Skill for Generating Chapter-based Practice Questions for Cost Engineer Question Bank · Module: Technology and Metrology
Maintained by SkillHub community publisher user_af163eaa (Version 1.0.1), this Skill is designed to generate practice questions for the "Construction Engineering Technology and Measurement" section of the Cost Engineer exam. It produces exercises organized by chapter, covering four core modules: engineering structure, engineering materials, construction technology, and engineering measurement. While a specialized professional track can be specified within a general framework, the questions are generated by AI based on a generic exam syllabus. The developers explicitly state that this is not an official resource and does not predict specific exam content. This article verifies its positioning, capabilities, and application boundaries: it is suitable for developers of exam preparation applications requiring chapter-level practice questions; however, the accuracy of the answers must be independently verified, and it cannot substitute for actual past exams. Since the official page and interface address could not be accessed during the writing of this article, installation commands, configuration fields, and usage examples are omitted. No speculative information is provided; instead, the article presents only verifiable details (e.g., classified under education, no API Key required, location identifiers) and directs readers to the official page to verify the enabling method.
Read MoreYunsike (aicloud-thought-proxy): Web AI Generates Plans, Local Agent Executes
# Introduction Calling Large Language Model (LLM) APIs is billed on a per-token basis. In a standard "plan first, then code" task, both the reasoning during the planning phase and the generation during the execution phase consume tokens, resulting in significant API costs. **Yunsike** (aicloud-thought-proxy) adopts a different division of labor: the AI in the browser interface handles planning, reasoning, and strategy development, while the local Agent is responsible solely for execution. The following is an introduction to this Skill. ## What is This Yunsike is a community-maintained Skill published by user_0611a1f0. It is categorized under ai-agent, is version 1.0.1, and does not require an API Key (requires_api_key=false). Its page on SkillHub
Read Moresleep-not-permitted: DSH Agent foreground waiting prohibited
sleep-not-permitted is a DeepSeek Harness profile-level plugin maintained by JJJACKsong. It rejects foreground sleep, tail -f, long wait_agent calls, and polling via system prompts and ctx.tools.guard, while guiding the use of run_in_background or terminating the current turn to await completion notifications. This article introduces the interception scope, the default 30-second blocking budget, interactive command guards, and the cordis.patch.yml configuration. It also covers the tarball installation and verification steps found in the README, and explains boundary conditions such as the lack of stdin for background jobs and the fact that completion notifications do not necessarily imply success.
Read MoreReview of DX-P3L3-02 PI-Level KOL Presentation: Segmented Review Skills for a Dual-Dimensional Pharmaceutical Matrix
This article introduces the DX-P3L3-02 PI-tier KOL presentation-script review and editing skill released on the Baitong AI service platform. The skill is positioned within a dual-dimension healthcare–enterprise matrix: P3 marketing promotion × L3 departmental decision-makers. It is currently at version 2.2.5, originates from enterprise, is categorized as professional, and does not require an API key. The article outlines its positioning, verifiable core functions, how to enable it (by referencing it via the coordinate @baitongai/dx-p3l3-02-pi-ji-kol-jiang-gao-shen-jiao; no installation command has been officially published), typical usage patterns, and applicable scenarios. It also cautions that this is a niche single-point skill with a small public data sample, and recommends validating the output with sample presentation scripts before integration.
Read Moredsh-plugin-2048: Play Full-Page 2048 in the Harness Main Area Without Switching Tabs While Waiting for the Agent
`dsh-plugin-2048` by JJJACKsong is a sidebar mini-game for the DeepSeek Harness Web UI. It features a full-page 2048 game in the main area, supporting control via arrow keys, WASD, or swipe gestures. It allows up to 30 steps of undo and persists game state in `localStorage`. The Host acts as a minimal shell and does not write to sessions or workspaces, requiring `dsh` >= `0.1.5-rc.2`. It can be installed by searching for "2048" in the Open Plugin Market or from a local directory using the Plugin Manager or the `dsh plugin add` command. Licensed under MIT 0.2.0, it is categorized as an entertainment plugin.
Read MoreBaoBiao · Software & AI Procurement Eye: Bid Information Query Skill for Software and AI Projects
软件与AI招采天眼 is a Skill in the business-ops category on SkillHub, published by 金里程信息, with underlying data provided by the 世舶招投标数据 API. It currently supports querying bidding lists and essential details, focusing on software, AI, cloud computing and digitalization projects. By default, it queries the last 30 days with 10 items per page, and the analysis mode defaults to the last 6 months. This article introduces its positioning, core features, prerequisites, typical use cases and usage limitations, including the requirement of an API Key, the disclaimer that it does not claim to cover all data categories, the caveat that keyword matches do not equal confirmed business opportunities, and that monetary amounts and regions are output as provided by the API without further normalization. It is suitable for sales, pre-sales and developers who need to regularly track software and AI project announcements.
