One-Click Generation of Product Detail Page Image Sets: Batch Process All Assets in One Run
Introduction to the “One-click Product Detail Page Image Set Generation” skill on SkillHub (@infimind/product-detail-page-image-set-generation), published by 极睿科技–AI Expert, categorized under design-media, with enterprise as the source. Targeting e-commerce operations, product photography, brand product teams, and live-commerce sales teams, the skill supports text-only generation and also supports adding reference images to control the subject, composition, and visual style. When used within the AI Hive environment, it automatically uploads reference images before generation, and after submission it automatically saves the task, checks progress, and downloads the images. Applicable materials include e-commerce hero images, product detail pages, advertising key visuals, posters, live-commerce materials, social “seeding” images, product retouching, background replacement, and character-consistency content, covering requirements such as selling-point images, spec images, scene images, material/texture images, long-form detail-page images, and PDP pages. The article explains the capability boundaries, the need for an API key, limitations related to vendor-side sourcing, and the missing install command, and it outlines a typical submission workflow and the intended user base.
Read Moredsh-whoop-easy-connect: Connect to WHOOP and View Health Dashboard in DSH
dsh-whoop-easy-connect is a DSH networking tool plugin maintained by ShawnRen57, integrating with WHOOP API v2. It requires DSH ≥0.2.1 and Node.js 24+. Authentication is completed via a four-step wizard on the settings page, with credentials and tokens stored locally. Opening the dashboard automatically synchronizes data from the last 30 days and supports token renewal. The dashboard centrally displays recovery, strain, sleep, and training records along with their trends. The installation command is `dsh plugin --profile desktop add github:ShawnRen57/dsh-whoop-easy-connect`; it can also be installed via the dshopen.org plugin market index. It is MIT licensed, and data requests are directed solely to api.prod.whoop.com.
Read MoreAutomotive Dealer RIPCAS Marketing Diagnosis Skill: Dual-Side Diagnosis Locating the Unique Principal Contradiction
The Auto Dealer Marketing Expert is a SkillHub community-sourced business-ops skill (version 2.0.0), located at @user_cae83fba/auto-dealer-ripcas-marketing, and requires no API key to use. Its core mechanism is two-sided diagnosis: on the marketing side, it identifies breakpoints across a six-stage pipeline—R Reach, I Interest, P Preference, C Conversion, A Advocacy, and S System; on the consumer side, it runs an eight-module checklist against customers’ verbatim statements. The outputs from both sides are merged to determine the single primary bottleneck. Decisions follow a conflict-handling stack: compliance > verified facts > brand tone > regional conversion, with methodological references numbered and facts tagged in three states (verified/unverified/unknown). Use cases include low lead volume, low store visits, low test-drive rates, stalemate after quoting, EV concerns, difficulty activating trade-ins, low referrals, content posting uncertainty, and post-event review gaps. Deliverables can include scripted talking points, dashboards, post-event reviews, and compliance checks. It is not intended to fabricate model parameters or sales figures, evade advertising laws or financial compliance, promise guaranteed sales outcomes, or replace final legal review and professional financial advice.
Read Moredsh-neuron-memory: Neuron-like Hierarchical Memory and 3D Memory Nebula Panel
dsh-neuron-memory is a memory plugin developed by shadowRjh for DeepSeek Harness. It buffers conversations in a hippocampal layer, performs scheduled nightly sweeps to write into a local JSONL memory store, and injects only the top N entries—matched by keyword or optional vector similarity—rather than the full context. It supports Ebbinghaus decay, spaced repetition, a handoff layer, and zstd-compressed historical session retrieval. Data is stored in ~/.dsh/memories/.neuron/ with no cloud telemetry. It ships with the neuron_* toolset and a 3D memory nebula panel with switchable themes. Install with `dsh plugin add dsh-neuron-memory` as described in the README. Compatible with DSH >=0.2.0-rc.2 and Node >=22.
