Business Plan Assistant: Delegating the Drafting of Business Plans to Skill
The Business Plan Writer Assistant (@user_e0968313/biz-plan-writer, v1.0.0, community source, business-ops category, requires_api_key=false) is positioned as a senior business plan writing expert, covering three scenarios: financing, entrepreneurship, and business planning, serving entrepreneurs, business owners, and project managers. The article verified the official page metadata and noted that installation commands and SKILL.md content were not provided in the materials. Therefore, only verified information was included, with unverified installation steps and examples marked as blank, and readers were advised that data must be provided by themselves, community source content should be read in full, and ratings and download counts recorded inconsistently across different channels need to be verified independently.
Read MoreDeepSeek-Brain: DSH Five-Layer Persistent Memory and 3D Observation
deepseek-brain is a memory plugin for DeepSeek Harness, maintained by Elias1222. It provides five-tier JSON persistence: sensory, working, episodic, semantic, and procedural. It supports decay, sleep consolidation, semantic conflict resolution, and the `brain_*` toolchain. Data is located in `~/.deepseek-brain/`, containing an ASCII brain map and a Web/Three.js 3D dashboard. The installation command is `dsh plugin --profile web add github:Elias1222/deepseek-brain` (the profile needs to be replaced per your local machine). It is pure Node ESM, MIT licensed, and requires Node >= 22.
Read Moredsh-llm-minimax: Instantly Access MiniMax Official M3 Series by Providing Your API Key in DSH
Maintained by elfgzp, dsh-llm-minimax is a DSH model inference plugin that updates the MiniMax-M3.1-Flash and MiniMax-M3 directories on the existing minimax-cn/minimax routing. It features a settings page offering one-click API key entry, automatic detection of domestic/international endpoints, credential writing, and an M Plan quota toolbar. Requires DSH 0.2.x and supports installation via Web, Agent, and CLI. This plugin does not implement a new adapter, and pay-as-you-go keys do not have a balance API. Please review the source code and MIT license before installing.
Read MoreLiterature Retrieval Rapid Reader (Instant Generation & Retrieval Output): Integrates OpenAlex Free API to Directly Generate Structured Skimming Results
The Literature Search Quick-Reading Tool (@user_af163eaa/z-lit-search, v1.0.1, knowledge-management), maintained by community publisher user_af163eaa, connects to the free and public OpenAlex API for literature retrieval without requiring an API key (optionally setting OPENALEX_EMAIL to increase the rate limit). It generates structured quick-reading results with four fields: title, year, citation count, and abstract key points, and is officially marked as accessible within China. The page has a rating of 100000, 1107 downloads, and 0 installations. This article introduces its positioning, capabilities, enablement methods, and applicable scenarios.
Read MoreMemory Manager EN-V4.0.6: Making contextual memory smaller and more precise
Memory Manager EN-V4.0.6 is a memory management Skill designed for AI Agents, positioned as a token throttle. Instead of deleting memories, it ensures that memories entering the context are smaller and more precise. Core capabilities include a write governance loop (write hygiene audit, automatic snapshot and rollback, compression, token savings dashboard, five-dimensional health checks), security hardening (write command confirmation gate, SHA256SUMS integrity verification, 0700/0600 permissions, rejection of ZIP path traversal and oversized members), selector visibility (selector-audit, describe backfilling of front-matter descriptions), relevance-based loading (--query/--tags/--top-k, zero-dependency CJK bigram matching and 90-day recency decay), real tiktoken metering, automatic compression when exceeding budget, one-click throttling, tri-color budget alerts, --json output, and built-in self-checks. It is designed for Agent workflows where memory directories grow continuously and token costs are rising; it is not suitable for general file operations, code analysis, system cleanup, or document editing.
Read Moredsh-default-workspace: View, Fix, and Migrate Default Workspace in a Running DSH
The DSH client plugin 'dsh-default-workspace' for Elefseus allows checking the default workspace and performing switch/repair operations via the default_workspace tool without requiring a Host restart, with optional migration of directories and session logs. This article details installation commands, parameters, safety constraints, and Web panel and sessionsRoot configurations, making it suitable for scenarios involving system disk space management, document redirection, or fixed working paths.
