Smart Chart Studio: A Skill for Requirement-Driven Chart Generation
smart-chart-studio is a community edition chart workflow Skill, published by @user_df210223, version 1.0.1, categorized under data-analysis. It frontloads requirement clarification: collecting business goals, chart types, data calibration, and format requirements through 4 rounds of Q&A, then automatically routing to 4 rendering engines via 14 intent categories, and performing visual enhancement and enterprise format compliance. It is positioned as "ask first, then draw," not a data-to-chart tool. The documentation indicates requires_api_key=false, meaning no additional API Key configuration is needed. This article describes its core capabilities, installation entry points, and expected interaction workflow, while pointing out limitations such as the four rounds of Q&A involving too many interaction rounds, and routing details needing to be referenced from the official page.
Read MorePatent Application Analysis and Drafting | JianShi AI: A Skill for Assisting in Writing Patent Application Documents
This is a community Skill named "Patent Application Analysis & Drafting | Jianshi AI," located at @user_fccf428b/patent-lawyer-agent, published by user_fccf428b, version 1.0.1, categorized as professional, with a rating of 100000 and 1111 downloads. It provides intellectual property guidance and patent analysis, assists in drafting patent application documents, and requires API key configuration. The official description explicitly states that it is not legal advice and that key conclusions should be confirmed with a licensed patent attorney. This article introduces its positioning, verified capabilities, confirmation steps before enabling, appropriate usage, and applicable limitations. The official page does not provide verifiable installation commands or reproducible examples, and this article does not fabricate any.
Read Moredsh-ps-preflight: Static Tiering and Auditing for PowerShell Prior to Agent Execution
dsh-ps-preflight is a DSH community admin-security plugin that performs zero-dependency, offline static risk grading (none/low/medium/high) through tools/pre-execute before invoking pwsh/powershell, covering rules such as deletion, overwrite, download, and credentials. The high tier requires secondary confirmation via the panel or harness, and all actions are written to a full JSONL audit log. It is ready to use out of the box with the default enforce mode, and also supports an audit mode that only logs without blocking. Based on the official README, this article covers the installation command, configuration options, panel usage, and capability boundaries, and explains how it differs from dsh-warden.
Read MoreSatisfaction Survey Questionnaire Generator: A Single-File HTML Survey with Scoring
This article introduces the community Skill on SkillHub: "Satisfaction Survey Questionnaire Generator (Satisfaction Survey)," located at @user_af163eaa/z-satisfaction-survey, version 1.0.0, category dev-programming, requires_api_key=false. Targeting scenarios that require satisfaction surveys (customer/employee/service evaluation), this Skill generates self-contained, single-file HTML surveys that include NPS 0–10 scoring, CSAT scale items, and open-ended questions. Scoring is computed locally on submission, with no backend or survey platform account required. It is triggered by mentions of "satisfaction survey / NPS survey / service evaluation form / satisfaction survey HTML." The article covers core features, how to enable it, typical usage, and applicable scenarios and caveats, including question type coverage, local scoring without server-side persistence, and current install count of 0.
Read Moredsh-pet-ledger: Pet Ledger Driven by Real DSH Session Events
dsh-pet-ledger is a DeepSeek Harness plugin where achievements are derived solely from real events such as `tools/result`, `agent/status`, and `session/event`, covering four series: continuous failure, long-running tasks, test red-to-green transitions, and commit pushes. Each achievement allows for the expansion of session and tool call evidence; demos feature a watermark and are not persisted to disk. Requirements include DSH version >=0.2.1-rc.2 <0.3.0, an MIT license, and data stored locally in `pet-ledger.json` by default.
Read MoreData Quality Diagnostics: A Health Check for Your CSV/Excel Files
Data Quality Diagnosis is a community-sourced Agent Skill (Coordinate: @user_053d27d0/data-profiler-zx, Version 1.0.0, Category: data-analysis) that analyzes CSV/Excel files, outputs descriptive statistics, provides five-dimensional quality scores (completeness, uniqueness, consistency, validity, and timeliness), detects outliers and formatting issues, and generates an HTML report. The article explains its positioning, five core capabilities, API key-free loading method, and typical usage scenarios, while highlighting usage notes such as the undisclosed scoring weights across the five dimensions, the community source, and zero installations.