Read MoreKnowledge Graph Generation for SuperStar Courses: Converting Teaching Documents into Importable Graph Data
Introducing the community Skill "Chaoxing Course Graph Generator": it extracts hierarchical knowledge nodes spanning six or more levels (up to seven levels) from PDF and Word documents, labels node types and semantic relationships, and outputs CSV/Excel files ready to be imported into new-version knowledge graphs on Chaoxing Online Courses. The article explains the Skill's positioning, core capabilities, its installation entry point and usage workflow on the same platform, and highlights caveats such as the official page lacking an installation command and the output structure pending spot checks. It is suitable for instructors and teaching assistants who need to build knowledge graphs in Chaoxing courses.
Read MorePaper Topic Recommendations (Free Edition): A Skill for Transforming Keywords into Actionable Topic Directions
This article introduces the Paper Topic Recommender (Free Version) skill on SkillHub, located at @org-liz6o019/paper-topic-recommender (Version 1.1.3). Published by Weipu Luhe, it falls under the education category and is sourced from enterprise. It requires no login, registration, or API key (requires_api_key=false). By inputting 1-3 keywords, it recommends 3-4 viable paper topic directions with established research foundations from a real academic literature database of hundreds of millions of entries. Each direction is accompanied by representative literature, with links to the main site for full-text viewing. After selecting a topic, users can visit the Luhe official website with one click to experience one-click content generation and literature download. The natural language triggers are "help me choose a topic" or "recommend paper topics". This article outlines its positioning, verified capabilities, access information, typical use cases, and applicable limitations. As the page does not provide installation commands or parameter fields, they are not fabricated here; please refer to the official page for confirmation.
Read MoreTreeCom (dsh-treecom): Managing Conversation Branches with a Tree Canvas in DSH
TreeCom is a DSH Web client plugin for Serenity-2026 (dsh-treecom, v1.0.3, MIT) that manages all branches of a conversation as a pan-zoomable canvas tree. It provides two entry points: a title bar and a "Conversation Tree" tab. Clicking a node only shows details, while "Return to this turn" switches the active branch. Features include hover preview, arbitrary turn branching, collapsible consecutive turns, and tree-based grouping on the left sidebar. Installation command: `dsh plugin --profile web add dsh-treecom` or `github:Serenity-2026/dsh-treecom`; a Host restart and page refresh are required.
Read Moredsh-matlab-runner: Run MATLAB in the DSH sidebar
dsh-matlab-runner is a DSH client plugin maintained by seally-web (MIT, v0.2.2). It provides VS Code-style editing and execution of .m files in the right sidebar, featuring session and batch modes, desktop figures, workspace variable writeback, local .mlx preview, and a session-shared matlab_run tool. The host launches MATLAB with system user privileges to avoid sandbox write failures in the MathWorks configuration directory. Web installation: `dsh plugin --profile web add dsh-matlab-runner`; for desktop, use `bin/install-local.sh`.
Read MoreKindergarten Curriculum Review Assistant: Transforming Children's Interests into Thematic Curriculum Resource Packages
Introducing the community Skill "Kindergarten Curriculum Review Assistant": by inputting children's interests and existing experiences, it cross-references domain goals from the "Guide to Learning and Development for Children Ages 3-6" with PCK (Pedagogical Content Knowledge) key experiences to generate a curriculum review resource package containing theme nature determination, name, resources, curriculum sequence, instructional strategies, and delivery formats. It explains its positioning, core functions, activation conditions, and applicable scenarios, noting that the output requires secondary judgment by teachers.
Read MoreSafety Module: Professional Practice: Registered Safety Engineer "Specialized Practice in Work Safety" AI Question Bank
Safety Module: Professional Practice is a community Skill within the education category on SkillHub, located at @user_af163eaa/safe-quiz-m04-practice (Version 1.0.1). Targeting candidates for the "Registered Safety Engineer" professional qualification exam's "Professional Practice in Work Safety" subject, it offers practice questions generated by AI based on the general syllabus. It explicitly states that it is for practice purposes only, and does not provide exam predictions or constitute official material. The tags indicate requires_api_key=false; since the description does not provide installation commands or reproducible usage instructions, please refer to the SkillHub page for details.
Read MoreIntelligent Comprehensive Analysis System for Class Performance: Delegating Multi-Batch Score Comparisons to AI
This article introduces the Intelligent Comprehensive Analysis System for Class Grades (@guomeng/score-insight-master, v1.1.0, released by Guomeng Digital, categorized under education on SkillHub). This Skill supports the upload of multiple batches of scores from entrance, midterm, final, and mock exams, with AI automatically performing cross-exam comparative analysis and progress trajectory analysis. It is a paid Skill that requires no API Key. The article outlines the core features, official installation entry point, and usage notes. It factually clarifies the installation commands and output formats not provided in the source materials, without fabrication.