Read Moredsh-token-panorama: Tracking Session Token Flow and Costs in DSH
dsh-token-panorama is a DSH client plugin maintained by jiuwu02 that reads logs from the platform's session persistence service to display cumulative usage, cache hit rates, peak-valley costs, heatmaps and trends, model and workspace distributions, and a collapsible session tree within the Token Panorama panel, supporting CSV/JSON export and incremental refresh. It also offers Composer context Clear/Compact and input box prompt Optimize. Requires DSH >= 0.2.1-rc.3 and a Web UI profile; install locally via plugin_manager install_bundle.
Read MoreWedding Planner: An Agent Skill for Chinese Wedding Preparation
The Wedding Butler is an Agent Skill based on the SKILL.md format, located at @user_4e23c311/wedding-butler, version 2.0.0, sourced from community, categorized under life-service, and published by user_4e23c311. It targets Chinese-style wedding preparation scenarios, triggered when users mention topics such as wedding planning, wedding budget, bride price and dowry, wedding banquet hotels, wedding planning companies, the four essential wedding vendors (photographer, videographer, makeup artist, and emcee), wedding photos, wedding attire, wedding process, bride pickup tradition, tea ceremony, wedding vows, invitation text, best man and bridesmaids, wedding speeches, pre-wedding anxiety, and wedding preparation disputes, or when uploading wedding-related materials. It provides planning advice and emotional support. It is not applicable for marriage registration legal procedures consultation, post-marriage relationship counseling, or matchmaking and dating guidance, nor does it recommend specific merchants or brands or substitute for user bookings and payments. The tag requires_api_key=false, enabling use without an API Key. The captured materials do not include the SKILL.md body text or installation commands; the official page should be verified before integration.
Read MoreMeditation Guide Skill: Translating Neuroscientific Evidence into Actionable Meditation Instructions
The Meditation Guide is a community-focused educational Agent Skill on SkillHub, identified by the coordinates @user_00c9b356/meditationguide (Version 1.4.0, published by user_00c9b356). Based on evidence from neuroscience research, it explains how meditation alters the brain and recommends suitable meditation types according to user needs (such as stress reduction, creativity, and sleep). It also distinguishes meditation from mind-wandering, flow states, and Taoist Qigong, while providing actionable practice steps. Trigger topics include "what is meditation," "how to meditate," "meditation for stress relief," "meditation for creativity," and "the differences between meditation and mind-wandering/flow." The skill requires no API Key, offering a low barrier to entry. Its capability boundaries are clearly defined: it does not write code for meditation apps, recommend meditation music, or sell meditation products. This article could not verify the official installation command (as the official page scraping failed), and this has been noted truthfully.
Read Moredsh-image-wheel-pan: Right-side Preview Image Scroll-to-Zoom and Left-click Drag-to-Pan
dsh-image-wheel-pan, maintained by jinmuxi1, is a DSH Web client plugin that adds Ctrl-free mouse-wheel zooming (approximately 1%–800%), left-button drag panning, double-click or “Fit” to fit the image to the pane, and 100% actual-size viewing for PNG/JPEG/GIF/WebP/BMP/ICO/SVG images in the right-side file preview. Install the npm package dsh-image-wheel-pan via the Plugins panel or paste the GitHub repository URL, then refresh the Web UI. It is automatically applied to newly opened image tabs; existing tabs can be switched via the preview header dropdown. The plugin only reads already-loaded image bytes, is released under the MIT license, and reviewing the source code before installation is recommended.
Read MoreIntroduction to the Financial Knowledge Base Skill: A Full-Process Skill for Knowledge Bases in the Finance and Insurance Verticals
The financial knowledge base is a community-sourced Skill in the SKILL.md format, with coordinates at @user_6399968a/finance-kb, categorized under knowledge-management, version 2.0.2, and tagged with requires_api_key=true, requiring an API Key for use. According to its official introduction, it covers the entire process of financial knowledge bases, dedicated scenarios for financial insurance, compliance, and metrics, supporting human-in-the-loop operations and traceable deliverables. This article only introduces its positioning and verifiable capabilities based on the retrieved materials. The official page is currently inaccessible, and the materials do not provide installation commands, configuration fields, or reproducible examples, so these are not included here—only its coordinate, category, and Key requirement are stated. It is intended for teams working on financial insurance-focused knowledge bases that need to incorporate compliance and manual review into their delivery workflows. Note that as a community-sourced resource, its maintenance status should be independently assessed before integration.
