DingTalk AI Table Application Development: Aligning AI Table Requirements with Business Semantics for Practical Implementation
For developers needing to build AI Table applications from scratch or perform incremental transformations within DingTalk AI Table. This Skill (ID: @user_718aae83/aitable-app-builder, version 0.1.4, source: community, category: office-efficiency, tag: requires_api_key=false) serves as the primary business entry point for DingTalk AI Table and multi-dimensional table operations. It should be prioritized over direct DWS calls when users request the creation, setup, modification, adjustment, refinement, optimization, diagnosis, or migration of AI Tables, or when handling table structures, fields, records, views, forms, automation, permissions, dashboards, or charts. The workflow involves first organizing solutions based on business semantics, then executing them via currently available DWS, DingTalk Suite, or DingTalk Work Suite, and finally reading back results for verification. This article outlines its functional boundaries, verifiable information (ID, version, tags, source), and the verification status of its installation and activation (noting that both official URLs were inaccessible during this round, and no installation commands were fabricated), while listing applicable scenarios and precautions.
Read Moredsh-next-prompt: Accept "Next Message" suggestions with Tab in the input box
dsh-next-prompt is an MIT-licensed plugin maintained by ihodorchenko and categorized under model inference. After each agent turn, the Host generates 2–3 candidate next user messages using a short LLM request or heuristics. The web GUI displays the best suggestion as ghost text in the empty input box; users can press Tab or click to insert it into the draft, and press Esc to hide it. It does not auto-send or modify session history. Supports installation via a link or git, with configurable model and language. This article describes the architecture, installation commands, configuration options, and usage limitations.
Read MoreBook Deconstruction Site: Splitting a PDF into Web-Readable Chapters
The "Book Teardown Station" is a Skill published by the SkillHub community that utilizes pdfplumber to extract text and join lines across breaks, identifying chapter boundaries. For scanned versions, it employs RapidOCR. It generates structured Chinese summaries for each chapter and outputs a self-contained HTML site featuring a table of contents drawer, chapter navigation, page-turning capabilities, and a reading progress bar. With one-click deployment on CloudStudio, it provides an online URL. It is designed for the chapter-based reading and guided review of Chinese financial and investment PDFs.
Read Moredsh-dock-attention: Harness Requires You in the Background, Dock Icon Will Bounce
**dsh-dock-attention** is a DSH Host client plugin maintained by *ihodorchenko*, compatible exclusively with macOS 12 and later versions. It monitors session events such as approval requests, agent inquiries, and turn completions. When the Harness window is not in the foreground, it launches a native helper application, **DSH Attention**, which briefly bounces in the Dock to request attention; clicking the tile returns the user to Harness. Since Host plugins cannot directly invoke the Electron Dock API, the plugin copies the helper to **Application Support** and triggers it via a file request. The installation command is `dsh plugin --profile desktop add "github:ihodorchenko/dsh-dock-attention"`, and it supports comprehensive configuration and self-check testing.
Read MoreGMP Document Review: Point-by-Point Verification Prior to Controlled Document Issuance
GMP Document Review (@baitongai/gmp-document-review, Version 1.1.2, released on the Baitong AI Service Platform, Source: enterprise, Category: professional) is designed for pending GMP-controlled documents within pharmaceutical companies, including SOPs, procedures, record templates, and forms. It streamlines the pre-publication verification workflow by reviewing documents item by item, comparing them against current enterprise templates, related documents, and applicable regulations. It outputs three types of results: a structured issue list, a "red-line" draft highlighting critical errors, and a clean reviewable draft. Before invocation, users must prepare the documents for review, enterprise templates, and at least one applicable superior document or regulation; superior documents and regulations are provided by the user, as the Skill does not include a regulatory database. Based on the general SKILL.md format, the Skill is tagged with `requires_api_key=false`, indicating no additional API key configuration is needed. It is suitable for self-inspection by SOP/procedure drafters before submission, pre-release quality checks by QA document administrators, and cross-reference consistency reviews by departmental SMEs. No installation commands, configuration fields, or differences regarding Cursor/Codex CLI/Claude Code were obtained from the official page, nor are commands listed in the main text. The review results are intended to aid revision and inspection; final release remains subject to QA decisions per enterprise document management procedures. The official page displays a score of 100,000, 917 downloads, and 0 installations. The zero installation count implies a very small user base, meaning there is limited community experience available for troubleshooting issues.
