DSH Little Whale Accounting Widget: Web Interface Balance and Customizable Bubbles
An introduction to the DSH bundle plugin `dsh-whale-widget`, maintained by ggxx001: it resides in the bottom-right corner of the Web interface, displaying DeepSeek balance and today's usage, peak/off-peak pricing, and per-round conversation consumption. It also supports modular bubble click sequences, styles and sound effects, docking and flipping, and multi-vendor balance templates. It provides the officially recommended installation command for the web profile and typical configuration paths, and explains the differences between this and the Codex product line, as well as precautions to review the source code and permissions before installation.
Read MoreOperational AI Assistant: A Universal AI Skill for Operational Scenarios
"Operations AI Assistant" is a general-purpose skill in SKILL.md format published by Kunlun Growth on SkillHub, located at @kunlungrowth/kg-biz60-022. This skill, categorized under business-ops (Version 1.0.0), requires no additional API keys. It is activated by trigger words such as "operations," "management," "supply chain," and "business" to standardize the handling pathways for operational issues, thereby minimizing the randomness in responses. This article describes the skill's positioning, verifiable capability boundaries, client-based activation methods, trigger-based verification techniques, applicable scenarios, and precautions.
Read MoreCycling Health Assessment: Integrating Physical Examination and Riding Data into Persistent In-depth Reports
The Cycling Health Assessment is a community Skill on SkillHub (v1.0.15, data-analysis), published by user_c660c653, with no API Key required. It integrates personal physical testing data (height, weight, body fat percentage, waist circumference, resting heart rate) with bike computer, watch, and smart trainer data (heart rate zones, cadence, distance, elevation gain, or power meter FTP/TSS) to generate an in-depth HTML assessment report featuring real calculations, visualizations, and actionable recommendations. The report can be browsed and saved locally. Its core value lies in using real data to identify misconceptions, such as training extensively yet still having high body fat, believing you are riding casually while actually straining in the threshold zone, or low cadence posing joint risks. The Skill also includes a built-in, purely front-end, offline data collection form. Once cyclists complete it, a self-contained data card with screenshots is generated, enabling analysts to standardize data collection and trace conclusions. The article introduces its positioning, core capabilities, installation entry point, typical usage, and applicable limitations; no installation commands beyond those provided in the material are given.
Read MoreZhengyan Patent Artisan: An End-to-End Workflow for Patent Mining to Disclosure Drafts
Zhengyan-patent-drafting-craft is a professional community Skill available on SkillHub, published by user_1567453e. The page version is 1.5.0, with a rating of 100,000, 1,073 downloads, 0 installations, and requires_api_key=false. Positioned as a patent drafting production pipeline, it enables one-click activation of the entire workflow—from technical mining and claim design to the generation of disclosure documents (.docx). It covers project skeletons, T1–T12 disclosure writing, terminology generalization, drawing self-inspection, quantitative scoring of grant barriers, novelty searches and OA response strategies, review defense, value assessment, and hierarchical matrix mining of R&D details, outputting both concise and comprehensive .docx versions. However, the documentation contains conflicting descriptions regarding headers, footers, cover fields, and formatting; therefore, sample verification is required before formal use.
Read MoreThesis Skill Bundle zjf: A Unified Entry Point for 18 Thesis Skills
The "Paper Skill Combo zjf" is a community-sourced collection of skills that consolidates 18 paper-related skills into a single entry point. It covers the entire workflow from literature search (CNKI, multi-source documents, arXiv) to submission and publication. This includes rapid paper reading, journal matching and selection, cover letter generation, seven-dimensional detection, three-round refinement, editorial perspective review, reference formatting, citation management, academic typesetting, academic translation, AI-detection-free polishing, and mechanism visualization. In conversations, it automatically matches corresponding skills based on keywords such as paper writing, submission, journal selection, polishing, references, translation, and AI-detection-free polishing. The combo itself does not require an API key. As of November 4, 2026, the official page shows a rating of 100,000, 1,090 downloads, and 0 installations, indicating that it is in its early stages. It is suitable for graduate students and researchers who need an all-in-one solution to handle every stage of academic paper processing.
Read Moredsh-pet: Raising a Transparent Animated Desktop Pet in DeepSeek Harness
dsh-pet is a DSH fun dress-up plugin maintained by the community directory ggxx001. It provides 97 transparent animation clips, drag physics, balance display, and closable murmur/dialogue features in both a Web interface and an optional Electron desktop mini-window. Based on the official README and package information, this article explains its positioning, core capabilities, installation command `dsh plugin --profile web add dsh-pet`, and typical operations, while noting considerations regarding model invocation, permissions, and Safari .mov assets.