Read MoreProject Planner: Turning Ambiguous Goals into Actionable Project Plans
Project Planner is a community Agent Skill categorized under "office-efficiency," maintained by user_8ee65e73. Based on the universal SKILL.md format, it is currently at version 1.2.4. It converges vague objectives into project plans featuring milestones, Work Breakdown Structures (WBS), risk registers, and schedules. Its outputs cover five categories: project plans, WBS, milestones, risks, and scheduling, while explicitly excluding coding and implementation. A key aspect of its design is forced routing: the Skill is activated immediately when users mention keywords such as "project plan," "WBS," "milestones," "task breakdown," or "roadmap," or when they directly paste, upload, or link raw text or screenshots. It does not initiate with clarifying questions or substitute with generic capabilities. Regarding configuration, it is tagged with `requires_api_key=false`, meaning no additional API key is needed for invocation. Official documentation does not provide specific installation commands, offering only coordinates and page addresses. In terms of usage, it is suitable for scenarios such as project initiation and aligning milestones and risk definitions across teams. It is important to note that it is not an execution tool; integrators must configure trigger rules correctly within the scheduling layer. The text also highlights an inconsistency regarding ratings and download counts: page scraping indicated a rating of 100,000 and 870 downloads (with 0 installs), which conflicts with the figures of 99,887 and 845 found in task clues. As two attempts to scrape the official page failed, real-time page data should be regarded as the authoritative source when citing.
Read Moredsh-ctrl-zoom: Ctrl + Scroll Wheel Zoom for Harness Web Interface
dsh-ctrl-zoom is a DSH client plugin maintained by jinmuxi1 that lets you zoom the entire Harness Web UI between 50% and 200% in 10% increments using Ctrl/⌘+scroll wheel, +/-, or Ctrl+0. The zoom level is persisted in localStorage and displayed as a pill in the bottom-right corner. Install it via Plugins using the package name dsh-ctrl-zoom, or with plugin_manager install_bundle. Built on CSS zoom, it makes no network requests and is licensed under MIT 1.0.0.
Read MoreRectification of Typesetting for Rational Expressions and Radicals in Junior High School Mathematics: Standardizing Fractions and Radicals in Lecture Notes and Exam Papers to Textbook Format
This is a typesetting correction Skill for middle school mathematics handouts, lesson plans, and exam papers, located at @user_3327fbbe/math-formula-standard-fix, maintained by community publisher user_3327fbbe, version 1.0.1, category education, source community, tagged requires_api_key=false. It standardizes irregular fractions and quadratic radicals in the text to the standard print format of the 2026 People's Education Press Grade 9 mathematics textbook, adjusting only the typesetting of mathematical symbols while retaining all original question stems, examples, solution steps, answers, and lesson structure. It is suitable for teachers and teaching researchers who need to batch standardize mathematical symbol typesetting before printing for use. Limitations: its capability scope is limited to fractions and radicals in middle school mathematics texts, with the standard based on the 2026 People's Education Press Grade 9 textbook, and it does not correct textual or numerical errors beyond typesetting. The official page does not provide verifiable installation commands and configuration fields, and no speculation is made in the main text; SkillHub metadata shows a rating of 100000 and 949 downloads.
Read MoreIntroduction to the AI Legal Advisor Skill
The AI Legal Consultant is a general-purpose SKILL.md skill published by community developer user_89b96750 on SkillHub, located at @user_89b96750/legal-consultant, version 1.0.0, categorized as professional, with the tag requires_api_key=false. It encapsulates three types of legal tasks—contract review, legal consultation, and document drafting—into a Skill with a fixed role, preventing inconsistent standards caused by ad-hoc prompt construction each time. The resource has a rating score of 100,000, 972 downloads, and 0 installs, with the source being community. Therefore, before integration, it is necessary to verify parsing and output quality in the target environment. The article does not provide installation commands or reproducible examples, as none were provided in the official resources, and emphasizes that model outputs should only serve as drafts requiring manual review and cannot replace the advice of licensed attorneys.