Read MoreWeChat Investment Blogger Distillation: Deconstructing 7-Dimensional Styles into Profiles
A WeChat Official Account investment blogger “distiller” (investor-distiller, @redfox-data/investor-distiller, v1.0.1, produced by Red Fox Data, classified under SkillHub data-analysis, score 100,000, 949 downloads) targets the repetitive work of manually reading through bloggers’ articles for style analysis. It takes WeChat Official Account article data as input and extracts a seven-dimensional “DNA” across seven dimensions—trading framework, market views, expression style, content depth, interaction patterns, hot-topic maps, and persona genes—and outputs a structured profile. It carries the tag requires_api_key=true, so you need to prepare an API Key before use. Trigger keywords cover distill investment bloggers, distill big-Vs, analyze investment blogger styles, style profiling, and similar phrases. Public materials do not provide installation commands or specific configuration fields. This article only describes the usage workflow and does not speculate. It is suitable for content teams doing account positioning and benchmarking analysis, investment research staff structuring bloggers’ viewpoints, and reverse-lookup bloggers by topic when building hot-topic maps. Limitations include: profile quality depends on sample size; the seven-dimensional DNA is a style abstraction rather than an endorsement of competence; and users must distinguish between original content and advertorials on their own.
Read MoreTaishan Ts-vision: Adding Image Recognition Capabilities to DSH Text-Only Models
Ts-vision (Taishan Image Recognition) is a DSH Windows desktop bundle plugin maintained by iguanren. It leverages the free vision model from Zhipu GLM for image recognition, enabling text-only models to invoke `ts_vision_describe_image` to retrieve image descriptions and perform reasoning during image-enabled conversations. The plugin panel supports Key configuration, model selection and ordering, and `ts_vision_diag` diagnostics. It requires DSH ≥ 0.1.7. Installation is performed by running `install-desktop.ps1` from GitHub, with a restart required to take effect. Licensed under MIT and distributed exclusively via GitHub.
Read Morepandoc-convert: A Skill for Document Format Conversion in Agents
pandoc-convert is a community skill that follows the general SKILL.md format, published by user_cea1fb81 on SkillHub. It is located at @user_cea1fb81/pandoc-convert-pro, version 1.0.0, categorized under office-efficiency, and tagged with requires_api_key=false. It enables an agent to complete tasks according to the Pandoc parameter system when users need to convert document formats. It covers conversions among Markdown, DOCX, PDF, HTML, EPUB, LaTeX, Typst, RST, AsciiDoc, Org, ODT, RTF, and ipynb, and supports single-file and batch conversions, citations, templates, table of contents, Pandoc installation assistance, and PDF engine troubleshooting. Official materials do not provide specific installation commands. The article provides the package coordinates, capabilities, and examples of common Pandoc usage, and outlines applicable scenarios and limitations.
Read Moredsh-loose-review: Narrowed review and runtime circuit breaking for destructive shell commands
IFIAnder's DSH plugin *dsh-loose-review* (MIT, v1.2.0) provides narrow-scope model review before execution and execution-time file deletion circuit breaking for destructive candidate shell commands. It can be configured as the permission preset "Auto Review (Loose)". It focuses exclusively on severe irreversible consequences; rule misses incur zero token cost. It operates alongside the official "Auto" experimental review. Installation: `dsh plugin add IFIAnder/dsh-loose-review`. Configuration and auditing are handled via `cordis.patch.yml` and `audit.jsonl`.