Read MoreDeepSeek Usage: Dual-Source Dashboard for Official Billing and Local Session Logs
DeepSeek Usage is a DSH plugin maintained by sdiwenchen-lang (npm: deepseek-usage, v0.1.6). By default, it reads host DeepSeek credentials without login and calls the platform API to display balance and per-day, per-model usage. The local mode aggregates usage from sessionPersistence sessions and estimates costs based on the 139 model pricing rates. The panel supports reconciliation, usage calendar, and sidebar summaries. On desktop, add the package name in Settings → Plugins. For Web/CLI, run `dsh plugin --profile web add deepseek-usage`. MIT licensed; review the source code before installation.
Read MoreComprehensive Study Planning Guide from Elementary to High School: K-12 Subject Planning & Q&A Skill
Introducing the community skill "K-12 Comprehensive Learning Planning Guide" (@user_93c6cff2/k12-learning-coach, v1.0.1, category: education, source: community). It structures the twelve-year journey from elementary school to the college entrance examination into three layers: stage diagnosis, error cause tracking, and targeted gap filling. Featuring built-in knowledge bases covering general principles, roadmaps, Chinese, Classical Chinese, mathematics, English, and chemistry, it is designed for scenarios such as parents or students consulting on learning methods, formulating study plans, analyzing exam papers, and receiving Q&A tutoring, requiring no API key. The article details its core functionalities, verifiable coordinates and version information, and the fact that no official installation command has been publicly released. It also outlines integration considerations, including the exclusion of subjects like physics from its coverage, the need to independently verify the accuracy and timeliness of the knowledge base content relative to the exam syllabus, and the discrepancy of over a thousand downloads with zero installations. Furthermore, it provides the official address.
Read Moredsh-skill-picker: Select and install skills in the DSH input box
The client plugin *dsh-skill-picker* by Sakura025 provides a persistent "Skills" entry for the DeepSeek Harness Web Composer. It allows users to browse by six development stages, utilize full-text search in both Chinese and English, and insert /skill names via single-click or multi-select operations. The plugin also supports in-panel installation to `$DSH_HOME/skills` from GitHub paths, SKILL.md URLs, pasted content, or the skills.sh registry. Installation requires extracting the 1.1.0 package into the profile's `node_modules/@local` and registering it at the end of `cordis.patch.yml`, followed by a full restart of DSH. It is compatible with DSH 0.2.x and Node ≥ 22.19.
Read MorePR and Advertising Assistance Tool for Finance: Monitoring of Announcements, Ads, and Public Opinion from 272 Data Sources
A monitoring and analytics skill for announcement/advertising and public-opinion information targeted at financial institutions such as funds, banks, brokerages, and insurers. It is maintained by SkillHub community publisher user_9d5a2a39, locator @user_9d5a2a39/app-home-monitor, category data-analysis, page version 1.0.13, internal version v10.8. It monitors 272 public data sources (198 apps, 35 institution websites, 20 financial media outlets, 12 financial forums, 7 self-media accounts), crawling only publicly accessible web/H5 pages behind no login, and it does not store session state or personal information. It supports collecting three ad formats: images, text, and video. On the image side, it covers carousel frame-by-frame OCR, full-page backgrounds, iframes, Shadow DOM, canvas, dHash deduplication, high-resolution candidates from srcset, and AVIF-to-PNG conversion. On the video side, it supports poster + src extraction and poster-frame OCR. Analytics include cross-platform comparison across multiple institutions, a competitor matrix, TF–cosine ad-copy similarity, ad-delivery intensity rankings, creative reuse, press-release exposure, sentiment analysis of forums, and public-opinion anomaly early-warning alerts based on five rules with period-over-period comparison, and it outputs a 13-section Word report. Runtime modes include one-off, incremental via --collect, scheduled via --schedule (falling back to the standard library cron implementation if croniter is unavailable), time-range queries via --query, targeted search via --search, ad-image sharing via --share, and a three-step conversational wizard. Version v10.8 increases MCP tools from 10 to 14, adding query_window, diff_runs_summary, app_trend, and competitor_trend; stdio falls back to pure-standard-library JSON-RPC; it fixes frame contamination, invalid imports, and non-serializable responses; removes zero-reference dependencies bs4/lxml; centralizes data-source counting into a single source of truth via source_stats(); and adds selfcheck.py for reconciliation. It is suitable for ad-delivery analysis, competitive monitoring, and public-opinion duty desks. Be mindful of limitations such as login walls, OCR misses, and rule-based alerting, and perform compliance assessment for proxy and User-Agent rotation.
