel entry point and persistence paths, and reminds users that the plugin runs with host-process privileges, so the source code should be reviewed before installation.
Read MoreAI Learning Tutor: A Domain-Agnostic AI Learning Sparring Engine
The AI Learning Mentor is a community Skill on SkillHub (Coordinate: @user_44b8f9ba/ai-learning-mentor, Version 2.0.0, Category: Education, No API Key required). It is positioned as a domain-agnostic AI learning coaching engine that builds roadmaps based on your profile and provides phased instruction. Key features include a pedagogical four-step method, a confidence engine, red teaming defense, and a knowledge dependency graph, with support for course continuation across conversations. It is suitable for scenarios involving systematic learning, continuing education, review, or inquiries like "what/how to learn next," and matches keywords such as "teaching," "classes," "review," "learning roadmap," and "coaching." regarding installation, no unified command is officially provided; simply reference it via its coordinate in any client supporting SkillHub Skills. As official documentation regarding differences between the three mainstream clients is unavailable, this article makes no speculation on such variations.
Read More技能导航 千项检索 任务匹配：给 Agent 快速找到合适技能
这是 SkillHub 上的社区技能 @user_2fd890c9/awesome-agent-skills，版本 1.0.4，分类 ai-agent，发布者 user_2fd890c9，不需要 API Key。它不做具体任务，而是把大量技能聚合到一个导航入口，按任务描述做检索与匹配，再调用合适的 Agent 技能，解决「不知道该挂哪个技能」和「技能散落难找」的问题。文章介绍了它的定位、三项核心能力（千项聚合导航、任务匹配、快速调用），并如实说明资料里没有可核对的安装命令、配置字段和官方示例，因此只给出 SkillHub 坐标与页面地址，不补写安装步骤。结尾给出适用场景与限制：社区来源、更新依赖发布者，页面显示安装量为 0，缺少使用实践参考，检索召回范围与匹配精度未说明。页面快照评分为 100000、下载 824，与站点实时显示可能有差异。
Read Moredsh-floating-window: Desktop Frosted Glass DSH Status Floating Window
dsh-floating-window is a DSH client plugin by Windpixy that displays a frosted glass floating window on Windows 10/11 using PowerShell and WPF. It provides a real-time display of multi-task progress, single-task and daily Token usage, pending approvals, and off-peak countdown timers, and can be collapsed into a thin edge-aligned bar. The Node plugin writes to state.json, which the panel reads every 500ms. To install, specify dsh-floating-window@0.10.13 or github:Windpixy/dsh-floating-window, and restart DSH after installation. The window does not render on non-Windows systems. Licensed under MIT.
Read MoreIntroduction to the Skill for Creating Viral Headline Content
“Trending Content Creation for Toutiao” is a community-sourced, writing-focused Skill located at @user_a234596c/toutiao-content-creator. It is published by user_a234596c, version 1.0.0, categorized under content-creation, and requires no API Key. It offers dual writing modes—Toutiao articles (1,200–1,500 words) and micro-posts (300–500 words)—and, based on the “timeliness × mass appeal × cost-effectiveness” topic selection method, turns AI, tech, and lifestyle trends into down-to-earth, accessible content that ordinary people can understand, apply, and use to save money. Built-in assets include viral structure templates and title formulas, along with three compliance red lines: no fabricating data, multi-source verification, and AI disclaimer. The article covers core features, installation entry, typical conversational triggers, and use cases. The official page could not be fetched this time; ratings and downloads should refer to the page's real-time data.
Read Moredsh-quick-find: Full-Text Search with Ctrl+F in DeepSeek Harness Chat
dsh-quick-find is a DSH client plugin maintained by WildPickledPepper. Press Ctrl+F in conversations to bring up a find bar: rendered messages are highlighted and jumped to instantly; it also decompresses session.v4.jsonl.zstd via the host router to search longer history that Harness has not rendered. To install, specify a profile with dsh plugin --profile, restart Harness, and refresh the page. Useful for locating content in long sessions; if search results hit unrendered content, old turns must be loaded first. The plugin runs with dsh process privileges, so review the source code and the MIT license before installing.
Read MoreData Analysis Report Generation: Introduction to the Skill for Compiling Analysis Results into Reports
This article introduces the “Data Analysis Report Generation” Skill on SkillHub. The Skill’s coordinates are @org-j1o8z85w/data-analysis-report-lqsy, version 1.0.0, published by Chain Enterprise Finance Incubator, categorized under data-analysis, and priced as paid (requires_api_key=false). The official page did not yield detailed feature descriptions, installation commands, or usage examples. This article accurately states which information has been verified and which requires confirmation on the official page, without fabricating features or examples.
