Turtle Diagram Process Skills: Applying the IATF Process Approach to the Diagram
Introduction to the "Turtle Diagram Process Skill" on SkillHub: coordinates @engicool/process-turtle, published by Engicool, version 1.0.3, categorized under design-media. Its core capability is generating IATF process approach turtle diagrams, supporting templates such as turtle diagrams (process approach six-box), swimlane diagrams, flowcharts, and timelines. It is suitable for drawing process turtle diagrams, SIPOC analysis, or exporting process visualizations. The tags indicate requires_api_key=false, meaning no API Key configuration is required. The article provides the official API and page addresses, and notes that the current SkillHub page displays a rating of 100000, 842 downloads, and 0 installations. Since the official page content could not be directly retrieved, installation details and template examples should be based on the official address.
Read Moredsh_workflow_fang: Running Governable Multi-Agent Workflows on DSH 0.2.1-rc.3
"PatTianFang/dsh_workflow_fang" is an adapted version of the upstream "omdsh-dev/dsh_workflow v0.1.2" tailored for "DeepSeek Harness 0.2.1-rc.3" and licensed under MIT. This plugin provides the "/workflow" slash command along with three tools: "workflow_list", "run_workflow", and "workflow_manage". It supports built-in workflows, six pattern types, and user-defined saved workflows, as well as run logging, pause-and-resume functionality, and scoped review capabilities. The installation command is `dsh plugin --profile desktop add "github:PatTianFang/dsh_workflow_fang#main"`, and Node.js version 22.19 or higher is required. As this is a non-official plugin, please review the source code and permission scope before using it.
Read More"Store Operations Daily Digest | Guo Yiyi: Simplifying daily store data into digests and next-day actions"
Introduce the community Skill “Store Operations Daily Report | Guo Yiyi”. It takes sales, foot traffic, inventory, staffing, and exception logs as input, and generates store daily reports, anomaly explanations, and next-day action suggestions, targeting store managers and regional operations personnel. Version v1.0.0, category office-efficiency, from community, no API Key required. The article outlines what it can do and what it cannot do, and highlights that users must prepare their own data and that next-day actions are suggestions only. The official page does not publicly provide verifiable installation commands, and this is stated honestly in the article.
Read MoreEnterprise Knowledge Base Building Consultant: Breaking "Building a Knowledge Base" Down into an Actionable Plan Card
Addressing the common trial-and-error approach of "connecting a vector database first" when enterprises attempt to transform fragmented documents, tickets, emails, and databases into AI-answerable knowledge assets, this article introduces the "Enterprise Knowledge Base Construction Consultant" Skill released by Kunlun Growth. It outputs an engineering methodology and a construction plan card covering six key stages: source data audit, segmentation strategy, metadata schema, hybrid retrieval, evaluation metrics, and governance operations. The article clarifies that it is not a runtime component and does not replace code implementation or data cleaning. It confirms verified metadata such as `requires_api_key=false`, notes that official documentation lacks installation commands and specific configuration differences for various Agent clients, and provides invocation examples constructed based on capability descriptions along with applicable and inapplicable scenarios.
Read More@pandayuuha/dsh-disable-preset-plugins: Configurably disable built-in plugins in native Agent presets
Introduces the community DSH plugin `@pandayuuha/dsh-disable-preset-plugins`: While keeping native presets such as 'standard', 'ptc', and 'minimal' unchanged, it marks specified built-in plugins as disabled via the Profile's `cordis.patch.yml`, supporting both global and mode-specific disabling. Includes a wizard, `dsh` CLI installation commands, configuration examples, and usage notes.
Read Moredsh-eli5: DSH Select-to-Explain without consuming main conversation context
dsh-eli5 is a DSH Web word-selection explainer plugin (v0.6.0) maintained by pandarayc. After selecting text, it invokes an auxiliary model via a standalone POST /eli5/ask call to explain in plain language, translate, find errors, or summarize. The explain and error-finding modes operate read-only within the workspace within a given budget, and they do not write to the Session. This article compares the two paths against the official README, describes the differences between the two routes, the radial menu, the installation patch, and typical usage. It also reminds users that the plugin runs with DSH process permissions.
