Weaver Project Case Materials: 0-12 Customer Cases Organized by Three-Tier Retrieval Sequence
This is the SkillHub community Skill "Weaver Project Case Materials," located at @user_040dcc29/weaver-customer-case-logo-wall. Maintained by user_040dcc29 (Version 1.0.6), it is categorized under "office-efficiency" and tagged with requires_api_key=false, meaning no API Key configuration is necessary. Based on the current project's industry, business, and customer facts, it curates between 0 and 12 customer case names following a three-tier Weaver retrieval sequence. For each case, it first analyzes the top image from Baidu; it attempts the second image only if the first has issues, strictly stopping at the second. Issues found in the top image are retained as "needs confirmation" flags but do not block the delivery of materials. The article provides the official address, the general SKILL.md activation method, and a typical four-step usage guide. It also outlines precautions such as manual verification, the fallback limit for the second image, the attribution of confirmation flags, and the stability of retrieval criteria.
Read Moredsh-sidebar-annotate: In-place image annotation in the DSH right sidebar and return to conversation
dsh-sidebar-annotate is a DSH client plugin (v0.7.0, MIT) maintained by HuaLQIn. In the right sidebar image preview, clicking Annotate overlays a canvas on the original image coordinates, providing six tools—Mark, Pen, Arrow, Rectangle, Square, and Triangle—with adjustable color and line width, undo/redo, mouse wheel zoom, and Space-drag panning. Clicking Send exports a PNG composed of the original image and ink (images with any side longer than 4096 px are downsampled) and attaches a numbered caption list, which enters the conversation via the official draft path. Installation: dsh plugin --profile <profile> add git+https://github.com/HuaLQIn/dsh-sidebar-annotate.git. Requires Node 20+. A full DSH restart is required after installation. Only common image formats are supported; no PDF support or erasing, etc. Please review the source code and license before use.
Read MoreEnterprise Business Fundamentals: A Guide to ERP Business Semantics and Fact Assessment in the Service Industry
Qiqi Business Foundations is a semantic and factual judgment guide for service industry ERP operations, officially released and maintained by Qiqi, available at SkillHub at @77hub/qiqi-business-foundations, with current version 1.0.3, categorized under knowledge-management, sourced from enterprise, and tagged with requires_api_key=false, meaning it can be used without any additional API Key configuration. It addresses the pain point of inconsistent definitions in service industry ERP scenarios: when you need to understand product capability domains, records, documents, master-detail tables, project-based business-finance phases, statuses, references, enumeration values, and time semantics, or need to calibrate statistical definitions for profit, budget, tasks, and work hours, or need to determine whether a set of records can support a certain business conclusion, use it to obtain consistent semantic foundations rather than relying on memory and ad-hoc conventions from various parties. Its core semantic coverage includes product capability domain division, organization of records and documents, the subordination and aggregation relationships of master-detail tables, the correspondence between project-based business and financial phases, document statuses and their transitions, references between documents and between records, meanings of field enumeration values, and temporal calibrations such as business dates and financial periods. There are three typical usage scenarios: first, understanding scenarios, where you use it to figure out what a functional domain contains, what stages a document's status transition goes through, or what an enumeration value represents; second, calibration scenarios, where teams use it to align definitions when they disagree on statistical calibrations for metrics like profit, budget, tasks, and work hours; third, conclusion verification scenarios, where you judge whether the records at hand are sufficient to support a business conclusion. It should be noted that this Skill covers semantics and judgment rules rather than being a data query tool, and it will not directly return business data; the semantic descriptions are based on the official introduction, and if there is inconsistency with the actual product behavior, the current product version should prevail. It is suitable for developers integrating with Qiqi business ERP, product and business analysts participating in requirement reviews, and operations or finance colleagues who need to draw conclusions from business data. The available material for this entry is limited; both the official page and API address are currently inaccessible, and installation commands are not included in the material. When using it, it is recommended to directly open the official page to confirm the installation method and the complete list of capabilities.
Read MoreWeChat Viral Article Generator: Packing the article creation workflow from topic selection to final draft into a single Skill
The WeChat Viral Article Generator is an Agent Skill developed by community developer user_d82fedf3. Located at @user_d82fedf3/wechat-viral-writer (version 1.0.0), it falls under the 'content-creation' category and originates from the community, tagged with requires_api_key=false. It integrates topic selection, content refining, rewriting, and sales conversion into a unified framework, requiring the caller only to provide a topic or source material. This article analyzes its positioning and verified capability boundaries, details the integration methods based on the standard SKILL.md protocol, and outlines three typical usage scenarios. Finally, it discusses applicable contexts and limitations.