Read Moredsh-market-card: Formatting Repository READMEs into Marketable Bilingual Market Cards
dsh-market-card is a DSH client plugin and standalone CLI maintained by dingchenhui0618-arch. It reads README.md, package.json, and docs screenshots, generates 1200px-wide bilingual (Chinese-English) market cards, and outputs self-contained HTML, 2x PNG rendered via Chrome/Edge, paste-ready card.md, and sourced card.json. Supports --strict and market-card.json to override heuristic extraction; no machine translation is performed. Installation: dsh plugin --profile <profile> add dsh-market-card.
Read MoreFinancial Morning Briefing: Translating Morning Summaries into Daily Sector Judgments and Trading Strategies
The Finance Morning Briefing is a daily A-share morning reading analysis assistant on SkillHub, published by user_74075f59, version 1.0.1, source community, category professional, tag requires_api_key=false. It is designed for pre-market morning reading scenarios, organizing the daily morning summary into sector judgments, clear operational recommendations, and risk warnings. It is suitable for investors who have a morning reading habit and hope to quickly form structured conclusions. The scraped materials do not provide a unified CLI installation command or client-specific differences for Cursor, Codex CLI, and Claude Code. When using it, you can first verify the version and activation method through the SkillHub coordinate @user_74075f59/hana or the official address. It should be noted that the materials do not specify data sources, backtest performance, or accuracy rate, and operational recommendations should not be regarded as guaranteed return promises.
Read MoreScientific Weight Loss: The Skill for Natural Language Fat Loss Logging
"Scientific Weight Loss" is a community Skill that follows the general SKILL.md format. It is located at @user_a6274128/slowshed, version 1.0.0, categorized as life-service, and sourced from the community. It draws on ideas from the books "You Are What You Eat" and "My Last Diet Book," as well as content from Bilibili creator "Good Person Songsong." The Skill breaks fat loss into three parts: scientific meal pairing, supportive check-ins, and quantifying results. Core features include recording weight, meals, and exercise in natural language; estimating calories for snacks; checking how much you can still eat today; and weekly fat-loss reviews. It emphasizes aiming for about a 1% body-weight reduction per week over a half-year period, eating staples, protein (meat/eggs/dairy), and fats every day, and that better sleep can speed up fat loss. It automatically converts portions and presents guidance in a conversational style, making it suitable for beginners who prefer not to measure by grams or understand complex calorie terms. The official page tag indicates requires_api_key=false, meaning no additional API key is needed to use it. The materials do not include a verifiable installation command; enablement should follow the official page or official address. Before use, note that it is a community-published Skill and does not constitute medical advice; results depend on the user’s consistent recording of meals, weight, and exercise.
Read Moredsh-liars-dice: Playing Chinese Liar's Dice in DSH, where Agents can either join the game or serve as the referee.
dsh-liars-dice is a Liar's Dice plugin for DeepSeek Harness developed by dingchenhui0618-arch. It supports 2–6 player hot-seat mode, Agent opponents, a Scoop Referee, replays of bidding and scooping, and a toggleable binomial distribution probability panel. To install, specify the profile using `dsh plugin --profile`, then run `add github:dingchenhui0618-arch/dsh-liars-dice`. Requires DSH >=0.2.1-rc.2 and <0.3.0. It is MIT licensed and has not yet been published to npm. It is suitable for local gameplay experiences and observing Agent bidding behavior. Note that the game table resides solely in the host's memory and is not an online room.
Read MoreFull-Paper Markdown Reader — Router: A Bilingual (Chinese-English) Markdown Reader Skill for Academic Papers
Introducing Full-Paper Markdown Reader — Router: A tool that converts papers from PDF, DOI, arXiv, publisher HTML, or pasted text into full-text bilingual Markdown with rendered formulas, repositioned figures/tables, and source anchors for each block. The article verifies the SkillHub resource page's coordinates, version, publisher, categories, tags, and description; explains core functionality, trigger scenarios, and usage considerations; and honestly acknowledges that the installation command could not be verified.
Read MoreDaily team debrief crucible: Transforming a single work log entry into a reusable reflection template.