Read MoreLegal and Regulatory Retrieval and Compliance Self-Check: A Verifiable Compliance Retrieval Method for AI
This community-published "professional" Skill addresses the pain points of unclear sources, inconsistent hierarchies of legal authority, and unverifiable regulatory validity encountered when AI addresses compliance issues. It covers four stages: hierarchy of authority, official sources, cross-validation, and conclusion output. The Skill requires the AI to first confirm whether the regulations are currently valid, identify their level of authority, and verify official sources before providing conclusions and self-check essentials. Labeled as `requires_api_key=false`, it does not rely on paid interfaces. The article outlines its positioning, capabilities, activation access, and usage limitations, while cautioning that it is a methodological Skill that does not contain its own regulatory database and that its output cannot substitute for professional legal advice.
Read MoreIn-depth Book Reading: An Orchestrator for Transforming Pure Text Books into Reusable Skills
Book Deep Reading is a community skill on SkillHub (@user_360c4d3a/book-to-deep, Version 1.0.0), categorized under ai-agent. It operates as an orchestrator running a two-phase workflow. Accepting user-prepared MD and TXT files and relying on nine user-confirmed book categories, Phase 1 constructs a wiki-style mapping of the book, while Phase 2 distills core mental models, methodologies, case patterns, and boundaries (including honesty boundaries) into ready-to-use skill artifacts. The workflow features step-by-step controls and review checkpoints, supporting retrieval and deep reading contemplation. When combined with question-generation skills, it can facilitate the creation of exam outlines, key topics, and test questions. This article covers its positioning, core features, installation and activation, typical usage scenarios, and usage considerations.
Read MoreExpert Negotiator: Translating Negotiation Theory into Actionable Diagnostics, Strategies, and Scripts
Introducing the community Skill "Negotiation Expert" on SkillHub. Based on frameworks including Harvard principled negotiation, BATNA, ZOPA, distributive/integrative negotiation, anchoring effect, TKI conflict modes, Cialdini's principles of influence, FBI tactical empathy, and game theory, it helps diagnose negotiation scenarios, formulate strategies, and generate templates and scripts when users mention keywords such as negotiation, consultation, price bargaining, salary negotiation, business negotiation, contract negotiation, procurement, conflict resolution, haggling, and related terms. The article outlines its positioning, theoretical foundations, trigger mechanisms, typical use cases, target audience, and limitations, and provides the official API endpoint and SkillHub page.
Read MoreImplementing Auto-play Bubble Mode for Voice and Video in WeChat Channels: Transforming Text into Audio-Video Documents
This is a community skill published on SkillHub, located at @user_29051052/voice-document-maker. Version 1.0.0, categorized under design-media, published by user_29051052, created by interccy, and tagged with requires_api_key=false. It addresses pain points in WeChat bot voice broadcasting, such as unknown voice file duration, poor format compatibility, and potential playback issues on mobile devices. The skill automatically converts text into TTS voice, pairs it with AI-generated background images, and synthesizes an MP4 video file. It is suitable for delivery via WeChat video cards, using the higher-dimensional format of video cards combined with audio to resolve issues associated with pure voice playback. The documentation provides the functional workflow and positioning but does not include specific installation commands, configuration fields, or invocation parameters; the text states this explicitly and defers to the official page for details. It is suitable for developers who need to broadcast text content via video cards on WeChat. Before use, users should verify the quality of the TTS voice, background images, and the maintenance status of the community source. The official URL is https://api.skillhub.cn/user_29051052/voice-document-maker.
Read Moredsh-bili-whale: Building an Agent That Browses Bilibili and Writes Diary Entries in DSH
dsh-bili-whale (0.4.6) is a DSH workflow plugin maintained by ggxx001: within scheduled time windows, it enables agents to read Bilibili videos (subtitles, danmaku, frame extraction, etc.), write dual diary entries, and push notifications to the owner via @ mentions in comments or private messages with limited quota. It provides a "fish-raising" settings page, dependency verification, QR code or Cookie login, with configuration stored in bili-whale.config.json. Installation requires Node 18+ and ffmpeg, added via dsh plugin link with DSH restart; sessions need ungrouped bili-whale with full access granted. Default risk control is conservative, single accounts can use @ sharing, while dedicated private message channels require mutual following between two accounts.