Read Moredsh-edit-rerun: Edit prompt history and rerun in branched conversation
`dsh-edit-rerun` is a DSH client plugin maintained by jinchenwu1220. It places a rerun button next to the input box that lists the actual user prompts from the session. Upon editing, it pre-fills a draft in a new branch via `sessions.fork` and reruns it, ensuring the original session's append-only log remains unmodified. The article covers installation commands, usage steps, limitations (such as the inability to rerun the first turn), and its distinct positioning compared to `dsh-plugin-message-edit`. It requires DSH 0.2.x and Node 22+; after installation, it must be added to the profile bundles and the application must be restarted.
Read MoreWeRead Co-reading: Turning Highlights into Thoughts, and Thoughts into Output
WeRead Reading Companion (formerly “WeRead · Idea Coach”) is an educational Skill in the SkillHub community, published by user_6b4bda16 (v1.10.0). It has 100000 ratings, 990 downloads, and requires an API Key. When it starts, it greets you and pops up a thought card (where you left off last time, the question to carry forward, and a line to chat about the spot where you paused). It automatically tracks the chapter you’re reading and asks only one question at a time, turning your underlines into ideas you can probe further. By default it produces a 200–300-character thought card, and only produces a longer piece when you call on it. When asked to consolidate, it can also evolve into a cross-book knowledge base. When you revisit a book, it returns your underlined passages exactly as they were. All outputs sync to the Reading Companion Notes folder in your iCloud “Reading” library.
Read MoreIndustry Research Assistant: Lifecycle-Based Industry Analysis Skill
The Industry Research Assistant is an industry analysis Skill designed for a general SKILL.md environment, located at @user_6e1f577b/leo-industry-research. It is version 1.0.0, sourced from the community, and classified as professional. The 'requires_api_key=false' tag indicates that no additional API Key configuration is needed for its use. The tool first determines the lifecycle stage of the industry, then dynamically adjusts the analytical focus across seven dimensions based on that stage: feasibility, scalability, defensibility, profitability, valuation, external factors (PEST), and industry prosperity. All information sources are attributed using a four-tier information grading method. It is activated via a single natural language command, for instance, 'Analyze the new energy vehicle industry'. This article describes its positioning, the seven-dimension framework, the significance of source grading, installation instructions, and typical use cases. It also clarifies its boundaries: it functions as an analytical framework rather than a data source, its determination of the stage relies on the quality of the input materials, and the specific definitions of the four-tier grading levels must be referred to the SKILL.md file within the package.
Read MoreEnterprise Information Security AI Advisor Capability Library: A Single Entry Point to 28 Knowledge Areas and 56 Reference Documents
**Enterprise Information Security AI Advisor Capability Library 🐉** This is a general-purpose SKILL.md format capability library tailored for enterprise information security scenarios. Maintained by community publisher user_2354702d, it is currently at version 1.0.1. The source is "community," classified under "it-ops-security," and tagged with "requires_api_key=false". The library decomposes enterprise security into 28 knowledge areas supported by 56 reference documents, covering scenarios such as security system construction, MLPS 2.0 compliance, security architecture design, defense-in-depth, cyber range exercises, SDL/DevSecOps, supply chain security, security training, the "Four-Diagnosis" security consulting method, risk assessment, security incident response, SOC construction, Zero Trust, data security, cloud security, AI security (prompt injection/red teaming/LLM architecture/compliance), IoT security, IAM, mobile security, threat intelligence, security metrics, ISMS documentation, security culture implementation, security project management, and security architecture review. Its core value lies in consolidating scattered reference materials into a single entry point, thereby reducing the cost of reassembling materials before initiating security tasks. It provides reference frameworks and documentation rather than ready-to-deploy security solutions; all generated outputs require human review.
Read Moredsh-coding-discipline: Imposing "Verify First, Small Changes, Evidence-Based" Work Constraints on DSH Coding Tasks
dsh-coding-discipline is a DeepSeek Harness workflow plugin (v1.0.0, MIT) maintained by JinBeiCN. It provides a Coding agent preset with the id `coding` (same tools as `standard`, reinforcing verify-first and minimal-diff practices) along with an on-demand loadable `coding-discipline` skill. To install: `dsh plugin --profile web add dsh-coding-discipline`; in a new session, either select the Coding preset or load the skill from the skill directory. According to the README, on a 16-task benchmark its solved rate matches the baseline at 15/16, and it is positioned as a guardrail rather than a capability booster. Requires dsh >= 0.2.1-rc.2 and Node >= 18.