Read MorePaper Abstract Generation Assistant: Make Abstracts More Academic and Clear
The Paper Abstract Generation Assistant is a community-sourced Skill for content creation, located at @user_6a6f1523/skill004-test (Version 1.0.0). It takes the original paper abstract as input and refines it into more professional and academic expressions, supporting both Chinese and English while optimizing logical structure and academic terminology. It is suitable for scenarios where users need to process paper abstracts, academic writing, and English abstracts before submission, revision, or academic reports. No installation command or configuration field is provided in the official materials, and the label indicates requires_api_key=false. When using it, users should confirm the client support method through the official address and note that it only performs abstract-level refinement; results may be limited if the original text lacks sufficient information.
Read MoreVideo to Notes: Turning Video Links into Two Notes
Video to Notes is a community Skill located at @user_60decab1/claw-video-notes, categorized under knowledge-management. It consolidates the entire video pipeline into a single call: download (YouTube, Bilibili, Douyin, Xiaohongshu, Twitter/X) → transcription → output of two notes (key points + full transcript) → automatic cleanup of audio and video files. Two transcription engines are available: a local Qwen3-ASR (Apple Silicon, free and offline) or a cloud-based OpenAI-compatible API, with "auto" as the default, preferring the cloud API when an online API is available. It triggers when a user provides a video link or a local audio/video file and requests download, transcription, or video summarization. The metadata notes requires_api_key=true; the installation command and differences across CLIs have not been verified.
Read MoreScreenshot Pro: Annotate UI screenshots with measured coordinates and verify with 15 assertions
screenshotpro, released by HtO404, is a DSH workflow plugin that emphasizes UI rectangles obtained via UI Automation or DOM measurements rather than manual tweaking. It provides a measure, render, and verify pipeline, outputting numbered boxes, arrows, pixelation, and legends, and validates the results with 15 assertion checks. It can be used as a standalone Python CLI or integrated with the DSH Screenshot Pro preset via pnpm add screenshotpro. It is suitable for manuals, process documentation, and other scenarios that require reproducible illustrative screenshots.
Read MoreMarketing Copywriting (Student Edition): Batch Production of High-Conversion Scripts for Physical Business Owners
Marketing Copywriting (Student Edition) is a copywriting Skill designed for brick-and-mortar store owners, published on SkillHub by community author user_151424ca (located at @user_151424ca/yingxiaowenan), version 1.0.0, in the content creation category, with the tag requires_api_key=false. According to its description, it focuses on producing marketing-oriented short-video voiceover scripts, organizes the full piece through a six-step persuasion chain, generates 500–650 characters per script, and produces 3–5 variants in one batch. It also provides five structural approaches for A/B testing by switching target audience, angle, or hook. The intended actions are hard-conversion goals such as adding a WeChat contact, visiting the store, placing an order, or signing up. It differs from soft content for account growth and claims to be usable in any industry. It triggers when users mention terms like marketing copy, marketing short video, livestream selling script, order-pushing copy, product-promotion video script, help me write a selling script, a post promoting an activity, conversion-focused copy, or paid-traffic ad copy. The provided materials do not include installation commands or differences across clients, nor do they provide officially reproducible examples; the body only explains how to access it and walks through one normal workflow. It suits physical-store owners, shop operators, and paid-traffic operators in high-conversion scenarios; it only generates copy and does not handle editing, voiceover, or ad-campaign setup. Its claimed advantages have not been officially validated, and if the input is vague, the output will be vague as well.
Read Moredsh-mobile-shell: DeepSeek App-style Mobile Shell for DSH Web UI
dsh-mobile-shell is an MIT-licensed client plugin maintained by Htao-uu that provides an official-style drawer, App-style input cards, attachment modals, and full-screen settings for the DSH Web UI when the viewport is ≤1023px, with no effect on desktop. The README recommends pairing it with dsh-android 1.4.0 (version 0.7.5+) to enable native camera, gallery, back button, and DSHNative recent photos; installing it standalone may cause issues. Installation: `dsh plugin --profile <profile> add dsh-mobile-shell`, then restart the harness and access via web. This is an unofficial plugin; please review the source code and license before installation.
