Building Construction Plan Xia: An Intelligent Assistant for Technical Bid Generation
This is a technical bid generation Skill for building construction, produced by "Peng Ge Liao Gongcheng AI" and hosted on SkillHub. Version 1.0.0, classified under "office-efficiency," and requires no API Key. It breaks down the drafting process for construction organization designs/technical bids into the following steps: tender document analysis, point-by-point verification of scoring criteria, generation of the construction organization outline (containing at least 11 standard chapters), chapter-by-chapter writing (defaulting to no fewer than 6,000 words per section), and multi-level reviews (three levels: section, chapter, and the entire draft), along with alerts for high-risk content. Adopting a "phased progression + phase confirmation" mode, every step is adjustable to avoid extensive rework after the full draft is completed. It supports PDF/Word uploads and can automatically fill in missing information based on conventional practices with one click when data is incomplete. It is suitable for technical bid preparation for residential, public, and industrial buildings, as well as specialized projects, catering to bidding specialists, technical directors, commercial personnel, and construction managers. **Please Note:** The generated output is a professional first draft. High-risk content—such as parameters for high-risk/complex engineering, resource allocation, and construction schedule milestones—must be manually reviewed and confirmed before bidding.
Read Moredsh-compat: Plugin Compatibility Layer and Pre-installation Check
**dsh-compat** is a DSH client compatibility layer maintained by Starsky16. Its **L2 loader** isolates modules to prevent faulty plugins from causing cascading white-screen crashes, while the **L3 shim** fixes transparent overlay issues and optimizes touch animations. The **L4 doctor** performs pre-installation checks on platforms and dependencies. It supports full capabilities on Termux (termux-dsh), automatic L2 detection on Linux/macOS, and is restricted to L3/L4 on Windows. Requires Node ≥ 24.2. Version 0.1.1, MIT License. **Installation example:** `dsh plugin --profile <name> add https://github.com/Starsky16/dsh-compat.git` Restart `dsh` to activate.
Read Moredocs: Shopify Integration Skill (Under Maintenance)
docs is a community-sourced Agent Skill categorized under dev-programming, located at @user_5e5107f8/docs. Version 1.0.0 was published by user_5e5107f8. While it claims the capability of "Shopify integration," the official description explicitly states that it is currently under maintenance, with founders@maton.ai listed as the sole support channel. Confirmed metadata indicates that the tag requires_api_key=true, meaning an API key is required for activation. The page displays a rating of 100,000, 833 downloads, and 0 installations; consequently, there is no practical installation data or community feedback available for troubleshooting. As the documentation does not provide installation commands, Skill directory paths, configuration fields, or reproducible invocation examples, this article will not cover these details or speculate on the specific scope of the Shopify integration. Starting with common failure scenarios encountered when configuring Skills for coding Agents, this article first outlines the current status of this Skill. It then lists verified fields and confirmed capabilities, highlighting two prerequisites for activation: confirming whether the Skill has resumed availability, and obtaining an API key while clarifying its permission scope. Finally, the article outlines applicable scenarios and precautions. It is suitable for teams currently working on Shopify integration who are willing to independently monitor the maintenance status. It cautions against integrating a Skill under maintenance into workflows requiring long-term stability and provides the official URL for reference.
Read MorePi Liansheng's Instructional Design Approach: Codifying Goal-Oriented Instructional Design into a Callable Skill
This summary describes a goal-oriented instructional design system based on Pi Liansheng’s *Instructional Design* (2nd edition), provided as “Community Skill,” Version 1.3.0. It categorizes education-related content with community sourcing and is contributed by user_fd7c20e2. The material covers all 16 chapters, including four methods of learning objective statements, Gagné’s five learning outcomes, Anderson/Bloom’s two-dimensional classification, a five-step task analysis method, and a six-step, three-phase, two-branch instructional design model. In addition, Chapters 7–12 focus on teaching strategies for six types of knowledge. The summary also outlines instructional delivery strategies, classroom management strategies, learning outcome measurement and assessment, and formative and summative instructional evaluation. It includes two complete geography examples—contour line reading and agricultural location factors. The tag `requires_api_key=false` indicates that no API key is needed. The text explains how to install and activate the system, typical calls, and limitations, including community-sourced content, limited subject coverage in the examples, and the official page being inaccessible for verification.