Read MoreQiuye Workplace: Tailoring Enterprise Training Cases into a Lesson Prep Board with WorkBuddy
WorkBuddy Lesson Prep_Autumn Leaf Workplace is a skill in the education category on SkillHub, published by community developer @user_c774f8e9 (version 1.0.2). By inputting customer information and an attendee list, it generates a Skill case table (approximately 10 entries, including name, function, suitable positions, teaching demonstration points, and development prompts), an expert setting table, and expert group configurations, all consolidated into a single-file HTML lesson preparation dashboard. The skill tag requires_api_key=false, indicating that no API key is needed. It currently has a rating of 100000, 1017 downloads, and 0 installations. Given the low installation count, pilot validation is recommended. The text introduces its core features, conversational usage, and applicable limitations. The official URL is https://api.skillhub.cn/user_c774f8e9/workbuddy-lesson-prep.
Read MoreBaZi Astrology Analysis Skill: Delegating Charting Rules to External Tools
Introducing the “Bazi Charting & Fortune Analysis” skill on SkillHub (coordinates @org-bzwfrdvh/bazi-skill, version 1.0.4, published by AIPDD). It collects the user’s name, birthday (Gregorian or lunar), birth hour, gender, and birthplace through interactive steps, then produces the Four Pillars and Eight Characters, the luck cycles (dayun) and annual cycles (liunian), auspicious/inauspicious spirits (shen sha), the life palace (minggong), and the fetal origin (taiyuan). Based on nine classics of Ziping Bazi and the perspective of Ni Haixia, it performs a comprehensive analysis of the Day Master’s strength or weakness, pattern/structure, elemental balance (tiaohuo), the Ten Gods, the six relations, health, career, and relationships, helping to avoid errors from general models attempting to chart the bazi themselves.
Read Moredsh-session-keyword-search: Cross-session substring keyword search
The built-in FTS5 (unicode61) in DSH is not well-suited for searching contiguous Chinese substrings. dsh-session-keyword-search, maintained by JimmyRowe13, takes over the sidebar search box with a locally-stored SQLite substring index, supporting multi-word AND queries, quoted phrases, result highlighting, and post-navigation target location. Incremental indexing is performed via the public sessionQuery API, with graceful degradation on failure. This article covers how to locate the plugin, installation commands, configuration, and the usage of verify/status, while also addressing plaintext indexing, permissions, and known limitations.
Read MoreHazard Identification Skill: Structuring Hazard Analysis of Work Activities into Systematic Outputs
Hazard Identification (@user_8dd18822/hazard-identification-skill, v1.0.0, Community Source, classified as professional) targets the production and operational activities of business entities. It consolidates the classification of hazardous and harmful factors, accident type determination, identification and grading of major hazard sources, and the output of the "Hazard Analysis Form" and "Hazard Identification Report" into a single, fixed entry point. Trigger words cover hazard identification, hidden danger screening, HAZOP, JHA, SCL, as well as standards GB 6441, GB/T 13861, and GB 18218. It is applicable to scenarios such as safety assessments, hidden danger screening, dual prevention system construction, work safety standardization, emergency plan formulation, and team-level safety training. It operates with `requires_api_key=false`, requiring no additional keys. The article is structured according to stylistic requirements: "What is this," "Core Functions," "Capability Boundaries & Input Requirements," "Installation & Activation," "Typical Usage," "Applicable Scenarios," "Notes," and "Conclusion." Regarding the captured data: installation commands, directory paths, configuration fields, and official reproducible examples were not provided in the retrieved materials. Furthermore, official pages and interfaces failed to be retrieved during this session. Therefore, these sections are honestly marked as "not provided," without fabricating commands or fields. Data points such as a score of 100000, 996 downloads, and 0 installations are based on the retrieved materials, with a prompt to conduct small-scale validation before integration. The conclusion states that this tool is suitable as a basis for material drafting and preliminary screening. However, final responsibility determination still requires manual review in accordance with current valid standards.
Read MoreAI Advisor for Construction Law Cases: Consolidating Legal Research and Document Drafting into a Single Skill
Addressing the pain point of language models fabricating legal provisions and case numbers in construction engineering legal consultations, this article introduces the community Skill "Construction Law & Case AI Advisor". Automatically triggered by topics such as construction contracts, tendering and bidding, affiliation, actual constructors, and priority right of compensation, it supports searching statutes and precedents, drafting complaints and defenses, and compliance health checks. It enforces prioritizing original texts from the local knowledge base "Practical Handbook for Courts in Handling Construction Engineering Cases" and prohibits fabricating case numbers or legal provisions. The article provides the coordinates for enabling it, typical use cases, and applicable limitations.
Read More