Read MoreBid Document Audit: A Skill for Decomposing Tender Review into Three Modes
**Bid-document-review** is a Business-Ops category Agent Skill published by the SkillHub community, versioned at 1.0.0 and maintained by *user_a2e8c8fb*, with *requires_api_key* set to *false*. It decomposes bid-document review into three modes: the **evaluator perspective**, which reviews responsiveness, conformity, and competitiveness item by item against the tender document; the **bidder-specific review**, which checks for rejection/invalid-bid risk items, typographical errors, grammatical errors, and formatting mistakes; and **similarity detection across multiple bid documents**, which identifies content duplication and abnormal consistency to help uncover clues of bid-rigging and collusion. Trigger scenarios cover bid-document health checks, tender-document inspection, tender-document review, rejection checks, bid-document similarity/collusion detection, evaluation review, and bid-document assessment. The article outlines the differences in review criteria across the three modes, provides natural-language trigger examples, discusses capability boundaries and selection considerations, and notes that because the official page is front-end rendered and the installation commands and configuration items could not be verified from the scraped materials, no inferences have been made.
Read MoreBuddy Daily Check-in Plus: Auto Check-in Skill for WorkBuddy Daily Points and Cat Travel
Buddy Daily Check-in Plus is an Agent Skill built upon the universal SKILL.md format and implemented purely in Python, originating from the SkillHub platform. It leverages the client's built-in host channel, WBIPC, to invoke official APIs to complete the WorkBuddy daily points check-in, dispatch the cat for travel, and retrieve travel points upon completion. Quotas and rewards are returned by the interface and vary with promotional activities. The skill supports the configuration of scheduled tasks; authorization headers are injected by the client, and credentials are held exclusively by the client. Trigger words include WorkBuddy check-in, daily points, check-in, credits, cat travel, and Buddy travel. Due to a failure in fetching the official address, the installation command and configuration fields could not be verified, which is stated faithfully in the text.
Read MoreKingdee Cloud Galaxy Professional Edition · Report Export: Automated Inventory Report Scraping with Playwright
This article introduces the community Skill on SkillHub: "Kingdee AI Cloud Galaxy Professional Edition · Report Export." Leveraging Playwright browser automation, this Skill exports Inventory Ageing Analysis and Inventory Inquiry tables from Kingdee AI Cloud Galaxy Professional Edition as Excel files saved locally. It supports checkbox filtering (zero inventory, disabled items, warehouse location) and scheduled tasks. Credentials are entered interactively upon first use, and the output directory is configurable. The article details its core capabilities, installation and activation methods (subject to the official page), typical usage, and precautions.
Read MoreNano Banana Pro Architectural Rendering Generation: From Text Prompts to Downloadable Images
This article presents the Nano Banana Pro Architectural Rendering Skill on SkillHub (@org-020jaoqn/lingdongai-nano-banana-pro-architectural-render, version 1.1.3, enterprise source, design media category, publisher Lingdong AI, platform score 100000, cumulative downloads 1073, tag requires_api_key=true). The skill primarily takes a theme description as input, with optional reference images and composition requirements, and outputs downloadable result image links. It is intended for "Nano Banana Pro architectural rendering generation" and synonymous requests; it is not used for construction submission drawings, structural calculations, or construction feasibility reviews. The article outlines its positioning and capability boundaries, required API Key, typical call input/output workflow, applicable scenarios and limitations; official documentation does not provide installation commands or complete reproducible examples, and this article does not fabricate commands. It is recommended to refer to the official SKILL.md on the official page for details.
Read Moredsh-uriauth: Provides a configurable fixed URI entry point for the DSH Web UI
dsh-uriauth is a DSH plugin maintained by hsushjk. On the server side, it exchanges the current process's launch token for browser cookies and provides a configurable fixed path (default: /trustme) for accessing the Web UI, eliminating the need to repeatedly copy `?token` from logs. It is disabled by default and must be enabled after acknowledging the warning in settings. It supports exchange/redirect modes and uri-auth.json configuration, intended for long-term development on private/intranet networks; on public networks, additional authentication is required. Install by cloning the repository, adding it to the web profile via pnpm, and including it in bundles. Compatible with @deepseek-ai/dsh 0.2.1-rc.5 and Node 22+.
