Xianyu Product Copy Optimizer: Enhancing Descriptions for Better Searchability and Inquiries
This article introduces the Skill "Xianyu Product Listing Copy Optimizer" (version 1.0.0, source: community, category: content-creation, tag: requires_api_key=false) released by community author user_32cf4d2a. It focuses on a single, well-defined capability: taking three types of information - product category, condition, and selling points - as inputs to generate product description copy optimized for the Xianyu platform. The goal is to enhance exposure and consultation conversion rates, helping sellers complete sales faster. The article outlines the tool's positioning, core functions, typical usage workflows, as well as applicable and non-applicable scenarios. It also cautions sellers to accurately state product conditions and defects, emphasizing that copy should not be used as a substitute for verifying factual accuracy. As the SkillHub page was unavailable at the time of writing, this article does not include unverified installation commands or configuration entries. Instead, it preserves verifiable coordinates, version, and source information, and provides the official page and API endpoint for readers to independently confirm.
Read MoreAnalysis of Market Size and Growth Trends in Target Therapeutic Areas: Delegating Market Estimation to a Reusable Skill
Released by the Baitong AI Service Platform (Version 1.0.0), this Market Insight Skill for the Prescription Drug Treatment Field is categorized as "professional" and requires no API key. It aggregates global and Chinese prescription drug data from IQVIA MIDAS, Minnet, and EvaluatePharma using unified standards to conduct market sizing, growth driver analysis, competitive landscape assessment, and trend modeling. The outputs are designed for R&D pipeline prioritization, product portfolio planning, and investor communication materials. As official documentation does not provide installation commands or configuration fields, this article lists only the coordinates, official URL, and SkillHub page, while providing invocation examples based on official scenario descriptions and notes on data standard considerations.
Read MoreSmart Stock Picking Score: Turning A-Share Codes into Technical Diagnostic Reports
Stock Score (@user_d767480f/stock-score) is an A-shares scoring Skill in SkillHub’s “professional” category. Published by user_d767480f (version 1.0.0), it requires no API key. Based on a four-component composite scoring model, it takes a stock ticker as input and outputs an overall score, an S/A/B/C/F rating, zone positioning (such as strong bullish, MA repair, or overbought-turning-weak), tactical buy points, trend-following signals, and a multi-dimensional breakdown. It addresses the high cost of building scoring models in-house and the difficulty of keeping scoring criteria consistent. The article covers the core features, output structure, installation and activation paths, typical usage, and applicable scenarios, and notes limitations such as the weights being non-adjustable and the conclusions not constituting investment advice.
Read Moredsh-memory-organizer: Move-Only Archiving for Fixed Memory Limit Exceeded
dsh-memory-organizer is a DSH memory plugin maintained by JiessieM, designed for existing markdown memory notebooks: when the fixed memory area exceeds the default 15,000 character threshold, it archives sections verbatim into the memoryORZ directory while leaving placeholders, preventing new memories from becoming invisible due to hard truncation from the file head by dsh-whale-loli and similar tools. It supports auto-discovery or path specification via files, persistent presence of the impression area, tiered migration by skill sections, and pre-write backups. Installation command: dsh plugin --profile desktop add dsh-memory-organizer; DSH restart required; operational status can be confirmed through heartbeat and configuration options such as threshold and intervalMs.
Read MoreMC All-in-One Workshop: A One-Stop Skill for Minecraft Java Edition Mod Development, Modpacks, and Packaging
MC All-in-One Workshop is a community Skill (v1.0.7, categorized under dev-programming) published by user_644a9d9f. Targeting Minecraft Java Edition mod developers and modpack authors, it consolidates project scaffolding generation, registration of items/blocks/block entities/creative mode tabs, generation of recipe/loot table/tag/model data, network packets and configuration, Mixin and client-side rendering, version compliance checks, low-version mod migration and upgrades, building modpacks from scratch (ecosystem verification/mod list/download scripts/datapack tasks/FTB manifests), auditing existing modpacks, and packaging & distribution (CF zip / mrpack / instance) into a single Skill driven by trigger keywords. It supports code styles for three major versions (1.20.1 / 1.21.x / 26.x) and four major platforms: Forge, NeoForge, Fabric, and Architectury.