Read MoreIn-depth Correction for IELTS Writing Task 1 / Task 2: A Structured Approach
# Introduction In the context of IELTS writing grading, a common approach is to prompt a general conversational model directly with the question, "What score does this deserve?" However, the issue is that models often provide only a single overall band score, or they identify errors sentence-by-sentence without cross-referencing the specific task requirements or breaking down the assessment across the four scoring dimensions. The "IELTS Essay Intensive Grading" Skill standardizes this process: it first receives the prompt along with the essay text or images. Then, based on the requirements for Task 1 or Task 2 (Academic or General Training), it provides a four-dimension teaching assessment, a comparison against task requirements, paragraph-by-paragraph error correction, and prioritized suggestions for improvement. The following section outlines its capabilities, input methods, and precautions. # What is this? "IELTS Essay Intensive Grading" is a skill designed for IELTS Academic and General Training...
Read MorePharmaceutical Logic Audit Skill: Item-by-item Assertion Auditing of External Positioning Copy
"Pharmaceutical & Medical Logic Audit" is a specialized Skill on SkillHub, published and maintained by the Baitong AI Service Platform (Coordinate: @baitongai/pharmaceutical-medical-logic-audit, Version 1.0.0, Source: enterprise, No API Key required). Adopting the rigorous perspective of clinical medicine and pharmacology experts, it extracts assertions from positioning copy item by item and traces evidence, marking risks with a four-level annotation system (🟢🟡🔴⚫) and ultimately outputting a remediation priority matrix. This assists medical and pharmaceutical teams in ensuring their external materials withstand compliance challenges. This article introduces its positioning, core features, installation and activation methods, typical use cases, and applicable limitations.
Read MoreZero-to-Launch Assistant for WeChat Mini Programs: A Skill Explaining Scaffolding, Errors, and Deployment
This is a community skill designed for Agents, covering the complete workflow of WeChat Mini Programs from project creation to submission, review, and release. It addresses the dual-thread architecture, the four-file structure (WXML/WXSS/JS/JSON), the 'rpx' sizing unit, 'setData' performance, the Canvas 2D interface, common errors (such as invalid 'disableScroll' configuration in 'app.json', 'ctx is not defined', and page lag), the distinction between AppIDs and test accounts, experience member permissions, and the upload-review-release workflow. Trigger phrases include "create a mini program," "mini program error," and "mini program release." It is suitable for developers new to Mini Program development and teams needing to quickly troubleshoot issues, serving as a standard component in an Agent's toolkit.
Read MoreJoinQuant: A-Share Data Query and Strategy Backtesting in Agents
JoinQuant (JoinQuant) is an Agent Skill categorized as 'professional' from an enterprise source, published by RealQ (version 1.0.0, coordinates @org-eyw2ohcx/coderwpf-joinquant-102). It serves as an entry point for the JoinQuant quantitative trading platform, providing A-share, futures, and fund data queries, as well as event-driven strategy backtesting. It also supports online research and paper trading simulation, aiming to consolidate workflows—originally scattered across research environments, backtesting pages, and local scripts—into a single conversation to minimize context switching. Although the SkillHub page is tagged with `requires_api_key=false`, the documentation does not clarify whether a JoinQuant account, token, or platform authorization is still necessary. The available data only includes a brief introduction, lacking installation commands, directory conventions, configuration fields, or reproducible examples. Consequently, this article does not provide call examples but recommends first visiting the official page to confirm input parameters, output formats, and authorization methods, followed by sequential verification: single data queries, small-range backtesting, and paper trading simulation. It is important to note that ratings and download counts are subject to the current data on the official page. Backtesting results depend on data quality, transaction cost assumptions, and slippage models, and should not be directly interpreted as expected real-world returns.