Read MoreIntermediate Economist Question Bank · Human Resource Management Module: Chapter-based Question Generation & Automatic Wrong Question Aggregation
This is a Skill tailored for candidates preparing for the Intermediate Economist (Human Resources) exam, published by community publisher user_af163eaa on SkillHub. Operating at version 1.0.1 and categorized under education, it is completely API Key-free. It generates chapter-wise Multiple-Choice, Multiple-Select, and Case-Analysis Questions for the "Human Resources Professional Knowledge and Practice" exam, with answers and detailed explanations attached to each question. Furthermore, it incorporates a self-testing mode and automatically logs missed questions into an error notebook, supporting re-practice as well as Word and PDF exports. The practice content itself is also exportable as Markdown or PDF. By addressing three core issues—chapter-bounded generation, exam-aligned question types, and lossless tracking of missed questions—it serves as a comprehensive practice aid. Please note that this is a practice tool, not a targeted prediction tool, nor is it officially affiliated. Its coverage is strictly limited to the Human Resources Management module. The quality of explanations depends on the underlying AI model; in case of any discrepancies, textbooks shall prevail. The reference material does not provide specific installation commands or details regarding differences in installation commands across different clients. It only provides coordinates and official addresses, so it is highly recommended to visit the official page to confirm the installation steps before execution.
Read MoreCapyWarm: DSH Original Capybara Emotional Companion Plugin
Released by GenXGSS, the DSH plugin CapyWarm utilizes three tools—capy_comfort, capy_praise, and capy_vibe—to provide bilingual (Chinese and English) comfort, celebration, and mood reminders featuring the original capybara character "Nuannuan" within conversations, supporting both ASCII and custom scripts. Install via the command `dsh plugin --profile web add github:GenXGSS/dsh-capywarm`, and configure languages or character names in `cordis.patch.yml`. Licensed under MIT with no affiliation to existing capybara IPs, it is designed to provide a lightweight enhancement to the developer experience within the Web profile.
Read MoreSoftware Requirements Analysis: Phased Requirements Gathering and Standard SRS Output
'Software Requirements Analysis' is a community-sourced dev-programming Skill, located at @user_0c3974d3/48491254512, version 1.0.0. It divides requirements collection into four categories: business, functional, non-functional, and interface requirements. Through staged interactive item-by-item input, it automatically organizes all information into a standard Software Requirements Specification (SRS) document. This Skill is suitable for solidifying interview content into a reviewable and acceptable requirements baseline during early project initiation. The Skill tag requires_api_key=false, requiring no API key configuration. The Skill has a SkillHub rating of 100000 and 979 downloads. Please note: Both the official webpage and API endpoint were unreachable during this verification. Installation commands, directory structures, configuration fields, and activation differences across various hosting environments could not be verified. In the original text, this is explicitly noted without fabrication. The quality of the final documentation still depends on the quality of input information during the collection phase; the four categories form a fixed framework, and special requirement dimensions require manual extension.
Read MoreShu Lie | BI Data Analysis Assistant: Turn Excel into Interactive Dashboards
"数猎 | BI Data Analysis Assistant - One-Click Excel Dashboards" is a data analysis Skill published by community developer user_10b0a9e9, version 1.1.5, categorized under data-analysis. It reads the column structure of Excel and CSV files, automatically designs charts and filters, generates interactive BI dashboards, and returns a no-login-required link, covering bar charts, line charts, pie charts, horizontal bar charts, detail tables, and other charts with filter linkage, suitable for BI analysis, operational analysis, sales analysis, business analysis, and other scenarios. The entire process uses a trial account, requires_api_key=false. The article introduces its capability scope, usage path based on the official description, applicable scenarios, and precautions, and honestly states that there was a failure in scraping the official page and that installation commands have not been verified.
Read Moredsh-qq-memory: Persisting QQ Conversations for qq-bridge and On-demand Memory Injection
dsh-qq-memory is a memory plugin written by Gentaine for DeepSeek Harness and qq-bridge. It syncs messages from social-v2.json to SQLite, injecting 1–8 previous conversations per turn based on keywords and time windows (with source.kind set to qq-table-memory). It tracks recall_count and turn_no, and performs periodic cleanup based on age, recall count, and per-turn interaction frequency. It does not invoke additional models or register new tools, making it suitable for scenarios that require locally retained, searchable QQ context with controlled table size. Installation is via profile link dependencies and the cordis.patch.yml configuration path, and requires DSH 0.2.x and Node.js >= 24.