Read MoreDistillation in One Sentence: Compressing Lengthy Content into a High-Concentration Summary
Distillation is a knowledge management Skill on SkillHub (located at @user_1f616208/yijuhua, version 1.0.0, source: community, tag: requires_api_key=false) that takes webpages, books, videos, images, or long-form text as input to generate a default high-concentration one-sentence summary, or full distilled content if explicitly requested. This article introduces its positioning, core functions, activation conditions, and use cases, and clarifies that since no installation command is provided in official documentation, no installation steps are fabricated.
Read MoreSupply Chain Audit: Generate SHA256 Manifest for Agent Plugin Supply Chains and Scan
Supply Chain Audit is a Skill listed on SkillHub with coordinates @org-02qudk26/chain, version 1.0.0, published by Zhiyu Technology, categorized under it-ops-security, and sourced from enterprise. It treats SKILL.md files, packaging scripts, MCP server definitions, hooks, and .claude/ configurations in the Agent environment uniformly as third-party software, generates a SHA-256 manifest and performs scanning for integrity record keeping and subsequent comparison. The official tag indicates requires_api_key=false, meaning no API key configuration is required. The official description is truncated at the phrase "generate SHA-256 manifest, scan", and specific targets, output format, and detection rules of the scan are not provided in the retrieved materials; the article notes that the official page should be consulted for accurate details. This skill is suitable for developers and security personnel who need to install third-party Skills, plugins, or MCP servers in Agent environments like Claude Code and establish integrity baselines.
Read Moredsh-bigfeiyu-focus: Big Fat Fish Gaze Follows the Mouse
dsh-bigfeiyu-focus is a standalone DSH plugin that adds mouse-gaze following to the web side of the dsh-pet "bigfeiyu", featuring perspective-based rotateY/rotateX transformations, parallax displacement, damped tracking, and an idle "looking around" behavior after 2.8 seconds of inactivity. It does not modify the core dsh-pet module and requires a browser for display positioning. Web installation: dsh plugin --profile web add github:panchalkmetkh249-alt/dsh-bigfeiyu-focus. MIT, v0.1.1.
Read MoreBid Document Review: A Single Skill Covers Both Integrity and Compliance
Bid Document Checker (@user_bda87ea5/bid-document-checker-james), released on SkillHub by user_bda87ea5, is a community Skill in version 3.0.0. It is classified under "business-ops" and tagged with attributes such as "requires_api_key=false". The skill performs checks and does not revise documents or generate content. Its core capabilities are divided into three levels. First, it offers built-in variance analysis across three bidding categories: engineering, goods, and services. Second, it provides five checklist categories: qualification, technical, commercial, formatting, and disqualification. These cover qualification certificates and performance thresholds, technical parameter responses and negative deviations, quotation, schedule, payment, and bid bonds, formatting requirements such as signatures, seals, binding, and page numbering, as well as mandatory clauses leading to direct bid rejection. Third, it includes a summary of common disqualification causes to prioritize the identification of high-risk issues. Typical trigger scenarios include verifying bid document completeness, checking responses against bidding requirements, and ensuring compliance with formatting standards. The article outlines the Skill's positioning and boundaries, the official access points retrieved, and four typical use cases derived from its functional scope: completeness checking, item-by-item response verification against bidding requirements, format compliance checking, and selecting check standards based on the bid document type. It also lists its target audience and five important notes: it is a verification tool rather than a generation tool; generic checklists cannot replace careful reading of the original tender documents; unreadable attachments may affect the results; community-sourced content requires manual review; and page statistical data should be re-confirmed when cited. As the official page does not provide specific installation commands or configuration fields, this was acknowledged transparently in the article.
Read MoreResume Optimizer: Resume Generation and ATS Optimization Skill for Job Search Scenarios
Resume Optimizer is an open-source Agent Skill on ClawHub, published by tomstools11. Currently at version 1.0.1 under the MIT-0 license, it consolidates four operations into a fixed workflow: resume creation, job-specific customization, analysis and improvement, and ATS-friendly PDF export. It includes three built-in format templates—chronological, functional, and combination—as well as an ATS optimization checklist and analysis scoring criteria. The article covers the core features, installation commands, typical invocation steps, and the usage of the PDF generation script, while also outlining applicable scenarios and limitations.