Read Moredsh-plugin-x6-core: Unified Integration of X6 Platform Core and DSH Web Side
dsh-plugin-x6-core (v6.0.0) is the core plugin for X6 on the DeepSeek Harness Web side, providing an X6-branded UI, OAuth quick configuration on the settings page, Global/China (mainland) access regions and X6 model providers, as well as a sidebar app plugin marketplace. To install, run `dsh plugin --profile web add` and pin v6.0.0, then restart `dsh web` and open it using the token URL. After uninstallation, the X6 provider and API keys are retained. The plugin runs with dsh process permissions; before installation, please review the GitHub source code and the MIT license.
Read MoreXiaoguo Quant Trading Ptrade All-in-One Assistant: A One-Stop Tool for Ptrade Strategy Development
The “Xiaoguo Quant Trading Ptrade All-in-One Assistant” is a community-sourced, one-stop Ptrade quantitative trading strategy helper published on SkillHub. Version 1.0.3, it is classified under “professional” and requires no API key. It supports backtesting parameter configuration, automatic generation of Python strategy code, in-depth analysis of trading logs, live-trading signal guidance, and performance attribution and risk assessment, integrating multiple stages from backtesting to analysis into a single workflow. Due to a failed official page scrape, this article is based on the collected materials. It does not provide installation commands or notes on editor differences, and only lists the official URL for readers to verify. As the Skill is newly released, its real-world effectiveness should be validated by users.
Read MoreInvoice Type Recognition for Contracts and Business Licenses: Classification, Tax Rates, and Issuance Types Based on Contracts and Licenses
This is a community-edition Skill published on SkillHub (@user_d533f65e/tax-invoice-advisor, v1.0.2, category: professional). It reads the user-uploaded contract and business license, and in conjunction with the latest 2026 tax laws described in the Skill’s own documentation (the VAT Law and its implementing regulations), it outputs the major invoice category, the secondary classification, the applicable tax rate, and the invoice type (special VAT invoice / ordinary VAT invoice). The description lists a set of trigger phrases (e.g., issuing an invoice, invoice classification, tax code, what is the tax rate, special or ordinary invoice, how to invoice mixed sales), and specifies that the Skill should be triggered when the user uploads a contract or mentions any invoice/tax-rate/invoicing need. The Skill uses requires_api_key=false and does not require any additional API key. The article explains the core features, official information, how to enable it based on SKILL.md, typical conversational usage, as well as notes regarding material completeness, policy updates, and that final invoicing should be subject to the tax authorities.
Read MoreInvoices & Bank Statements Automatic Reconciliation Skill: Automate Line-by-Line Manual Matching
Maintained by community publisher user_77824951, this is a general-purpose Skill in the SKILL.md format (Version 1.0.0, Category: Professional). It automates the line-by-line matching of sales invoices with customer receipts, and purchase invoices with supplier payments. It supports two-way reconciliation for full digital special VAT invoices, full digital general VAT invoices, corporate bank transaction records, and WeChat Work/Alipay merchant transaction flows. The tool identifies red-letter reversals and voided invoices, summarizes current account discrepancies, and includes standard bookkeeping entries, outputting a reconciliation table ready to be copied directly into Excel. Designed for bookkeeping agencies, full-time finance staff at SMEs, and part-time accountants for individual businesses, it accommodates both small-scale and general VAT taxpayers. Tagged with `requires_api_key=false`, it does not require an external API key. However, as the official page scraping failed, installation commands and field formats could not be verified; no information has been fabricated in this text. It is recommended to run a test with a small dataset first.