Read MoreQCC Skills: Assisted Generation of QCC Presentation PPTs
The "Quality Control Circle (QCC) Skill" (@engicool/qcc-activity-assistant, version 1.0.1, published by Engicool, categorized under office-efficiency) is an auxiliary generation skill designed for QCC project presentation scenarios. It is intended for use when users need to prepare for QCC presentations, generate presentation PPTs, write speech scripts, or conduct pre-presentation checks. The text outlines its four categories of capability boundaries, available coordinate and version information, and the fact that no API Key is required. It also describes the workflow to be followed after project materials are finalized: organizing materials, generating the PPT, writing the script, and performing pre-presentation checks. Additionally, it faithfully notes the absence of installation commands, configuration fields, and reproducible examples within the retrieved data, avoiding the provision of unverified steps.
Read MoreBuilding a Skill for DingTalk AI Table Demo: High-Fidelity Multi-Dimensional Table Interactions
The "DingTalk AI Table Demo Building" skill is maintained by community publisher user_3a3ef0f8. Classified under dev-programming (Version 1.0.1), it is designed to build and continuously iterate high-fidelity interactive demos for multidimensional tables and AI tables. By referencing DingTalk AI Table, Feishu Multidimensional Table, Airtable, or user design drafts, it implements authentic product interactions such as fields, views, sorting, freezing, primary fields, drag-and-drop, and modal dialogs. This article outlines core features, typical building workflows, use cases, and precautions. No explicit installation command was provided on the official page within the crawled data. The tag `requires_api_key=false` indicates that no additional API Key is required.
Read MorePressure Vessel Design: Embedding the 8-Step Workflow of GB/T 150 and ASME VIII into a Skill
This skill covers the full-process of pressure vessel design, establishing an eight-step standardized workflow from finalizing design conditions to outputting design documents, based on **GB/T 150-2024**, **ASME BPVC VIII**, and **TSG 21-2016**. Trigger keywords have been expanded from approximately 10 to over 40, covering Chinese terminology, English abbreviations, standard codes, and colloquial inquiries. Applicable scenarios include new design, in-service checking, calculation report preparation, material substitution justification, welding procedure qualification, inspection plan formulation, standard clause inquiry, and ASME design for export projects. Released to the community by user_a8c636dd, this skill is categorized as 'professional', requires no API Key, and is currently at version 1.0.1. This article introduces its positioning, trigger keywords, typical usage, installation access, and important notes.
Read MoreChief Brand Officer Masterclass: End-to-End Framework for Delegating Corporate Brand Content to a Single Skill
An end-to-end brand content workshop for group brand directors, marketing heads, and CBOs, positioned as the “digital alter ego” of Chinese enterprises’ Chief Brand Officer. Provided with requirements, it delivers brand content solutions from strategy to execution, covering 24 scenarios: long-form WeChat Official Account articles, executive speeches and remarks, short video scripts, overseas social media posts (LinkedIn/Facebook/Instagram), domestic news briefs, large-event copywriting, brand slogans and naming, media interview Q&As, creative briefs, channel enablement scripts, business correspondence, award submissions, managerial documents, investor relations and financial report narratives, employer branding, crisis communications talking points, ESG reports, Party-building communications, internal notices and policies, instant messaging group notifications, legal documents (brand-style polishing), multilingual globalization, and more. Version 1.0.3. Category: content-creation. Community-sourced. Publisher: user_e699094d. Coordinates: @user_e699094d/dla01. No API key required. This document outlines its capabilities, typical use cases, applicable scenarios, and limitations, and notes that the official installation command is not provided in the retrieved materials; refer to the official address for the correct installation command.