Read MorePayroll Expert: Salary Adjustment and Compliance Calculation Skill for Local Excel Data
The Salary Calculation Expert is a community-sourced general skill (coordinate @user_e0ebcb6a/hr-salary-calculation-expert, version 1.0.0, category business-ops, SkillHub rating 100000, 998 downloads, tag requires_api_key=false). It targets HR teams and developers with existing local Excel salary data, consolidating four types of calculations into reusable processes: salary adjustment plan generation, salary inversion detection, market competitiveness benchmarking, and salary compliance review. It does not rely on external APIs, all analyses are based on local Excel, and results are deterministic and explainable, making it suitable as repeatable analysis steps rather than one-off scripts. The article introduces its positioning and capabilities, verified coordinate and tag information, and frankly states that the retrieved materials do not provide installation commands, configuration fields, or step-by-step reproducible examples, so it does not fabricate usage and recommends referring to the official page for details. It also outlines applicable scenarios and precautions: result quality depends on the structure and criteria of the input Excel; market benchmarking requires consistent criteria; it is community-sourced and published by user_e0ebcb6a; salary data is sensitive, so attention should be paid to permissions and data desensitization. The article concludes by reviewing its value and providing the official address.
Read Moredsh-team-dev: Step Budget Guard and Packaged Skills for Development Teams
The `dsh-team-dev` bundle in JieCore is a workflow-type plugin package for DeepSeek Harness. It provides per-turn tool-calling soft/hard thresholds (default 20/25) for Agent Teams' Teammates, rejecting non-whitelisted calls before the hard ceiling and guiding handoffs; the Lead is unrestricted. The same package installs the `team-dev` and `team-boot` skills. Drawing on the official README, this article covers installing the bundle and `cordis.patch`, configuration options, log verification, and probing the "Step budget exhausted" behavior. It also reminds readers to run under process permissions and to self-check MIT license compliance.
Read MoreAiToEarn: AI Content Marketing Skill for One-Click Distribution to 14 Platforms
AiToEarn is an AI content marketing agent Skill, located at @user_1a99cb00/aitoearn, version 1.0.0, from the community, published by user_1a99cb00, categorized under content-creation. It leverages the MCP protocol to distribute content to 14 global platforms with one click, eliminating the need to write separate integration code for each platform. Tagged with requires_api_key=true, credentials must be prepared before enabling it. This article introduces its positioning, core features, installation access points, and applicable scenarios, and transparently marks unverified information (platform list, call parameters, installation commands), recommending reference to the official page for accurate details.
Read Moredsh-island-bridge: Approve and Navigate DSH Sessions from the macOS Notch Panel
dsh-island-bridge is a DSH client plugin maintained by JiayangSUNwinter that projects running root sessions and pending operations to a native SwiftUI Dynamic Island panel on macOS 14+. The approval interface occupies the entire island, with verdict verbs consistent with DSH’s allowed-once/rejected; clicking a task row or source wakes the desktop and opens the corresponding session. Data is transported via in-process seams and stdio, while navigation and responses utilize the same-origin authenticated route without opening additional ports. Installation is available via the plugin management page or `dsh plugin --profile desktop add github:JiayangSUNwinter/dsh-island-bridge`, requiring DSH >= 0.2.1-rc.2 and Xcode CLT. macOS only, MIT licensed.