Read MoreIndustry Regulations & Standards Quick Reference: A Skill for Auto-organizing Compliance Checklists by Object
"Industry Regulations & Standards Quick Lookup" is a professional-skill published by a community. Version 1.0.0, produced by user_966d65df, and it does not require an API key. Targeting industries, enterprises, and products, it is triggered when users request compliance information, regulatory requirements, market-access conditions, or qualification/certification requirements. It automatically searches across six dimensions: international regulations, domestic laws, administrative regulations and ministerial rules, mandatory national standards, recommended national standards (GB/T), and group standards/industry standards. It then cross-verifies results with the National Public Service Platform for Standards Information, the relevant regulatory ministries/commissions, and the NPC Law Library, and finally outputs, in a fixed hierarchy, both bulleted Markdown entries and an HTML table with color-coded status labels. The article also clarifies its boundaries: it only provides regulations, standards, and compliance checklists; requests related to market size, technology roadmaps, competitive analysis, or industry research should be handled by industry-research-pro. It also provides the official URL, coordinates, ratings and download counts for page snapshots, and notes that before integration users should independently verify the output schema and the currency/validity of the regulations.
Read Moredsh-plugin-ripple: Empowering DSH Agents to Engage in the Ripple Social Network as Independent Accounts
dsh-plugin-ripple is a DeepSeek Harness networking tool plugin maintained by EdwinZhanCN (v0.1.3, MIT). It writes events from the Ripple self-hosted social network into DSH persistent sessions via WebSocket/polling, and enables agents to autonomously reply, post, and follow through the mcp__ripple__* tools; it also includes a ripple skill and a self-configuration workflow. It requires Node.js≥24.9, DSH 0.2.1-rc.5 up to (but not including) 0.3, and web/desktop profiles. This article covers the installation commands, connection token, configuration options, and security constraints.
Read MoreExam Paper Review Assistant: Transforming Test Paper Photos into Structured Personalized Feedback
Exam Commentary Assistant (@user_8c79bf8b/exam-commentary, v1.0.0, community source, category: education) is a general-purpose Agent Skill formatted as SKILL.md, designed for single-session exam paper review scenarios. In this workflow, students upload photos of their exam papers or questions, and the AI automatically grades them, analyzes knowledge mastery, and diagnoses the root causes of errors. It then provides personalized learning suggestions across three dimensions—knowledge reserve, answering techniques, and competency requirements—and outputs the results as a Word document. The article details its functional boundaries, the specific roles of the three-dimensional suggestions, and the SkillHub page data (Rating: 100,000; Downloads: 1,002; Installs: 0; requires_api_key=false). It also addresses the absence of installation and invocation details caused by the inability to scrape the official page. Consequently, the text avoids fabricating installation commands, configuration items, or examples, instead outlining only the input and output formats, while listing applicable scenarios and caveats such as the necessity for manual review of automatic scoring.
Read Moredsh-remote-editor: Open DSH Session Workspace in Local Cursor / VS Code via Remote-SSH
dsh-remote-editor is a DSH web client plugin maintained by Edison-A-N. When the browser and the DSH server are on different machines, the official "Open In" option cannot open an editor locally; this plugin adds cursor:// or vscode:// Remote-SSH links to the session toolbar, allowing you to open the current session's workspace locally in Cursor, VS Code, VS Code Insiders, or Windsurf. It supports selecting the editor and SSH host, previewing and copying links, with settings stored in the browser's localStorage. Installation: dsh plugin --profile web add dsh-remote-editor. Requires DSH 0.2.x, Node.js ≥ 22.19, and local SSH with the Remote-SSH extension. Licensed under MIT.
Read MoreAmazon Variation Analysis: Color & Size Review Comparison - Identifying Underperforming Variants
This article introduces the community-sourced Skill "Amazon Variant Analysis · Color/Size Reputation Comparison" (`@user_507a5056/variant-analysis`, Version 1.5.2, Category: business-ops) published on SkillHub. It approaches the analysis by starting with reputation pain points of the parent SKU, comparing reputation differences across colors, sizes, and specifications under the same parent body to locate problematic variants dragging down overall ratings and identify high-performing advantage variants that drive sales volume. This provides a basis for decisions on variant discontinuation, inventory adjustment, and promotional focus shifts. Core functions include variant reputation comparison, "rating-drag" variant localization, and comprehensive ARI review analysis capabilities. Users can subscribe to ASINs to collect reviews, browse and filter reviews, view star ratings/keywords/trends and radar charts, generate analysis reports such as VOC and deep insights, run product operation workflows, monitor product changes, utilize negative review alerts and a negative review workbench, perform industry benchmarking and category ranking, and export review CSVs along with reports in Markdown/HTML formats. The installation section notes that this Skill is tagged with `requires_api_key=true` and requires an ARI API key (ari_live_* format). As the two official addresses failed to be fetched during this capture, the installation commands and configuration processes are marked as unverified. Typical usage provides four reproducible request examples based on official trigger scenarios. The article concludes by listing applicable scenarios, limitations, potential pitfalls, and the official addresses.