Read MoreTikTok, Amazon, YouTube, Facebook, Instagram, Reddit, LinkedIn, X: Social Media & E-commerce Insights (By RealTrace)
This article introduces the "Tiktok, Amazon, YouTube, FaceBook, INS, Reddit, LinkedIn, X Social E-commerce Insights--By RealTrace--" Skill published on SkillHub. Released by Hangzhou Cangxuan Technology Co., Ltd., this is a business-ops research Skill (Version 1.0.6) that requires a connection to RealTrace MCP and is tagged with `requires_api_key=true`. It covers 10 scenarios, including Amazon review queries, review insights, and multi-ASIN price comparison; TikTok comment demands, content briefs, and TikTok Shop opportunity cards; as well as social media demand validation, product gaps, influencer-product fit, and cross-platform content briefs. The article outlines its positioning, core capabilities, precautions regarding MCP connection and session authorization during activation, reproducible usage paths, and applicable scenarios and limitations.
Read Moredsh-connect-modelscope: Integrating ModelScope Inference and Magic Bean Balance in DeepSeek Harness
dsh-connect-modelscope is a DSH bundle plugin maintained by IFFCheckPass that registers the ModelScope OpenAI-compatible endpoint as the modelscope provider. Its settings page saves the ms- token and selects models, while the chat bar displays a refreshable ModelScope credit balance. It supports storing tokens in the credential service, reusing the MODELSCOPE_API_KEY, and fetching the model list in real-time. Installation: dsh plugin add dsh-connect-modelscope --profile. Please note limitations such as pay-per-request billing, real-name verification with Alibaba Cloud binding, and the absence of unit price display.
Read MoreSeedance: Solidifying the Seedance 2.0 Prompt Structure into a Reusable Generation Workflow
Seedance is a universal SKILL.md format prompt generation Skill published by honeybee1130 on clawhub. Version 1.0.0, it is categorized under content-creation with the coordinates @clawhub_honeybee1130/seedance. It solidifies the six-part "Subject-Action-Camera-Style-Audio-Constraints" structure of Seedance 2.0 into a fixed workflow, solving issues of subject drift, motion blur, and inconsistent styling caused by loosely structured video prompts. Core capabilities include a six-part prompt structure (where the order is a hard constraint), @ tag reference material citation (capped at 9 images, 3 videos, and 3 audio clips, covering four use cases: character consistency, first/last frames, motion transfer, and style reference), three prompt templates (single-shot scenes, multi-shot narratives, action sequences), six advanced techniques for clean, high-speed motion (Continuous Shot Lock, Physics-Motivated Camera, Environmental Anchoring, Scale Progression Arc, Sensory Render Instructions, Rhythm Through Verbs), and a selectable list of negative prompts. Installation is performed via `openclaw skills install @honeybee1130/seedance`. The tag `requires_api_key=false` indicates that an API Key configuration is not required, though model service credentials are still necessary to actually call Seedance 2.0 for video generation. Note: This Skill only generates prompts and is not a video generation interface. The rating of 99229, category design-media, and 901 downloads suggested by SkillHub clues do not match the rating of 100000, category content-creation, 9973 downloads, and 316 installations found in the scraped data; this article prioritizes the scraped data.
Read Moredsh-mnemon-sync-monitor: Keeps Local Markdown Knowledge Bases Synchronized with the dsh-mnemon Documentation Layer
The Documents Source in dsh-mnemon does not track changes in the knowledge base after import. Idlehanker’s dsh-mnemon-sync-monitor watches Markdown directories used by Obsidian, Logseq, etc.; after a debounced merge, it updates Project Documents under an index lock and can optionally distill content into Memory Spaces. Based on the README, this article outlines installation commands, cordis.patch.yml configuration, the settings page and Host API, and highlights considerations such as prune, distill.command, and process permissions.