Read More汇报 PPT 内容汇总与整合助手：把多份 PPT 合成一份总汇报
介绍社区发布的 Skill「汇报 PPT 内容汇总与整合助手」（@user_c160f2b7/report-ppt-integration，版本 1.0.0，分类 office-efficiency，来源 community，评分 100000、下载 982）。它面向用户上传多个 .pptx 后需要汇总、整合、合并成一份总汇报，或需要对多份 PPT 做内容分析、逻辑梳理、汇报顺序建议、去重合并、风格统一的场景，支持汇总PPT、整合PPT、合并PPT、多份PPT、总汇报、汇报逻辑、PPT整合、merge pptx、consolidate ppt 等触发词。核心是五阶段工作流：逐个解析 PPT 内容、逐文件生成内容卡片、提出总汇报逻辑顺序草案、生成整合版总汇报 PPT、后期微调，其中逻辑顺序草案必须等待用户确认后才进入生成，生成依赖 python-pptx。文章说明了定位与触发条件、五阶段流程、依赖要求，并如实指出官方资料未提供安装命令与可复现示例，接入方式取决于宿主对 SKILL.md 的支持，同时给出输入格式、环境依赖、强制确认环节和非文本元素保留等使用注意。
Read MoreWriter.Skill: Eliminating AI-sounding content in Chinese writing
Writer.Skill (@user_59e2247b/writer-skill, v1.0.0, community-sourced, categorized under content-creation, no API Key required) uses the writing habits of modern and contemporary Chinese writers—such as Lu Xun, Wang Zengqi, Lao She, Shen Congwen, Yu Hua, Wang Xiaobo, and A Cheng—as a baseline to constrain the Chinese output of large language models: prioritizing specific nouns, employing varying sentence lengths, and avoiding clichés. It triggers for a wide range of content including articles, prose, fiction, WeChat official accounts, essays, film and book reviews, speeches, and copywriting, as well as requests for "removing the AI tone," "sounding human," "improving literary style," "eliminating robotic feel," or "polishing and rewriting." Even if users do not explicitly specify a style, the tool should be consulted whenever the output is intended as a long-form Chinese text for human readers. The article outlines the tool's positioning, core features, official URL, and activation methods. It provides directly reproducible examples of writing and polishing, and clarifies applicable scenarios, limitations (it does not alter facts, and the baseline acts as a constraint rather than a template), as well as inconsistencies found in the page data.
Read MoreAgent Project Management Skill: Context Management for Long-Term Projects and Multi-Agent Collaboration
Agent Project Management Skill is designed for long-term projects and multi-user, multi-Agent collaboration. Published by user_2731c9a4, version 9.0.0, sourced from the community and categorized under ai-agent, it requires no API Key. Core capabilities include on-demand context restoration, clarifying task scope and responsibilities, recording changes, verification and handovers, and maintaining bounded indexes and searchable history. It is suitable for initialization, taking over, interruption recovery, parallel development, and phase completion. The document notes that installation commands are unverified and usage is inferred from official descriptions; users should confirm details on the official page before use.
Read MoreDSH Client-side Payslips and Model Meal Allowance Estimation
The DSH community client plugin, **dsh-salary-meter** (Working to Fatten the Fish), provides a draggable pay stub on the right side of the interface. It accumulates wages for the current day by shift, displays a shift-end countdown and work status, and estimates the day's costs for specified models under DeepSeek official/account routing based on submitted session records. It supports expanding a "Whale Girl" WebGL animation, with parameters and statistics stored in the DSH data directory. Both desktop and Web versions can be installed via `github:Jiang-Xiaoxing/dsh-salary-meter#v0.3.0`; costs are estimates, so please review the source code and MIT License before installation.
Read More7-Question Method for Book Topic Evaluation: Using Seven-Dimensional Scoring to Decide if a Book is Worth Publishing
This article introduces the "Book Topic Evaluation 7-Question Method" Skill released by Sun Sanye, a book topic evaluation tool based on the "Seven Questions for Publication Topic Selection." It breaks down subjective topic judgments into verifiable evidence grading, seven-dimensional scoring, mandatory requirement assessment, risk diagnosis, and verification actions, generating an auditable Excel scoring sheet. Input materials can include seven categories: synopses, author profiles, table of contents, sample chapters, manuscripts, competitor books, and channel information. The text outlines the Skill's capability boundaries and applicable scenarios, along with implementation considerations such as its community source, zero install count, and no requirement for an API key.
Read MoreIntroduction to the Web Motion Design Expert Skill
The Web Motion Design Expert is an Agent Skill for designing web motion solutions and optimizing performance, published by community publisher user_442551c3. Version 1.0.0, it is categorized under dev-programming. It covers five major scenarios: micro-interactions, transition animations, parallax scrolling, loading states, and data visualization animations. It includes a CSS/JS animation solution selection matrix, an easing curve library, a performance checklist, and accessibility constraints, and is used for three types of tasks: designing motion solutions, optimizing animation performance, and evaluating motion quality. The article introduces its positioning, core capabilities, installation and activation methods (official URL, no API key required), typical usage, and applicable scenarios along with caveats (community-sourced, low installation count, official details still need to be completed).