Daily Review Alchemy Furnace (Daily Review Alchemy Furnace, @user_69f3cf82/professional-capital-system) is an Agent Skill designed for personal work review, released by user_69f3cf82. Sourced from the community and categorized under business-ops, it is at version 1.0.7 and tagged with requires_api_key=false, indicating no additional API Key configuration is needed. It functions by accepting a single sentence about the day's work and executing the IPO-D-G-A closed loop (Input → Probing → Output → Decision → Growth → Asset). It generates a review card with impact grading and ROI layering, while depositing project memories, cognitive breakthroughs, and SOP assets into a personal capital account. Distinguished from generic chat-based reviews, it embeds a deterministic adjudication script. This script determines the five-tier impact level (S+/S/A/B/C) in a single sentence, automatically triggers three levels of ROI inquiries, and constrains output quality using Q1-Q8 quality gate self-checks. There are four trigger methods: direct input of work logs or daily reports, commands for daily/weekly/monthly reviews, retrieval of project memory cards, and requests for ROI analysis or strategic decision support. This article outlines core functions, official links, and the SkillHub page, along with usage instructions. It also highlights two limitations to note: input quality dictates the effectiveness of grading and ROI inquiries; discrepancies in scoring and download counts may exist depending on the statistical timeframe, so official pages should be treated as the definitive source.
Read Moredsh-compress-bench: Quantized Context Compression via Mirror A/B in DSH
dsh-compress-bench is a local-image A/B measurement tool provided by dingchenhui0618-arch for DeepSeek Harness. It pins the model and temperature inside a read-only fixture sandbox, running once with compression disabled and once with compression enabled. It outputs locally estimated input tokens, turn count, tool success rate, fact coverage, and answer differences, and generates JSON for dsh-compress-bench/report@1 along with a comparison SVG. It is suitable for weighing token savings against trade-offs with reproducible evidence before switching models or compressors. Requires DSH >=0.2.1-rc.2 <0.3.0; install via plugin management from github:dingchenhui0618-arch/dsh-compress-bench or a local build directory.
Read Moredsh-tailnet-ingress: Connecting Loopback Web UI to Local Private and Tunnel IPv4
The DSH client plugin dsh-tailnet-ingress, maintained by dingcang, listens on local private networks, CGNAT (including Tailscale 100.x), and the 198.18.0.0/15 IPv4 range, transparently piping TCP connections to a Web listener on 127.0.0.1. By default, autoDetect reads the network interface table, and the port follows the dsh web setting. It can be installed via npm or github:dingcang/dsh-tailnet-ingress. Multi-address reachability does not imply API access authorization; it must be combined with --trusted-host. MIT licensed; version 1.2.2 as listed in package.json.
Read MoreContract Legal Compliance Review: Integrating Compliance Pre-screening of Chinese Contracts into Agent Workflows
This article introduces the Skill “Contract Legal Compliance Review” (@baitongai/contract-legal-compliance-review), published by the Baitong AI service platform, version 1.0.3, in the professional category. It reviews risks in Chinese contracts across four dimensions—legal compliance, legal validity/effect, rights and obligations, and completeness—and verifies legal bases by checking the original text of current regulations through the Legal Star API. It supports TXT, Markdown, DOCX, standard PDF, and scanned PDF inputs. The materials indicate requires_api_key=true, so an API key must be prepared first. The official page does not provide installation commands or reproducible examples; the body text states this as-is and does not fabricate configuration options or invocation code. This Skill is not intended for due diligence on business registration, litigation, credit, seals/signatures, or approval workflows. Its output is for preliminary screening and does not constitute legal advice.
Read Moredsh-tui-workspace-sidebar: Workspace Inbox Sidebar in dsh-TUI
The community plugin dsh-tui-workspace-sidebar provides a right-side inbox panel for dsh-TUI 0.17.0, displaying running, unread, pending-answer, and recently read sessions grouped by workspace. It supports jumping, marking as read, and creating new sessions within a workspace. Installation command: dsh plugin --profile dsh-tui add dsh-tui-workspace-sidebar. Current version 0.1.2 (MIT). The plugin uses only public host seams, persists state locally to ~/.dsh-tui/workspace-sidebar.json, and performs no network telemetry.