Read MoreAuraLife (dsh-auralife): DSH Digital Life Plugin
AuraLife is a DeepSeek Harness plugin maintained by hsushjk that implements a persistent cross-session identity within the Harness. Driven by three hormones, physical energy, and persistent obsessions, it proactively initiates conversations. Its memory grows and decays as a versioned chain of events and facts. When capabilities are lacking, it can hot-load custom self-written "organs." It supports text, images, emojis, voice, and files, as well as image generation and spatial journaling. Data is stored in `~/.dsh/plugins-data/auralife/`. Based on the README, this article outlines the `file` dependency installation, access via port 2332, permission settings, and nginx reverse proxy considerations, and points to secondary development documentation such as `DEV.md`.
Read MoreGMP Compliance Inspector: Self-Assessment Checklist for Pharmaceutical and Food Production Processes
The GMP Compliance Checker is a Skill designed for large model Skill platforms, located at @user_bda87ea5/gmp-compliance-checker-james. Published by user_bda87ea5, it is currently at version 3.0.0 and is sourced from the community. Classified under "professional," it is tagged with `requires_api_key=false`, meaning no additional API Key application is necessary for use. It addresses the issues of missing items and inconsistent standards when checking GMP regulations item-by-item prior to audits or surprise inspections (for-cause inspections). It features a built-in checklist of over 60 items across 9 major categories, covering high-frequency defect areas identified during inspections. Furthermore, it supports cross-referenced checking between China GMP (2010 Revision) and FDA 21 CFR. The three usage scenarios defined in the official description are: checking GMP compliance, self-inspection prior to audits, and verifying production process compliance. As the provided materials do not include installation commands, directory paths, or configuration fields, nor do they provide official examples with parameters or output formats, this article accurately identifies and states this gap without writing unverifiable commands. The materials also do not mention differences across the three client environments, so it is recommended to perform actual testing after import. The article serves as a reminder that: the checklist is static built-in content and needs to be verified against current versions following regulatory updates; the Skill performs checklist verification and cross-referencing but cannot replace QA judgment regarding specific batch records or on-site conditions; and the 60+ check items are not detailed in the text, so if specific categories are of concern, a run should be conducted first to confirm coverage. The SkillHub page displays a rating of 100,000 and 1,001 downloads, though these figures are subject to change over time.
Read MoreAccount Opening Voice Guide: Batch Conversion of Steps to Voiceover
The community skill "Account Opening Guide Voice - Steps to Speech" (@u_319b171b/broker-account-voice, version 0.1.1, category content-creation) converts pre-written account opening steps into spoken narration item by item, generating one voice clip per guideline. The account opening step audio, notification tones, and counter opening audio are uniformly recorded based on the standard account opening narration script. It delivers a package of 8 to 20 voice clips at once and does not require an API Key. The article explains its positioning and core capabilities, reminding users to prepare the step text themselves and pay attention to the quantity range. Since the official page and interface were inaccessible during this review, installation and configuration details remain unverified; only the official address is provided for confirmation.
Read MoreChan Theory Buying Point Scanning Closed Loop: Integrating Buying Point Scanning, Next-Day Conditional Orders, and HTML Reports into a Workflow
"Chan Theory Buy Point Scanning Closed-Loop" is a skill released by user_7f21f988 under the professional category on SkillHub. Version 1.0.0 and sourced from the community, it requires no API key. Targeting the STAR Market and ChiNext, it scans the first, second, and third type buy points based on Chan Theory, supports rolling reuse of watchlist groups, analyzes single-day patterns, and generates next-day conditional orders and HTML reports. Trigger scenarios include running Chan Theory buy point scans, generating next-day conditional orders, setting up automated buy point scanning, and screening stocks for Chan Theory buy signals. Official documentation does not provide installation commands, directory paths, or configuration fields; the integration method must be confirmed according to the host Skill's loading mechanism. This article introduces its features, activation methods, typical usage, applicable scenarios, and limitations.
Read Moredsh-aurachat: An IM-independent DSH AI Companion Plugin
dsh-aurachat is a DSH client-type plugin maintained by hsushjk that provides an AI companion on a fixed port or in the sidebar, featuring markdown-based personas, mood and four memory blocks, sharded chat archiving, space dynamics, voice and emoji packs, and multiple TTS/STT/image generation backends. Models are drawn from the DSH list. The recommended installation command is dsh plugin --profile desktop add dsh-aurachat. The README notes that the project has been archived, with dsh-auralife as its successor. Please review the MIT source code before installation; the plugin runs with DSH process-level permissions.