Read MoreWorkBuddy Buddy Daily Task Skill: Complete Check-in, Dispatch Cat, and Reward Collection in One Go
This Skill consolidates three tasks—WorkBuddy Buddy Station daily check-in, sending the Meow to travel, and collecting travel reward points—into a single execution. It introduces an auto-collection feature scheduled 4 hours and 5 minutes after rewards are credited, preventing missed collections. By utilizing the client's built-in local authentication proxy to call official APIs, it ensures no credentials are read or stored, strictly using the currently logged-in account. Beyond daily execution, it supports querying current travel status, verifying the consistency between the local authentication proxy channel and the account, and troubleshooting failures in check-in or travel automation. Trigger words cover natural language commands such as "Auto check-in WorkBuddy," "Daily points collection," "Buddy Station," "Send Meow," "Send Cat travel," "Auto send Meow to travel," "Collect travel rewards," "Claim travel points," "Get rich Meow," and "Set up WorkBuddy daily auto task." The official address is https://api.skillhub.cn/user_a9fd6ac9/workbuddy-buddy-daily, categorized as ai-agent.
Read MoreAssistant for Formulating Quality Planning Schemes for Municipal Utility Tunnels — Focusing on the Tunnel-Specific Plan
"Municipal Utility Tunnel Project Quality Planning Assistant - Peng Ge Liao Gongcheng AI" is a skill designed to assist in the preparation of quality planning for municipal integrated utility tunnels, cable tunnels, trunk tunnels, branch tunnels, and underground space tunnel projects. Published by user_833fca90 on SkillHub (Version 1.0.0, Source: community, Category: professional), it does not require an API Key. Addressing the common issues of simply applying municipal road or building templates, which results in plans filled with empty rhetoric about "strengthening management," this tool first guides users to define the project type, total tunnel length, cabin cross-section, construction method, pipelines within the tunnel, and auxiliary systems. It then generates quality planning schemes and supporting tables that are suitable for approval, executable, verifiable, and dynamically updatable. A preliminary draft can be generated with a minimum of six data points, while simultaneously listing materials that need to be supplemented. It supports creation from scratch, review of existing plans, sectional writing, and the generation of quality risk analysis tables. It does not replace the approval processes for construction organization designs or high-risk special schemes, nor does it fabricate specification numbers, tunnel lengths, cabin quantities, or award commitments. To enable the tool, the skill file must be imported into a platform that supports Skill loading.
Read MoreFission Distribution Scheme Design (Pro Edition): Integrating Compliance Review into Distribution Scheme Generation
This Skill targets healthcare, e-commerce, private-domain, and community-based businesses and designs compliance-first, viral distribution plans based on product, customer segment, goals, budget, and channel information. Its deliverables include factual verification, goal and financial projections, distribution roles and incentive design, compliance review for two-tier distribution, execution timeline, risk inventory, and diagrams in Mermaid syntax. The price is RMB 0.99 per call, and the price is shown before payment, but a formal plan is generated only after payment authorization is approved. Where data are insufficient, any assumptions are explicitly labeled, and model estimates must not be presented as guarantees of returns.
Read Moredsh-restart-switch: One-click restart of the Cordis kernel or the entire application in the Composer toolbar
dsh-restart-switch is a DSH client plugin maintained by icymint0220. It provides a restart menu to the right of the Composer input area: users can choose to recompose the profile via HMR (preserving the Host and the window) or relaunch the entire DSH (this interrupts tasks and requires confirmation). It supports installation via npm and GitHub and can be configured with kernelDelayMs and appDelayMs. Requires DSH 0.2.1-rc.5 or above in the 0.2.x line; kernel restart relies on live HMR. MIT licensed, version 1.0.0.