Read Moredrawio-to-visio: Locally convert Draw.io diagrams to editable Visio VSDX
drawio-to-visio is a Skill published on the SkillHub community, maintained by user_400639b0. It is categorized under dev-programming with version 1.0.1. It addresses the specific pain point where Draw.io/diagrams.net diagrams (.drawio/.xml) provided by clients need to be delivered as Visio-editable files, but confidentiality constraints prevent uploading content to online conversion sites, and manual redrawing is too time-consuming. It works by utilizing local browser automation to parse the input and output editable VSDX files, while performing structural and page content validation to ensure the result is a genuine Visio document rather than a screenshot. Trigger scenarios include converting, exporting, batch converting, or delivering VSDX files. The page indicates requires_api_key=false, meaning no API key application is required. This article organizes its core functions, reproducible usage, and applicable boundaries based on the official introduction. It also truthfully notes that the retrieved materials lack specific installation commands and configuration fields, suggesting users refer to the official page and SKILL.md for authoritative information.
Read MoreBusiness Cooperation and Negotiation (Distribution Channels) Expert: Writing Airline Negotiation Processes as a Reusable Skill
This article introduces the community Skill "Business Cooperation & Negotiation (Distribution Channels) Expert." It organizes the negotiation process into three stages: preparation, consultation, and contracting. Leveraging a structured methodology, it facilitates partnership expansion between airlines and stakeholders such as airports, OTAs, and corporate clients. The objectives are to achieve resource complementarity, expand route networks, enhance service capabilities, and diversify revenue. The article outlines core functions, integration approaches derived from the profile, applicable scenarios, and precautions. Note that the official page failed to fetch, and installation configuration has not been confirmed; the full text of SKILL.md must be reviewed before deployment.
Read MoreQBuddy (dsh-qbuddy-prompt-enhance): One-click rewrite of drafts into actionable prompts in the input box
QBuddy is a DSH client plugin that provides an enhancement button on the right side of the input box. It reads the current draft, utilizes a single independent LLM call to rewrite it into a clearer and more actionable prompt, and writes it back to the input box without participating in the conversation history. It supports four built-in presets expanding from Qoder to WorkBuddy, unlimited custom presets, and quick switching via dedicated menus for enhancement-specific models and thinking intensity. The desktop version must be installed via settings or `install_bundle`; the Web/CLI version is available at `github:dhdkfkfkf/dsh-qbuddy-prompt-enhance`. It is licensed under MIT and maintained by dhdkfkfkf.
Read MoreAndroid Image Processing: Compression, Conversion, and Batch Optimization - A Standard Workflow for Android Developers
This article introduces the Android-focused image-processing Skill “安卓图片处理 压缩转换 批量优化” (coordinates: @user_a10ab3ad/imagecraft-android). Maintained by the community publisher user_a10ab3ad, it is categorized under dev-programming. The current version is 1.0.3, and it is tagged requires_api_key=false. The Skill is positioned as a standardized workflow for image compression, format conversion, and batch processing, helping developers reduce repeated trial and error around parameter selection, format decisions, and edge cases when uploading profile pictures, sharing images, or batch processing tens to hundreds of images. Core capabilities include image compression, format conversion (including common Android formats such as WebP), and batch optimization. In terms of installation, it is distributed using the general SKILL.md format. Official documentation does not provide a specific installation command, so the article describes only the general approach: obtain SKILL.md, place it into the client’s Skill directory or load it as documented, and trigger the task in a conversation. As requires_api_key=false, no API Key configuration is required. Regarding typical usage, official documentation does not provide reproducible commands, scripts, or parameter examples; the article therefore refrains from fabricating samples and instead outlines the process of specifying the task type, executing according to the workflow, and verifying the results. Use cases include Android developers who need batch image compression and conversion, teams seeking reusable workflows, and scenarios where an AI assistant handles image tasks. Notes include that dependencies on libraries such as Glide, Coil, and BitmapFactory are not specified; batch memory and threading strategies are not detailed; and because the data comes from a community release, information should be verified independently. The article provides the official URL and the SkillHub page URL for further reference: https://skillhub.example.com/skill/@user_a10ab3ad/imagecraft-android and https://example.com/official/skill-page.