Read Moredsh-batch-review: DSH Batch Review — Submit Structured Conclusions in One Go
The DSH client plugin `dsh-batch-review`, maintained by palx666, provides the `batch_review` tool: Agents pass multiple pending items at once; users address them individually via the Web UI panel or native cards before submission, receiving `verdict`, `note`, `filled`, and `report_text`. The plugin automatically injects routing rules (using `batch_review` for multiple items, `ask_user_question` for single items), with unfilled items defaulting to fail-closed. The panel is served by the host at `/api/batch-review`, requiring no additional ports. Requires DSH 0.2.1-rc.3–rc.5 and Node.js 20+.
Read Moredsh-ppt-studio: Preview, edit deck.json, and import PPTX in the DSH sidebar
dsh-ppt-studio is a DeepSeek Harness client plugin that provides 16:9 DOM preview with thumbnail bars, in-place text/table editing, python-pptx element-level PPTX import, and a session-bound AI chat sidebar in the workspace `.ppt-studio/deck.json`. Installation utilizes the official release tarball via `dsh plugin --profile web add`. Import requires Python 3 and python-pptx, with optional LibreOffice rasterization for chart pages into read-only full-page images. Based on the repository README, this document explains usage, deck format, and known limitations, while noting that the plugin runs with the same permissions as DSH.
Read Moredsh-tribot-team: Using Casting to Switch Between Five Bot Methodologies in DSH
dsh-tribot-team is a DSH workflow plugin maintained by P3NN4L. It registers the /tribot skill and provides casting pills alongside the input box, allowing switching among Moderator, Researcher, Engineer, Steward, Reviewer, and Team modes, with zero-config, MIT-licensed ease. Install with: dsh plugin --profile desktop add dsh-tribot-team. After restart, use /tribot or the menu to call the casting. The README explains context economics guidelines and task cost comparisons for handoff documentation, and notes that it should be installed either with or instead of a Tribot skill-type installation.
Read Moredsh-study-ask: Follow-up questions via text selection in the main chat; independent right sidebar for inquiries without cluttering the conversation list.
`dsh-study-ask` is a pure client-side plugin for DSH maintained by P1gG0dG0d, compatible with desktop and Web versions 0.2.1-rc.3. In the main conversation, selecting text and clicking "提问" (Ask) opens a reusable Q&A conversation in the right-side native panel, with selected text automatically inserted as a quote. Q&A sessions are hidden in the left panel list via CSS. The plugin supports KaTeX rendering for quotes, deterministic session IDs, and cookie mapping. Installation involves linking a local directory as demonstrated in the example. Released under MIT license, version 0.1.0.
Read More"Drive Axle CAD Drawing Creation: Turning 'designing a drive axle' into actionable drawing outputs."
"Drive Axle CAD Drawing Generation" (@user_5453610a/drive-axle, v1.0.0, community source, design-media category) is an Agent-oriented Skill designed to converge colloquial requests—such as "drive axle," "generate drive axle drawings," "make a drive axle," or "drive-axle"—into a clear entry point for drafting tasks, thereby avoiding the need to explain requirements from scratch every time. The metadata indicates requires_api_key=false, meaning no key is needed for integration. Public information only provides trigger phrases and a brief introduction, without specifying the output drawing types, formats, or parameter input methods; specific details must be checked in the SKILL.md after installation. Additionally, no installation commands, directories, or configuration fields are provided, so users should refer to the official page for instructions. The retrieved metrics show a rating of 100000 and 874 downloads, but 0 installs. Due to the lack of available usage feedback, the output format and precision should be verified locally before formal production.
Read MoreTINKER: Desktop Toolbox Skill for AI Agents
TINKER is a desktop toolbox Command Line Skill designed for AI agents. It provides over 100 general capabilities in a plugin format, covering categories such as JSON, hash, base64, regex, files, clipboard, converters, and editors. It is maintained by community publisher user_39d41190 (located at @user_39d41190/tinker, Version 1.0.2), classified under dev-programming, sourced from the community, and requires no API Key. The core philosophy dictates that when a user's requirement can be met by an installed plugin (including scenarios involving user "@" mentions, references pointing to the Skill, explicit requests to use Tinker, or simply a need for desktop utilities), the agent should prioritize discovering and reusing plugins rather than rewriting one-off scripts. Verified capabilities include plugin lifecycle operations (open, list, scaffold), MCP tool calling, agent-browser UI debugging, and controlling the Tinker application from the CLI. As official installation commands and a complete plugin list are not provided in the materials—and should be referred to official sources—verification is required. The installation count is displayed as 0, indicating low deployment volume, so independent verification is recommended.