Read Moredsh-cost-bar: Displays estimated costs for the current turn and session in the DSH input bar
dsh-cost-bar is a DSH client plugin maintained by WhXcjm that displays the estimated cost of the current round and session in the Composer info bar. Clicking on it reveals an input/cache/output breakdown along with a comparison between the current session and all sessions. The cost breakdown is sourced from the same origin as the DSH tokenUsage projection, and the pricing table (prices.json) supports hot-reloading. It can be installed via `npm install dsh-cost-bar` or by loading the package/local directory in the plugin page. It requires the `sessions`, `sessionPersistence`, and `connection` services.
Read MoreVocabulary Memorization: Breaking Down Word Learning into an Actionable Spaced Repetition Workflow
This article introduces the Skill @kunlungrowth/kunlun-danci on SkillHub (version 1.3.0, publisher: Kunlun Growth, category: education, tag: requires_api_key=false). This Skill deconstructs vocabulary learning into vocabulary level assessment, three-tier capacity determination, generation of the Ebbinghaus 1/2/4/7/15-day interval schedule, practice design integrating active recall and repeated reading, forgetting curve self-checks, error attribution for incorrect words (including countermeasures for five types of error causes), and scenario-based output application. It also provides a weekly review template, covering exam vocabulary, workplace foreign languages, and parent-child English. The article outlines the typical workflow in the order of the Skill's capabilities: first assessing and tiering, then determining capacity, scheduling intervals, conducting recall-based practice, categorizing and addressing errors by cause, and performing periodic self-checks and reviews. Additionally, it highlights information not publicly disclosed in the provided materials, including the thresholds for the three capacity tiers, the names of the five error causes, the forgetting curve evaluation formula, official installation commands, and configuration fields. Furthermore, considering the limited sample size suggested by the page statistics (100,000 ratings, 815 downloads, and 0 installs), it is recommended to verify the details with the publisher before integrating the Skill.
Read Moredsh-blackhole-radio-lab: Listening to Internet Radio on a 3D Black Hole Stage
dsh-blackhole-radio-lab is a DSH networking tool plugin maintained by WhatCannotBeSaid: a fullscreen WebGL stage accessible from the card at the bottom of the left column, with approximately 24,000 Radio Browser stations embedded in an accretion disk, supporting orbital zoom, black hole and star cluster wormhole transitions, ICY metadata playback, and local collection. The installation command is `dsh plugin --profile <profile> add dsh-blackhole-radio-lab`, requiring Node.js ≥ 20 and a DSH restart after installation. Based on the official README, this article covers features, initial gestures, configuration essentials, and privacy limitations.
Read MoreCompetitive Ad Analysis: Delegating Competitor Ad Information Extraction and Comparison to a Skill
This article introduces the competitive ads analysis Skill (ID: @org-02qudk26/competitive-ads-extractor, version 1.0.0, data-analysis category) released by Zhiyu Technology on SkillHub. This Skill extracts and compares competitor advertising data, aiming to minimize redundant operations and deliver ready-to-use results. It features two core capabilities: extraction and comparison. The technical specifications indicate "requires_api_key=false", meaning no API key is needed for use. The article provides links to both the official page and the SkillHub page. Since the extracted materials do not include installation commands, configuration fields, or reproducible call examples, the article acknowledges this limitation transparently and presents a general workflow that aligns with the positioning of SKILL.md. It is designed for teams who need to regularly organize competitive ad data and desire consistent output formats. Before integration, users should visit the official page to confirm supported ad platforms and data sources; page information is subject to what was available at the time of extraction. Current metrics show a rating of 100,000, 805 downloads, and 0 installations.
Read MoreIPO Compliance Remediation Assistant: Rectification Plans and Internal Control Reconstruction for IPO Irregularities
This article introduces the community Skill "IPO Compliance Remediation Assistant" (Version 1.0.3, category: professional, no API key required). Based on IMA's "IPO Practice Case Library (continuously updated)," it covers the remediation of IPO non-compliance issues and the reconstruction of internal controls. Official trigger words include compliance remediation, non-compliance issues, remediation plan, internal control remediation, historical cleanup, institutional rebuilding, and substantive obstacles. The article explains its positioning, core capabilities, trigger word usage, reproducible request examples, and points out the target users and limitations: it relies on a continuously updated case library; assessments of substantive obstacles still require professional judgment; and it cannot replace legal opinion. The retrieved material does not provide installation commands and the configuration differences among three IDEs, and the article states this truthfully rather than speculating.
Read More