Read Moredsh-lion-pet: A Breathing Xing Xiao Shi Desktop Pet in DeepSeek Harness
dsh-lion-pet is a third-party desktop pet plugin for DSH, maintained by Rumen233, which displays "Xing Xiao Shi" (DSH Lion) in web pages and Electron desktop mini-windows. It features idle breathing, random walks, drag-and-throw interactions, click responses, and a right-click menu, with persisted position and display modes. It does not invoke models or consume tokens. Based on the official README, this article covers local `file:` installation, display modes (`both`/`desktop`/`web`/`off`), configuration via `localStorage` and `config.json`, as well as optional Chrome MV3 extension usage and licensing notes.
Read MoreWorkBuddy C Drive Migration Expert: Relocating AI Tool Data Off the C Drive
# Introduction By default, WorkBuddy and codebuddy CLI place the data directory at `%USERPROFILE%\.workbuddy`, that is, inside the user directory on the C drive. After the AI tool has been running for a long time, files such as session records, caches, and indexes keep piling up, gradually eating away at the free space on the system drive. Manually moving the directory is risky, however: a direct cut-and-paste may break configuration references, and an accidental deletion could result in data loss. The Skill introduced below, "WorkBuddy C Drive Migrator," is designed to handle this. It performs a health check first, then carries out the migration, and supports rollback. It requires no administrator privileges at any point, nor does it require manually editing configuration files. # What is this "WorkBuddy C Drive Migrator" is a community-sourced Skill, located at
Read Moredsh-github-trending: Browse GitHub Trending Structured Reports in Harness Web UI
## Input Information: ## dsh-github-trending dsh-github-trending is a DSH Cordis network plugin maintained by Royalsan. It scrapes GitHub Trending daily, weekly, and monthly lists and supports language filtering. Using a rule engine, it generates fields such as summaries, growth rates, tech stacks, highlights, and risks. It provides a card panel with pagination and filtering capabilities in the Web UI sidebar, and registers the `github_trending` tool, the `/github-trending` command, and the `GET /dsh-github-trending/report` endpoint. This article covers installation methods (local, GitHub, npm, Release tarball), configuration items such as tokens, and typical usage, as well as notes on caching, API limits, and process permissions.
Read MoreProcess Mapping: Breaking Down the Current State into a Prioritized, Actionable, and Standardized Improvement List
# Foreword The most common way process optimization fails is by leaving the problem at the level of "it feels a bit clunky." Without time baselines, you cannot determine if a wait is ten minutes or three working days; without thresholds, you cannot distinguish between normal queuing and a bottleneck; without prioritization, the optimization list simply becomes a long to-do list with no owner. The Skill introduced below transforms "process mapping" from a one-time discussion into a reusable workflow consisting of six fixed steps: Boundary Definition, Current State Collection, Swimlane Diagram Mapping, Identification of Four Bottleneck Types, Optimization Design, and Implementation & Consolidation. # What is this `@kunlungrowth/kunlun-process-mapping`, named "Process Mapping," published by "Kunlun Growth," version 1.3.0, source ent
Read MoreTeaching and Research AI Assistant: Introduction to Skills for Teaching and Research Scenarios
The Teaching Research AI Assistant is an "education" category Skill published by Kunlun Growth on SkillHub. Currently at version 1.0.0 with an "enterprise" source, it is triggered by keywords such as teaching research, tutoring, instruction, and education. Its tags indicate that no API Key is required. This article outlines its positioning, verified information, activation methods, and important considerations: official documentation provides brief functional descriptions but lacks installation commands and examples; given its installation count of 0, small-scale validation is recommended prior to deployment.
Read Moredsh-docx-format: Automatically Comply with Official Document Format in Word Generated by DSH
Romebody's DSH workflow plugin, dsh-docx-format, ensures that generated .docx files automatically comply with the "General Standard for Official Document Format" through a three-layer architecture consisting of prompts, the docx_format tool, and the tools/result hook. Layout formatting is implemented in lib/docxfmt.py and by default intervenes only when user messages contain trigger words such as "official document." It provides check/apply operations, docx_read functionality, and npm installation commands. Suitable for official document automation scenarios, it requires review of the source code and process permissions before installation.
Read More