Read Moredsh-tri-model: Configuring a "Planning → Execution → Review → Iteration" Multi-model Pipeline in DSH
dsh-tri-model is a DSH-native workflow plugin maintained by starry-sky-blip that chains planning, execution, review, and automatic rework into a configurable pipeline. Parameters and models are configured under “Settings → Multi-Model Collaboration,” including endpoints, protocols (OpenAI/Anthropic/Gemini), models, and tiers; API keys are stored in the DSH credential store. It provides four tools including tri_run and the commands /tri, /tri-doctor, and /tri-models. npm installation: dsh plugin --profile desktop add dsh-tri-model; you must restart DSH after installation. It’s ideal for use cases involving multi-vendor models within the Harness, reducing external orchestration. Please review the source code and MIT license in advance.
Read Morebig8: Metaphysical Q&A Skill Regarding Feng Shui, Physiognomy, BaZi, Astrology, Divination, and Almanac
big8 is a general-purpose SKILL.md skill designed for image-based Feng Shui reading, face-reading (physiognomy) analysis, BaZi (Four Pillars) calculation, zodiac/horoscope lookup, daily hexagram divination, and traditional Chinese almanac queries. Developed by community developer user_15292d5a, it is at version 1.0.0 and sourced from the community. It is categorized under life-service, and the tag information is: requires_api_key=false. It consolidates image-based Feng Shui and physiognomy analysis, BaZi casting from birth dates, zodiac fortune lookup, ritual divination (casting hexagrams), and almanac queries into a unified entry point, making it suitable for developers who need to mount life-service metaphysical Q&A capabilities in conversational products. This article verifies its core functions based on the official brief and notes that where the materials do not provide installation commands, configuration fields, or reproducible examples, third-party commands should not be copied blindly. It also cautions that output accuracy must be validated independently, and it highlights the boundaries of metaphysical content, image privacy, and runtime requirements.
Read Moredsh-auto-archive: Idle Session Auto-Archiving and Projection Cache Cleanup
dsh-auto-archive is a DSH local plugin maintained by StanElka. By default, it calls the official archiveSession to archive sessions that have been idle for 14 days, and for sessions that have already been archived and have been idle for approximately 90 days, it only deletes the rebuildable session projection cache without deleting transcripts, attachments, or workspace files. It supports 30-minute inspection, pinning, whitelist protection, per-operation limits, and dry-run. Configuration is located at ~/.dsh/auto-archive/config.json, and it provides a settings panel and a local loopback API. Installation requires running `dsh plugin --profile desktop add github:elk-9527/dsh-auto-archive` followed by restarting DSH.
Read MoreRoot Cause Analysis and Fix Guidelines for Python Exception Logs: From Traceback to Resolution Steps
This article introduces the community Skill "Python Exception Log Root Cause Localization & Fix Guide" (@user_2fd890c9/exception-logger, Version 2.0.3, categorized under dev-programming). It parses complex Python multi-layer tracebacks and exception chains to pinpoint the initial root cause and provide actionable repair guidance. The Skill is usable without an API Key. The article covers core functionality, activation instructions, input formats, typical usage, applicable scenarios, and limitations.
Read MoreComposition Experts - Prompt Engineering: Implementing Character Consistency and Composition Control within Prompts
This Skill targets AI drawing prompt engineering, covering five model syntax families: natural language, conversational, MJ-Niji, SD, and domestic APIs, while supporting five output formats: images, storyboards, comic panels, video, and 3D. The core is a precision control workbench for weighting and ratios, featuring a character anchor table, composition and shot references, pose-emotion mapping, a reference image capability matrix, multi-character spatial relationships, storyboard grid rules, speech bubble positioning, LoRA management, a temporal consistency checklist, a 3D generation matrix, video prompt templates, conversational editing workflows, negative prompt libraries, and a three-level P0/P1/P2 verification checklist. Maintained by the community publisher user_2cbe9405 (Version 2.8.0) and categorized as "design-media" on SkillHub, it does not strictly require an API key. As the official installation command is not provided in the materials, actual results require independent verification.