Read Moredsh-linked-workspace: Connecting Remote Directories into DSH using SSH
dsh-linked-workspace is a DSH client plugin that provides read-only remote directory browsing via SSH/SFTP, along with a linked workspace that pairs local and remote directories. Within a session, the linked_workspace tool enables direct reading and writing of remote files, while the right sidebar's dual-tree editor supports file transfer, git change views, and optional mirror synchronization. No DSH installation is required on the remote end. It targets DSH 0.2.x and depends on dsh-better-sidebar v0.24+. After installation, configure it in Settings → Connections or Add Workspace, then restart the desktop application to load the Host plugin.
Read MoreMistake Radar: Turning K12 Mistakes into Root-Cause-Based Remediation Plans
Error Question Radar (@user_65b9a06e/studybuddy-plus, K12-EKG-PLPR) is an Agent Skill in the education category of the SkillHub community. Version 1.0.0 requires no API Key. It processes not the archiving of incorrect answers, but the root causes of errors: it first classifies mistakes by error reason (the official source emphasizes that "carelessness" cited by children is not necessarily true carelessness), maps them to knowledge points for diagnosis, and then traces back to prerequisite dependencies—for example, a failure in quadratic equations may stem from an incomplete understanding of linear equations. It subsequently generates remediation strategies, organizes tasks into a personalized daily learning path of approximately 20 minutes, and tracks progress. It incorporates the "Five No's Principle" (No Identity Storage, No Ranking, No Labeling, No Prediction, No Harm) as output constraints for scenarios involving minors. The article explains the pipeline relationship of six capabilities, the activation method for the generic SKILL.md, and reproducible usage patterns on the input side (specifying grade level, question text, student's answer, and daily duration). It should be noted that the official webpage content failed to be retrieved during this capture; the SKILL.md source text was unavailable in the materials, resulting in the omission of installation commands and configuration fields. Page ratings and download counts are based on capture snapshots and are subject to change over time. Applicable scenarios include parental test paper analysis, batch identification of common knowledge gaps by teachers, and pre-exam review. Limitations include the fact that diagnostic quality depends on input quality, the scope of subject coverage is not publicly disclosed, and products requiring long-term account tracking and ranking capabilities would need to modify its default behavior.
Read MoreFoundation Project Cooperation Information Summary: Consolidating Scattered Foundation Recruitment Notices into a Verifiable HTML Page
# Introduction Those involved in matching public welfare projects often face a challenge: information regarding "partner recruitment" and "project collaboration" from foundations is scattered across various institutional websites, WeChat official accounts, and third-party platforms, lacking a unified entry point. Manual collection involves opening links repeatedly, verifying expiration dates one by one, and consolidating duplicates—a time-consuming process prone to errors. This Skill streamlines the workflow of retrieval, verification, and aggregation. By inputting a specific topic, it generates a filterable HTML resource page where links can be clicked to verify information. # What is this "Foundation Project Collaboration Resource Aggregation" is a resource compilation Skill designed for the public welfare sector. It is located at coordinates `@user_b687eed0/jianxing`, published by the community account user_b687eed0, currently at version 15.15.15, and sourced from co...
Read Moredsh-superagent-inbox: Local Folder-Driven Super-agent, No External Account Required
dsh-superagent-inbox is an extension developed by JHOpenDev that adds an inbox Source to the official dsh-superagent bundle. By configuring local directories ops/in and ops/out, it triggers Router evaluation, ticket creation, and Worker execution via JSON messages, with results written back to the folders, eliminating the need for GitHub or Feishu bots. The article covers the directory protocol, message fields, Policy mapping, as well as installation via git clone and cordis.patch.yml, Local inbox configuration in the panel, and typical delivery commands. It is suitable for local testing and script integration; the plugin runs with dsh process permissions, and source code and MIT license should be verified before deployment.
Read MoreInvestment Banking Document Polishing Assistant: Language Refiner for Prospectus Style
The Investment Banking Material Polishing Assistant is a community-sourced, general-purpose Skill in SKILL.md format. Targeting scenarios such as prospectus chapters, inquiry responses, and research or due diligence reports, it optimizes language by converting colloquial expressions into standard written form, removing traces of AI generation, promotional tone, and emotional embellishments. It streamlines expression to align with the objective and restrained style typical of prospectuses. The tool focuses strictly on polishing; it does not supplement data or assess compliance. It outputs a clean Word document accompanied by a .md comparison file for sentence-by-sentence verification. This article introduces its positioning, core functions, installation and activation methods, typical usage, applicable boundaries, and provides the official link.