Read MoreLiberal Arts Thesis Writing: End-to-End Paper Writing Skill for Core Journals across All Liberal Arts Fields
"Liberal Arts Thesis Writing" is an end-to-end Skill for writing core journal papers across eight disciplines: Literature, History, Philosophy, Law, Economics, Management, Education, and Journalism & Communication. It targets CSSCI, Peking University Core, SSCI, and authoritative domestic journals. Each discipline features a theoretical map, evidence standards, common pitfalls, and target journal tiers; Journalism & Communication specifically includes 24 in-depth theoretical templates. The system integrates a unified Chinese command entry, environment diagnostics, and network timeout/retry guardrails. The core workflow supports offline operation and includes auxiliary scripts for AI-pattern removal, percentage-based grading, reference verification, Crossref validation, and formatting or similarity checks. The entire process is governed by a 13-gate quality control system.
Read Moredsh-advanced-settings: Modify all plugin configurations in Harness using a VS Code-style page
dsh-advanced-settings is a DSH client plugin maintained by genius-alray. It provides a VS Code-style advanced settings page in the Settings panel, automatically generated from the Cordis Config schema of loaded plugins. It supports search and "Modified" filtering, with volatile fields saved immediately via mutate, and regular fields reloaded after confirmation through the configEditor. Secrets are server-side masked and writable only via loopback. It is installed via `plugin_manager install_bundle → dsh-advanced-settings` and accessed at Settings → Advanced settings. MIT-licensed; the README notes it has been tested on DSH 0.2.2-alpha.2.
Read MoreHigh-end Visual Design: Executable Design Specifications for AI Agents
"High-End Visual Design" is a design guideline Skill for AI Agents that provides professional design specifications and best practices via SKILL.md, constraining the Agent's output in visual design tasks and reducing stylistic inconsistency issues. It is published by SkillHub community publisher user_3c6cb52e, version 1.0.0, category design-media, requiring no API key. The page displays a rating of 100,000, over 976 downloads, zero installs, and limited community validation samples. The retrieved materials do not provide installation commands and usage examples; we faithfully present the information as is and make no extensions.
Read MoreYa Xian Zhi: Da Liu Ren Jin Kou Jue Intelligent Prediction System — Integrating Liu Ren Charting and Divination Texts into a Single Skill
Yiazianzhi · Da Liu Ren Jinkoujue Intelligent Prediction System (username: user_501ebc00; slug: yaxianzhi-jinkoujue; version 7.3.1; source: community; category: life-service) is a Skill for intelligent chart casting and divination for Da Liu Ren Jinkoujue, released by user_501ebc00, with requires_api_key=false. It is designed to address four pain points: automatically inferring when there are too many categories (33) to choose from; providing three-layer plain-language explanations when interpretive terminology is unclear; offering local history and multi-session comparisons instead of discarding after casting; and calibrating to the second using VSOP87 astronomical chart casting combined with true solar time when time input is inaccurate. The divination side includes a library of over 1,100 interpretive statements, 14 symbolic interpretation frameworks, and 23 algorithmic engines. It supports techniques such as Five Moving Lines, Three Moving Lines, Four Jue and Four Bai (four extinctions and four defeats), spiritual star linkage, void reversal interpretation, triple harmony bureau affairs interpretation, entering the tomb judgment, triple assembly bureau, 16 techniques for deriving affairs from four-position generation and restriction, continuous coriander (lian ru), celestial net and earthly net, monthly break and daily break, cutting the barrier, fu yin and fan yin, Sixty Jiazi Bell, Youdu and Ludu, host-guest victory or defeat, directional auspiciousness or inauspiciousness, current year and flowing year. Public materials only provide coordinates and metadata, without installation commands or configuration fields. This article accurately describes this situation and does not fabricate installation steps.