Read Moredsh-bili-wallpaper: Using Bilibili Live Streaming or Video as the DSH Web GUI Background
The st21a1-maintained DSH fun skin plugin overlays Bilibili livestreams or on-demand videos as the background behind the chat interface in desktop/web profiles. It supports QR-code login, resolution selection, blur and mask adjustments, multi-part (P) support, and progress memory; a host-side proxy handles Referer and HLS issues. Recommended installation: dsh plugin --profile web add dsh-bili-wallpaper. Requires Node.js >= 22.12.0 and DSH ^0.2.0-rc.1. Credentials are stored locally, with no additional npm runtime dependencies.
Read MoreFully Controlled C Drive Space Saving, Cleanup, and Migration: Putting C Drive Cleanup and Migration into a Deterministic Process
By pre-packaging complete workflows, read-only scanning scripts, and safety rules, this Skill reduces the number of interactive rounds required for Windows C drive cleanup. It also supports migrating data for large-scale applications, such as WeChat and WeCom, from the C drive to the D drive. The process emphasizes item-by-item confirmation, resource release following backup verification, and reversible deletion via the Recycle Bin, while utilizing Junctions to ensure programs can access original paths. According to the official description, this approach compresses a single cleanup session from over 15 rounds to approximately 5, reducing Token consumption by about 90%.
Read MoreVideo Generation Skill: Rendering HTML Assets into Promotional Videos with HyperFrames
The Source-to-Video (source-to-video) skill is a community Agent Skill based on HyperFrames, identified as @user_8bae2f22/source-to-video. Classified under design-media and currently at version 1.0.1, it accepts raw inputs such as HTML pages, workspaces, and landing pages. It covers the entire workflow, including material understanding, design token extraction, composition, Chinese voice-over, and rendering. By default, it produces a 30-second, 1920x1080@30fps promotional video with narration, replacing manual screen recording and editing processes. The label requires_api_key=false indicates that no additional API key configuration is necessary for invocation. Use this skill when users provide HTML files and request video or promotional content, or when referencing Hyperframes. Installation commands and platform-specific configurations are not fabricated here; please refer to the official page for detailed information.
Read MoreDSH My Phone: Accessing DSH on Your Home Computer via Mobile Browser
DSH My Phone (ssy-in) is a community plugin for DSH clients. It acts as a pairing and proxy entry point on the local machine. Users can access the native DSH interface by scanning a code with a mobile browser to establish trust. It supports LAN and same Wi-Fi access, with an optional Cloudflare fast tunnel for external network access. The installation command is `dsh plugin add dsh-my-phone --profile desktop`. The management page is accessible only locally at `127.0.0.1:8080/__myphone/`. Based on the GitHub README, this article introduces the pairing workflow, `state.json` configuration, security principles, and troubleshooting paths. It also highlights that entry permissions equate to local DSH capabilities, and LAN or tunnel features should be disabled when not in use.
Read MoreAI Large Model Expert | FLUX Image Generation Alternative | AI Image Generation and Editing: Fulfilling Text-to-Image and Reference Image Editing with a Single Skill
This Skill combines text-to-image generation and reference-based image editing into a single call. Designers, retouchers, and e-commerce visual artists can provide direct text descriptions and attach a reference image when control over the subject or style is required. The Skill automatically handles the entire workflow on AI Hive, including reference image upload, task submission, progress monitoring, and result download. AI Hive is part of the product ecosystem of Beijing JiRui Technology Co., Ltd., established in 2017. Model integration supports asset uploads, real-time model configuration, and price snapshots, offering three routing strategies: COST_FIRST, SPEED_FIRST, and SUCCESS_FIRST, which prioritize cost, speed, and success rate respectively. Officially marked as `requires_api_key=true`, a valid key must be configured before invocation. Use cases cover e-commerce main images, product detail pages, advertising key visuals (KV), posters, sales and promotional imagery, product retouching, background replacement, and character consistency.