Read Moredsh-tech-companion: Gently Completing Technical Expressions in DSH Conversations
## dsh-tech-companion is a DeepSeek Harness workflow plugin maintained by jf143757-max, which injects static behavioral guidance via bundle. While completing tasks, it gently supplements terms and distinguishes similar concepts at the end of responses with "Teacher Zhizhi reminds you:" and provides precautionary notes before high-risk operations. It supports Hermes-sourced SKILL.md. The installation command is dsh plugin --profile desktop add github:jf143757-max/dsh-tech-companion, with an engine requirement of dsh >=0.2.1-rc.3 <0.3.0, verified on 0.2.1-rc.3. It does not store chat records or make additional model calls.
Read MoreSingle-Cell Sequencing Analysis Skill: Delegating QC and Clustering to Managed Executors
bio-single-cell-integrated is a community-sourced Skill in the education category, published by user_f4fc8899, at current version 1.4.22, with a SkillHub rating of 100000 and 909 downloads. It encapsulates one of the most routine segments of single-cell analysis into a managed executor: computing total counts per cell, the number of detected genes, and the mitochondrial ratio; counting QC-passing cells; and performing k-means clustering after log1p(CPM) normalization. The input is documented JSON, and the output is structured content, which can be returned as Markdown when provided. The official documentation explicitly states that it does not run upstream local toolchains and does not operate third-party accounts; its only deliverables are verified server-side responses. This article introduces its positioning and verified core capabilities, and notes that installation and usage documentation are missing, so the official pages should be treated as authoritative.
Read MoreAgent's Garden · File Organizer: Let Agent Tidy Up Messy Directories
Agent's Garden · File Organizer (@user_69a6e2fe/agent-garden-file-organizer) is a community-sourced, general-purpose SKILL.md-format Skill, version 1.2.0, published by user_69a6e2fe and tagged requires_api_key=false. Its official description lists four capabilities: automated classification, deduplication, structural optimization, and junk cleanup, aimed at long-accumulated scattered files in Downloads, Desktop, and project folders. The article only presents information based on verified sources, and notes that the official page and API could not be fetched at the time of writing, so the installation command, configuration fields, and the enabling steps in Cursor, Codex CLI, and Claude Code cannot be verified; therefore, no installation command is provided. It also gives applicable scenarios and usage notes: because moving and deleting files are involved, it is recommended to first validate classification and cleanup rules in a single directory; the official documentation does not specify classification rules, deduplication algorithms, supported file types, or criteria for identifying junk; requires_api_key=false only indicates the presence of the tag and does not equate to no external dependencies, so users should read the original SKILL.md before use. The official URL and the SkillHub page are included at the end.
Read Morecosmetics_packaging_enterprise_brands: Turning Major Overseas Cosmetic Groups into Prospect Lists for Printing and Packaging
Introduction to the community Skill `cosmetics_packaging_enterprise_brands`: Guides research on major overseas cosmetics groups and established brands to generate a prospect list for printing and packaging. Requires no API Key; includes installation, activation, typical usage, and scope of application.
Read MoreDocument Formatting: Standard-Level Formatting of Manuscripts in Accordance with GB/T 1.1-2020
This article introduces the SkillHub community Skill "Document Formatting" (coordinate @user_9b7fb39b/fhgwdpb, version 1.0.0, category office-efficiency, requires_api_key=false). This Skill performs rigorous professional formatting on documents in strict accordance with the "Directives for standardization - Part 1: Rules for the structure and drafting of standards" (GB/T 1.1-2020). It can be activated by trigger words such as "format", "format document", "format by national standard", "GB/T formatting", "standard document formatting", "format for review submission", and "generate formatting scheme", or it can automatically organize draft documents uploaded by users according to national standard specifications. The article outlines its core capabilities, official address and activation methods, two typical use cases, as well as applicable scenarios and limitations: it does not perform polishing, translation, or content rewriting; requests solely for changing fonts or font sizes are not applicable; and it is recommended to cross-check against the original standard text before submission for review.
Read More