Read Moredsh-feishu-cli: QR Code Login and Driving the Official Feishu CLI in DSH
dsh-feishu-cli is a DSH client plugin maintained by hoyin-law that wraps the official Feishu lark-cli. It provides sidebar QR code login, a command console, and exposes feishu_cli and feishu_auth to Agents. It covers 18 business domains, 200+ commands, and full OpenAPI invocation. Based on the GitHub README, this article describes installation commands, the two-step scan-code process for initial setup, typical CLI and tool parameters, as well as considerations such as permissions, confirmation for high-risk write operations, and environment variables.
Read MoreWorkbuddy Persona Assistant: Defining Personality and Work Habits for AI Assistants
Workbuddy Persona Setup (@user_99bfbad2/persona-setup) is an AI agent skill published by user_99bfbad2 in the SkillHub community. It performs a preliminary background diagnosis to determine professional background, technical/programming proficiency, and generational preferences. Based on this, it provides rapid or in-depth customized intelligent presets, helping users define an AI assistant's personality, working methods, communication style, and task execution preferences in one go. The customization results are written into identity configuration files (SOUL.md, IDENTITY.md, USER.md) and the memory file (MEMORY.md), allowing subsequent sessions to inherit these settings from the files rather than relying on users to repeat instructions. Users can trigger the skill using conversational commands in Chinese or English, such as "reset personality" or "adjust communication style." It is suitable for all demographics, both technical and non-technical, particularly those born between the 1980s and 1995.
Read MoreFutures Market Briefing: A Comprehensive Review of the Domestic Futures Market
Futures Market Brief is a community-sourced Skill (@user_0ad34d0b/futures, version 1.0.0, categorized under data-analysis, no API key required). It is designed for the domestic Chinese futures market and covers major actively traded categories including ferrous metals, non-ferrous metals, precious metals, energy and chemicals, agricultural products, new-energy products, and financial futures. It strings together a pipeline for fetching market data, computing technical indicators (moving averages / MACD / KDJ / volume-price analysis), generating alerts lists and sector-level advance/decline statistics, and producing Markdown/HTML reports with candlestick charts. It can be triggered via natural-language instructions such as “review futures,” “futures recap,” or “generate a futures daily report.” Fundamentals and key same-day news are supplemented by the model together with references. The article introduces its layered feature set, trigger methods, how to enable it via SKILL.md, as well as applicable scenarios and usage considerations.
Read More`dsh-recent-sessions`: View "Recent" Sessions Across Workspaces at the Bottom of the DSH Web Sidebar
dsh-recent-sessions is a DSH Web plugin maintained by hotheat that displays cross-workspace "recent" sessions at the bottom of the workspace tree via a portal: sorted in reverse chronological order by last activity time, filtering out blank, archived, and subagent sessions, with support for collapsing, pagination, and native StateDot states. To install, run dsh plugin --profile web add github:hotheat/dsh-recent-sessions and restart dsh web. The block is hidden in single-list or search modes; the plugin runs with dsh process permissions, so please review the MIT source code before installing.
Read MoreSales & Marketing Assistant: Transforming Repetitive Business Tasks into Structured Outcomes
Sales & Marketing Assistant (@org-02qudk26/salesforce-marketing-cloud-automation-nt, v1.0.0) was published by Zhiyu Technology, classified under SkillHub as business-ops, and tagged with requires_api_key=false. Its official positioning is to help white-collar workers handle repetitive tasks related to business automation and quickly obtain structured, reusable work deliverables. This article checked the collected materials: rating 100000, downloads 964, installs 0. An install count of 0 means there is no public validation. The name points to the Salesforce Marketing Cloud problem domain, but the materials do not state whether it directly calls Salesforce interfaces, nor do they list API Key requirements. The official page is currently inaccessible, and the article does not provide installation commands or typical usage examples; instead, it explains verification paths and first-trial recommendations: break tasks into small pieces, confirm the output format first, and perform compliance confirmation before connecting to internal data.
Read Moredsh-claude-codex-mode: Read-Only Discussions, Code Changes Delegated to Codex
dsh-claude-codex-mode is a DSH client bundle maintained by hotheat that separates requirement discussions from code modifications within the Git repository. Discussions are handled in read-only mode by the DSH main model or Claude Code CLI, while repository changes are delegated to Codex CLI only after explicit user execution, intercepting structured write tools via a write guard. The plugin page allows switching between API and claude-code-cli modes and provides prerequisite checks for Node, Codex login, Anthropic, or Claude Code. The installation command is `dsh plugin --profile web add github:hotheat/dsh-claude-codex-mode`, requiring DSH >= 0.2.1-rc.2 and Node.js >= 20.
Read More