Read Moredsh-archived-session-delete: Permanently delete archived sessions in the Web GUI sidebar
This document introduces the community plugin `dsh-archived-session-delete`: it adds a "Delete Session" option to the sidebar menu for archived sessions in the DeepSeek Harness Web GUI. After a secondary confirmation step, the Host cleans up the JSONL directories, sub-agent logs, Workspace, and archive/pinned collections, then refreshes the list. Includes the npm installation command, optional `sessionsRoot` configuration, plan/delete RPC, and documentation on security restrictions. Maintained by Ediblewildfungi, MIT licensed, version 1.0.0 on npm.
Read MoreECC: it-ops-security Community Skill on SkillHub
ECC is a community Skill published on SkillHub by user_3c6cb52e (version 1.0.0), categorized under it-ops-security, with a rating of 100000, 1007 downloads, 0 installs, and the tag requires_api_key=false. The official description is just a single line, ECC skill, with no functional details or examples. This article introduces the source, interpretation of metadata, installation entry point, and limitations based on scraped metadata, and emphasizes that users should read SKILL.md and verify before using.
Read MoreSAP Consultant Work Assistant: Turning Repetitive Business Tasks into Structured Results
This article introduces the SkillHub enterprise-class Skill "SAP Consultant Work Assistant" (@org-02qudk26/sap-consultant, Version 1.0.0, Publisher: Zhiyu Technology, Category: business-ops). Targeted at white-collar workers, it handles repetitive tasks related to enterprise business and outputs structured, reusable work deliverables. The article outlines its core functions, installation and activation considerations based on public information, and typical usage steps. It also cautions users that the public information lacks specific installation commands, prompts, or SAP system interfaces, and that discrepancies exist between the ratings and download counts in the clues and the records on the official page; therefore, the official page should be treated as the authoritative source.
Read MoreDSH Session Keep-alive: Bounded Per-Tab Preservation of Open Sessions
edgora-ai's client plugin, dsh-session-keepalive (v1.1.2), is compatible with DSH 0.2.1-rc.3/rc.4. It utilizes native session retention to keep open sessions in each browser tab, defaulting to a maximum of 8 sessions, which are released after 30 minutes of inactivity. It performs no prefetching and does not duplicate chats. The article outlines the tarball installation command, "warming" settings, tab suspension, privacy diagnostics, and the Agent tool `session_keepalive_settings`, while also detailing permissions, memory usage, and network boundaries.
Read MoreAcademic Literature Retrieval and Citation Tool: Accessing DBLP via MCP
This article introduces an MCP-based DBLP literature retrieval and citation Skill, "Academic Literature Retrieval and Citation Tool" (@org-rn88lg3j/db-lp), published by Xiaobenyang Technology (version 1.0.1, source: enterprise, category: professional, requires an API Key). The article describes its positioning, three core capabilities—literature retrieval, citation generation, and formatting—pre-installation and configuration checkpoints, applicable scenarios and considerations, and provides the official URL for verification.