Read MoreGooseHost Static Website Hosting: Registration, Login, Site Building, and Deployment in Skill
# Introduction Deploying a static site typically involves registering an account in a browser, logging in, creating a new site, and uploading files—operations scattered across various pages. For developers already utilizing Agent Skills, consolidating these steps into a single conversational command allows for a seamless process from registration to deployment without leaving the current environment, thereby saving significant back-and-forth. The following is an introduction to GooseHost Static Site Hosting, a community project that encapsulates the static site hosting workflow as a Skill. # What is this GooseHost Static Site Hosting is an Agent Skill based on the universal SKILL.md format, identified as @user_a3ed06bc/goosehost. It is maintained by the community publisher user_a3ed06bc, and the current version
Read MoreBusiness Modeling Expert: Handing Enterprise Capability Inventory to a Model's Skill
Business Modeling Expert is a community-published Agent Skill located at @user_ce145526/ba-capability-051, version 1.0.0, categorized under business-ops. It requires no API key. This skill targets the systematic identification, analysis, planning, implementation, and optimization of enterprise and organizational capabilities, making it suitable for capability assessment, process mapping, and reviewing existing outputs. Based on official metadata gathered, this article covers its positioning, verified metadata, activation notes, and applicable use scenarios. The installation command has not been verified due to the inaccessible page.
Read Moredsh-model-behavior: Injecting Editable Behavior Prompts by Provider and Model
dsh-model-behavior is a DeepSeek Harness community plugin (ssuper1, MIT 0.2.0) that appends editable behavior prompts after the agent request parses the real provider/model, without replacing the core system prompt. It supports grouping by provider, individual model selection within groups, multiple numbered prompts and per-model specific text, scope settings for main sessions/sub-agents, and inclusion/exclusion of model IDs. The settings panel allows previewing the final injected content. The installation command is `dsh plugin --profile <profile> add github:ssuper1/dsh-model-behavior`, and a desktop restart is required after installation.
Read MoreDSH Orchestrator: Hard-Isolation Scheduling with Strong Model Planning and Cheap Model Execution
DSH Orchestrator is a workflow-oriented DSH plugin maintained by SShiRkiT. Through a scheduling mode panel, it decouples Main Agent and sub-agent model routing, and uses tool gates to restrict the main agent to planning and acceptance only. It supports parallel delegation via `delegate`, batch pipelines via `orchestrate`, as well as zero-token session log inspection and `send_message`-based correction. Installation requires running `dsh plugin --profile desktop add SShiRkiT/dsh-orchestrator#main` and restarting the client. It is suitable for scenarios involving cost control, preventing the main agent from directly executing tasks, and long-running Workers requiring observability. Please review the source code and the MIT license before use.
Read MoreKarpathy's Perspective: Distilling His Mental Frameworks into Actionable Skills
"Karpathy Perspective" is a community Skill that distills Andrej Karpathy's mental frameworks. Coordinates: @user_00c9b356/karpathyperspective. Published by user_00c9b356. Version 1.0.1. Category: knowledge-management. Tag: requires_api_key=false (no API Key configuration required). The content is structured in three layers: 5 mental models (Ghosts vs Animals, Minimalism/Constructionism, Verifiability Framework, Computational Systems Metaphor, Technological Empowerment), 10 decision heuristics, and a complete "expression DNA." Additionally, it includes 7 ideological turning points that trace the trajectory from CS231n in 2015 to joining Anthropic in 2026. The material claims to be based on over 60 primary sources (blogs, GitHub, podcast interviews, Twitter/X). This Skill is request-triggered; it activates only when a user explicitly requests an analysis from "Karpathy's perspective/way of thinking/framework." Typical phrases include "use Karpathy's perspective," "what would Karpathy think," or "from Karpathy's standpoint." It will not activate for mere name mentions without a request for perspective, purely factual inquiries, or daily conversations unrelated to AI/Deep Learning. Given that both the official page and the SkillHub page failed to scrape during the drafting of this article, no unverified installation commands or configuration fields are provided; only confirmation that it follows the standard SKILL.md format is given. Typical usage involves providing specific technical issues while indicating the dimension of judgment, such as using its framework to analyze technical routes, evaluate team engineering habits, or check unverifiable assumptions in a plan. Note that this is a distillation of a thinking framework rather than a database of statements; it is not suitable for looking up original quotes, and conclusions naturally carry the bias of the framework itself. The official address is found in the main text.