Read MoreCollege Major Planner Agent: Integrating RIASEC Interest Tests and Undergraduate Program Information
This planning-oriented Skill, designed for undergraduate majors in Chinese universities, is maintained by the SkillHub community publisher user_82fbf76f at @user_82fbf76f/major-advisor. Classified under education and sourced from the community (Version 1.1.0), it does not require an additional API Key. Using the Holland RIASEC interest test as an entry point, it encompasses three core capabilities: major selection advice, explanation of major content, and details on core curriculum and employment prospects. The Skill is configured to trigger automatically even if the user does not explicitly request a test or consultation. It covers a wide range of inquiries, including college entrance examination preferences, major selection, recommendations, interest testing, queries regarding curriculum or job prospects, major transfers, postgraduate directions, and questions asked by parents on behalf of students. The article is structured into sections on positioning, core functions, installation and activation, typical usage, applicable scenarios and precautions, and a conclusion. Due to a failure in fetching the official page and interface during this session, specific installation commands and differences between editors were omitted; instead, the SkillHub page and metadata interface addresses are provided, with a recommendation to consult the SKILL.md file before integration.
Read MoreGame Character Design: Generating Character Settings and Four-View Drawings at Once
Game Character Design is a design-media Agent Skill, located at @org-kanlf9wn/okflow-game-character-bible, produced by okflow and published by Guangxi Liancheng Digital Technology Co., Ltd. It is an enterprise-sourced skill in the design-media category, version 3.2.1. It generates a character design document based on an input character description and also produces four views—front, back, side, and top—eliminating the need to generate the setting and views separately and then reconcile them afterward. The tag indicates requires_api_key=false, meaning no separate API key application is needed to enable it. Current rating is 100000, with 1241 downloads and 0 installs; usage is still limited, so it is recommended to pilot it in a small scope before broader rollout. The materials do not provide an official installation command, directory path, configuration fields, or reproducible examples. This article states these gaps as-is and recommends referring to the official page for the latest details.
Read Moredsh-senses: Adding Eyes, Mouth, and Ears to the DSH Agent
dsh-senses is a DSH community plugin (MIT, v0.1.0) maintained by Geang-yy. It integrates "eyes" (on-demand screenshots and native Windows OCR), "mouth" (edge-tts speech playback after replies), and "ears" (wake words and owner voiceprint). Installation requires pip dependencies and running `dsh plugin add` with the official URL. After installation, users must enroll their voiceprint and run a diagnostic check via `doctor`. It supports half-screen regions and a `look(text)` function to save tokens, with individual toggles for these three features available in the settings page. Designed primarily for local Windows development, the plugin runs with the privileges of the DSH process; please review the source code before use.
Read MoreSecurity-Arsenal: Penetration Testing Tools Categorized by the Cyber Kill Chain
security-arsenal is a community-released security-focused Skill. Its current version is 1.0.2, categorized under it-ops-security, with the tag requires_api_key=false. It uniformly classifies 96 open-source security tools based on the seven stages of the Cyber Kill Chain (Reconnaissance, Weaponization, Delivery, Exploitation, Persistence, Command and Control [C2], and Actions on Objectives), supplemented by a dedicated Web security category. It provides tool lists in scenarios such as downloading penetration testing tools, setting up attack and defense exercise environments, mentioning specific tool names (e.g., Nmap, SQLMap, Metasploit, BurpSuite, Gophish, CrackMapExec, Mimikatz), or inquiring about tool selection for specific stages. This article covers its positioning, scope of coverage, trigger mechanisms, typical example queries, and applicable boundaries. It also outlines the limitations of this list-based Skill: it does not include the tools themselves, nor does it handle installation or updates, and users must confirm their authorization scope before use.
Read MoreEnglish Vocabulary Flashcards Lite: Batch Convert JSON Word Lists to Printable Black & White PNG Flashcards
English Vocabulary Flashcards Lite (@user_1745aec9/vocab-cards-lite, 2.0.4-skillhub, community/education category, rating 100000, 1019 downloads) is a 68KB plain-text Skill without built-in fonts; it utilizes the system's DejaVu for English and IPA, and NotoSansCJK for Chinese. It generates black-and-white print-optimized PNG cards from JSON word data in a single click. These include main cards as well as optional sub-cards and Baidu Baike QR code cards, requiring no API key. It is applicable when users mention terms such as word cards, English flashcards, flashcard, or vocab cards. A typical scenario involves batch converting word lists containing phonetic symbols, parts of speech, definitions, collocations, example sentences, and cultural background into printable card images. This article describes its positioning, core functionality, trigger scenarios, installation and activation methods (since official documentation does not provide installation commands, please refer to the SkillHub page or SKILL.md), and typical usage. It also highlights limitations such as the plain-text version's reliance on system fonts and its output being optimized for black-and-white printing.
Read More