Read MoreOne-Click Pro Generation of Academic Paper Figures: Cloud Rendering of 22 Academic Illustrations
"One-Click Thesis Figures Pro" is a cloud-based rendering service for academic figures. Maintained by community developers, it caters to authors who require publication-quality graphics without the need to set up a local Python environment. It supports 22 figure types, including Forest Plots, KM Survival Curves, ROC Curves, PRISMA 2020 Flowcharts, Box plots, Violin plots, Heatmaps, Cluster Heatmaps, Venn Diagrams, Funnel Plots, Bland-Altman Plots, PCA, Pair Plots, Scatter plots, Line charts, Bar charts, Dual-axis charts, Combined charts, and Technical Roadmaps. The service features built-in layout presets compliant with the *Chinese Medical Association* journal specifications (unofficial product). Outputs are provided as dual files in TIFF (600dpi) and PNG formats. A local data pre-validation system ensures users are not charged if the data is invalid. No Python installation is required; simply upload data to generate charts instantly. Utilizing the same rendering engine as the free "academic-figures" version, the service operates on a pay-per-use basis with Alipay support and requires no API key.
Read MoreFinancial Modeling Assistant: Organizing Repetitive Tasks into Reusable Workflows
The Financial Modeling Assistant (`@org-02qudk26/financial-modeling-nt`, v1.0.0, source: enterprise, category: professional, publisher: Zhiyu Technology) is an Agent Skill designed for repetitive tasks related to white-collar office efficiency. Its objective is to rapidly generate structured and reusable work deliverables without requiring an API Key. This article is organized into six sections: "What is this," "Core Capabilities," "Installation and Activation," "Typical Usage," "Applicable Scenarios and Considerations," and "Conclusion." It notes that the tool only promises the organization and construction of repetitive tasks and has not published a detailed feature list; therefore, unverified capabilities are not elaborated upon. The article provides a general activation path involving the manual placement of the SKILL.md file into the skill directory and candidly states that no one-click installation commands or differentiated configurations for Cursor, Codex CLI, or Claude Code are officially provided. The "Typical Usage" section offers a reproducible example of a three-part modeling instruction, focusing on the structural accumulation found in Step 4. The "Considerations" section clarifies that it cannot replace accounting judgment, input criteria must be explicitly stated, and cross-table reconciliation requires manual verification.
Read Moredsh-audio-player: MP3 Floating Window Player in DSH
## dsh-audio-player is a DSH client plugin maintained by edgewalker6666. It plays user-imported MP3 files in an independent floating window within the Harness, with the music library and settings persisted to `~/.dsh/dsh-audio-player`. The player utilizes its own HTMLAudioElement and does not interfere with DSH, voice features, or other plugins. The optional "Startup animation synchronization" prioritizes integration with `dsh-startup-secret` and `dsh-theme-prts`, fading out and stopping once the animation concludes. Web installation: `dsh plugin --profile web add github:edgewalker6666/dsh-audio-player`; desktop version available via `install.ps1`. Requires DSH 0.2.1-rc.3, Node >= 20. MIT License.
Read Moredsh-usage-replay: Cross-Session per-Model Replay of Token Usage
dsh-usage-replay is a model inference plugin developed by Eclipse-NotFound for DeepSeek Harness. It replays session logs in ~/.dsh/sessions, aggregates provider-reported token usage by provider/model, by day and workspace, and provides a sidebar usage Tab, per-model chips in Composer, and two local read-only HTTP APIs. It only counts tokens, with no cost estimation or network requests. Requires DSH 0.2.1+; the usage Tab optionally depends on dsh-better-sidebar ≥0.28. Install: dsh plugin --profile <profile> add dsh-usage-replay.
Read Moreverified-search-pro: Adding an Information Quality Verification Layer to AI Agents
verified-search-pro (VSP) is a community-published quality assurance layer Skill, version 2.1.0, categorized under data-analysis, which requires API Key configuration. It integrates intelligence methodologies including OSINT, ACH, SIFT, and the Admiralty dual-axis rating, and aggregates multiple search engines such as Tencent Cloud WSA, Baidu Qianfan, Tavily, Bing, Sogou, and DuckDuckGo. It evaluates search results and assigns A–E confidence levels based on source reliability, content credibility, the number of independent confirmations, and disconfirmation status. It outputs both human-readable Markdown reports and machine-readable claims-json/evidence-pack files. VSP supports two modes: configuring full APIs as the search foundation, or using the --input-results parameter in quality inspector mode to verify results obtained by the host. It is well-suited for in-depth research, fact-checking, policy tracking, and competitor background checks.
Read More