Read Moredsh-task-groups: Task Grouping and Activity Reminders for the DSH Sidebar
dsh-task-groups is a DSH client plugin maintained by spike237 that provides custom task grouping for the sidebar (naming, color coding, drag-and-drop sorting, creation within groups, and bulk mark as read/archive) along with an activity bell panel that centrally displays running, pending confirmation, and unread conversations. Installation requires DSH 0.2.1+ and Node.js 20+, and can be done via `dsh plugin --profile desktop add dsh-task-groups`; state is stored in DSH_HOME/task-groups/state.json. Project view switching is preserved, licensed under MIT.
Read MoreChinese Document Formatting Cleanup: Unifying Full-width/Half-width Characters, Punctuation, and Paragraph Spacing
Introducing the “Chinese Document Formatting” skill on SkillHub: published by a community author, user_ae3bd8e0; categorized under office-efficiency; version 1.0.0; no API key required. The article describes how it targets Chinese document layout tasks—unifying full-width and half-width characters, fixing punctuation spacing, and standardizing paragraph spacing—and truthfully notes what has been verified, as well as installation commands and usage details that have not been verified, along with applicable scenarios.
Read MoreWeChat Official Account Publishing Assistant: Assistance for Layout and Publishing of Industry Research Articles
The WeChat Official Account Publishing Assistant is a content-creation Skill published by Queqiao Technology on SkillHub (Version 1.0.0, Coordinate: @queqiaotech/wechat-service-publisher-pro, Tag: requires_api_key=true). Tailored for the publishing pipeline of industry research articles, it supports original text reading, hierarchical summarization, sorting of industry chain targets, image2 poster and cover generation, body text formatting, saving to WeChat drafts, readback verification, and pre-publication confirmation. The retrieved materials do not provide copyable installation commands or configuration fields; the official page and coordinate must serve as the basis for activation. As specific models, interfaces, image parameters, and draft authorization methods are unspecified, users should confirm the API Key and refer to the official page prior to use.
Read MoreChemical Identification and Data Extraction: A Spreadsheet for Inputs to Downstream Engineering Calculations
This article introduces the SkillHub community Skill: "Chemical Identification and Data Extraction." This Skill identifies chemicals involved in projects against seven major regulatory catalogs and performs pressure vessel toxicity classification (HG/T 20660-2017). It outputs a classification identification table, a 16-column chemical property table, and a 17-column hazardous chemical characteristics table (Table 3.1-1 from the Design Special Section on Safety Facilities). These are mandatory forms for safety assessments and serve as the primary data source for seven downstream Skills, including electrical explosion-proof selection, gas alarm value setting, general layout fire classification, HAZOP, process package design, pressure vessel design, and material balancing.
Read Moredsh-token-dash: Clearly View Token, Budgets, and Subscription Quotas in DSH
dsh-token-dash is a DSH Web plugin maintained by spartawhy117. It consolidates pay-as-you-go balances and monthly budgets from seven providers (including DeepSeek), as well as subscription quotas for GLM and Kimi, within input box pills, overlays, and full-page dashboards. It provides local cost estimation and real token counting, supporting time ranges from the current day to full history, along with JSON export and import. Installation requires DSH 0.2.1 or higher; the recommended command is `dsh plugin --profile desktop add dsh-token-dash`. Credentials are retrieved via the host's credentials system, and data is stored in `$DSH_HOME/dsh-token-dash/`.
Read More