Read MoreFive-Star Instructional Model Lesson Plan Design Assistant: Generating Lesson Plans and Presentation PPTs based on the First Principles of Instruction
This article introduces the SkillHub Skill "Five-Star Teaching Model (First Principles of Instruction) Lesson Plan Design Assistant" (@user_72b898e0/fpi-instruction-design, version 2.25.21, category: education, tag: requires_api_key=false). Targeting teachers and trainers in universities, schools, and vocational education, this Skill drafts instructional design (lesson plans) in Word documents based on the Five-Star Teaching Model (Merrill’s First Principles of Instruction) and optionally generates presentation PPTs for lesson explanations and classroom teaching. It is triggered when users mention keywords such as Five-Star Teaching Model, Five-Star Teaching Framework, First Principles of Instruction, Merrill, FPI, or when they require the generation of Five-Star teaching lesson plans, lesson explanation PPTs, or teaching PPTs. The article outlines the Skill's positioning, core features, installation and activation methods (referring to the official address and SkillHub page), typical keyword-triggered use cases, as well as applicable scenarios and precautions.
Read MoreRenovation Pitfall Guide: Skills for Reviewing Quotes and Supervising Construction on Behalf of Homeowners
"Renovation Pitfall Prevention Guide" is a paid life-service Skill located at @org-d48rnc0s/zhuangxiu-fangkeng-pay, version 1.2.0, published by Xinmi City Guoguo Network Technology Studio (Individual Business). It positions itself as a private renovation strategist standing on the homeowner's side, offering line-by-line quote breakdowns (low-balling tactics, deliberate omissions, unit traps, brand and model fraud), prevention of over-added items in high-risk areas (plumbing and electrical outlet locations, waterproofing height, pipe casing drilling, garbage disposal), explanations of semi-packaged, fully-packaged, and labor-only contracts, mandatory renovation contract clauses (caps on additional charges, delay compensation, payment schedules, holding the final payment for three months), water and electrical pressure testing and inspection, hollow tile detection in masonry, bathroom water-tightness testing, paint inspection, as well as checklist-based guidance on avoiding material selection pitfalls for tiles, flooring, bathroom fixtures, cabinets, and doors, recognizing warning signs of renovation company abscondment, and understanding the difference between a deposit (定金) and an advance payment (订金)—all without requiring an API Key. This article covers its origin and positioning, core features, installation and activation methods, and examples of typical queries, while also outlining applicable scenarios and precautions that require verification against local market conditions before use.
Read Moredsh-workspace-running-indicator: Collapsed Workspace Title Bar Session State Restoration Indicator
After DSH collapses the workspace, the title row displays only the folder icon, hiding the waiting dots, loading animations, and completion dots of internal sessions. The icrefin client plugin **dsh-workspace-running-indicator** (v0.3.0) overlays the three standard markers onto the icon cell of the collapsed title. It reads from `sessions`, `workspaces`, and `uiSession` projections and requires DSH 0.2.1-rc.5. To install, run `pnpm check` and `pack` in the repository to generate a `.tgz` file, then install via Settings → Plugins and restart. There are no configuration options; the completion dot is cleared only when the parent row is expanded or the session is opened.
Read MoreTrading Behavior Volatility and Expected Stock Returns: Stock Ranking Based on the Coefficient of Variation of 20-Day Turnover Rate
This article introduces a data-analysis skill published by Zero-One Intelligence on SkillHub: the "Trading Behavior Volatility and Stock Expected Return" skills (located at @zerone-iteration/turnover-volatility-factor, version 1.0.1; paid but requires no API Key). It utilizes the Coefficient of Variation (CV = standard deviation / mean) of a stock's turnover rate over the past 20 days as a trading behavior volatility factor. The skill automatically executes market data retrieval, applies research report screening rules, performs volatility stratification, and determines the expected direction of long and short positions. It ranks the stock pool with a priority on low volatility and outputs directly usable stock selection conclusions. The article successively outlines the positioning, core functional workflow, installation and activation methods, reproducible typical invocation flows, as well as the target audience and limitations. These limitations include the sensitivity associated with the short 20-day window, the tendency for CV to be amplified when the average turnover rate is low, and the fact that ranking conclusions do not constitute a guarantee of returns. Finally, the official URL is provided for further verification.
Read More