Read Moredsh-llm-token-usage: Tracking Token Usage and Reference Costs in DSH
dsh-llm-token-usage is a DSH client plugin that uses ctx.sessionQuery to perform read-only aggregation of the harness's LLM token usage and reference costs, and provides a left-side dashboard, slash commands, session capsules, and an llm_token_usage tool. It supports day/model/workspace/session reports, CSV and HTML export, platform balance and local price list estimation, and can automatically sync reference prices during plugin updates. This article draws on the GitHub README and package.json to introduce installation commands, typical usage, configuration options, and security guidelines.
Read MoreDSH Soul Seed: A Workflow Plugin for Local Persona, Mood, and Memory
dsh-soul-seed (v0.11.10) is a workflow extension for DeepSeek Harness released by DGcoin0415. It generates a persona.md via a Web panel using a ten-question prompt, and features a six-dimensional mood system, long-term memory, personal ledgers, and a memory tree (with root, branch, leaf, and pruning mechanisms). It also provides CLI tools such as 12-hour session pagination, vector recall, soul pack migration, and response filters. Based on the GitHub README and SkillHub entries, this article outlines installation notes, typical commands, and use cases.
Read MoreInvestment Positioning and Solution Planning for High-end Residential Real Estate: An AI Planning Assistant for the Full Process
This is an AI planning Skill designed for residential real estate investment decision-making, published by community developer user_f1869d8e on SkillHub. Versioned at 1.0.0 and categorized under business-ops, it requires no additional API key. Modeled on top industry methodologies, it covers six core capabilities: six-dimensional city analysis, five-dimensional weighted scoring of land parcels (Location 30% / Amenities 25% / Planning 20% / Market 15% / Development 10%), customer profiling for five distinct segments, product positioning and unit type ratios, ten-dimension "Good Housing" schemes, and comprehensive economic calculations with sensitivity analysis. Built with an index of over 100 national and industry standards, a timeline of "Good Housing" policies (2023–2025), and a compilation of methodologies from leading real estate enterprises, it generates seven types of professional reports: city analysis, land parcel assessment, customer profiling, product positioning, "Good Housing" planning, economic calculation, and investment decision-making. The accompanying article outlines its capability boundaries, activation methods, and application guidelines.
Read MoreBlog Writing Assistant: Writing Blogs in the Author's Own Voice
The Blog Writer assistant on SkillHub (coordinates @tomstools11/blog-writer, v0.1.0, MIT-0) was released by tomstools11 on ClawHub, categorized under content-creation, with a rating of 100,000 and 14,291 downloads. It calibrates tone, vocabulary, structure, and length using the author's published articles in references/blog-examples/, following a five-step workflow: collecting topics and materials, drafting according to style-guide.md, iterating based on feedback, publishing to Notion (TS Notes library, Type=Writing, with a hardcoded database ID and project page — the skill specifies that failing to execute this step counts as failure), and storing the final draft in the examples library while maintaining a 20-article cap. Installation uses openclaw skills install @tomstools11/blog-writer, with no API key required. Note that the examples library is initially empty and needs to be populated with the author's articles first.
Read MoreTeams Message Automation Management | Jianshi AI: Integrating Channel Messaging and Workflow Automation into a Single Skill
This article introduces the community Skill “Teams Message Automation Management | Jianshi AI” (coordinates @user_fccf428b/microsoft-teams-automation, version 1.0.1, category office-efficiency). The Skill packages Microsoft Teams channel message sending, formatted content push, and team collaboration workflow triggers into callable capabilities, replacing repetitive integration work that typically requires hand-written bots or scripts. The article verifies the publisher attribution, three core functions, and the prerequisite requires_api_key=true, and notes that at the time of writing both the official page and the SkillHub page failed to fetch. Therefore, it does not provide unverifiable installation commands or configuration items, and only provides the official address as the entry point for enabling it. The article concludes by outlining applicable scenarios: it is suitable for people who need to perform repetitive message pushes and workflow triggers. Limitations include: it is community-sourced and permissions must be evaluated independently; the version is relatively early; an API key must be provided by the user, and the key channel is not explained; and the number of installations is 0.
Read More