Read MoreGovernment Publicity Data Analyst: Clarifying Government Communication Data with Unified Standards
Retrospectives on government publicity effectiveness often stall due to inconsistent metrics: for a single policy interpretation, WeChat reports reads, Douyin reports plays, and offline events report attendance, resulting in three mismatched datasets. Furthermore, rising traffic does not necessarily imply public comprehension. This Skill, based on the universal SKILL.md format, is published and maintained by community publisher user_b4a8b79e (Version 1.1.0). Sourced from the community and categorized under data-analysis, it carries the tag requires_api_key=false. SkillHub displays a score of 100000, 882 downloads, and 0 installations. It employs a unified metric framework to analyze the reach, engagement, conversion, and service effectiveness of government openness campaigns, policy interpretations, event communications, and public service content. It is suitable for multi-platform data aggregation, metric alignment, content retrospectives, anomaly explanations, and executive briefings. In practice, it must adhere to the principles of minimum necessity, aggregation prioritization, and personal information protection. It distinguishes between platform exposure, authentic reach, and public service outcomes, avoiding the use of a single traffic metric to substitute for social impact, and refrains from manipulative profiling of vulnerable groups. Because the official page does not provide installation commands or configuration fields, this article does not include unverified procedural instructions.
Read Moredsh-done-alert: Windows Taskbar Completion Alert and Notification Sound
dsh-done-alert is a DSH desktop plugin (Windows) built on dsh-badge-notify. It overlays green/orange corner badges on the DSH taskbar icon and plays alert sounds for completions and to-dos, adopting the same meaning as the small indicator dots in the sidebar. It supports management via config.json and a loopback management API, and is compatible with dsh-study-ask. Install it with: dsh plugin --profile desktop add <local-absolute-path>.
Read MoreExpert Management 3.0: An Introduction to the AI-Empowered Skill for Expert Resource Management
Expert Management 3.0 — AI Powered, Ensuring Full Support from Experts is an Agent Skill published by the Baitong AI Service Platform, and it is an Agent Skill in a generic SKILL.md format (Version 1.0.0). It is categorized under business-ops in SkillHub; its source is tagged enterprise, and it is tagged requires_api_key=false. It is built for operations and marketing teams who maintain expert assets such as KOLs and KUs over the long term. It uses AI assistance to organize expert rosters, profiles, and follow-up records that are scattered across spreadsheets and documents, so that they can be turned into agent-executable, standardized workflows. The official description states that its capabilities include, but are not limited to, KOL management and KU management. The publicly available materials do not disclose a specific task list, input/output formats, installation commands, or differences in how they are enabled on each client. This article does not speculate or fabricate use cases, and introduces only verifiable information from the official page, while also highlighting key considerations such as zero installations and sensitive-data compliance considerations.
Read Moredsh-session-tabs: Switch Sessions with Browser-style Tabs at the Top of the DSH Web GUI
dsh-session-tabs is a pure client-side plugin for DSH that provides a full-width session tab bar below the Web GUI title bar: one tab per session, grouped by workspace folding. It supports closing tabs without deleting sessions, creating new sessions, and remembering the folding state via localStorage. Based on the repository README, this article outlines installation methods (Release tgz, jsDelivr, github:, and local directory), a table of typical operations, and engine and web platform limitations. It also reminds users that the plugin runs under the permissions of the local DSH process, so the source code and MIT license should be verified before installation.
Read Moredsh-chat-search: Use Ctrl+F to Search in DeepSeek Harness Conversations
dsh-chat-search is a DSH client plugin maintained by P-PPJ for Desktop 0.2.x that enables in-conversation keyword search with Ctrl+F in already open chats. Rewritten around the data-conversation-* DOM, it provides body text highlighting, previous/next navigation, a hit list, and incremental counts as results are found. It does not expand collapsed sections or change the reading position. For installation, use the Release or CDN (direct link or local file) of dsh-chat-search-1.0.1.tgz and refresh the page afterward. Note limitations: it may conflict with older packages of the same name, and collapsed content not present in the DOM cannot be searched.
Read MorePipeline Welding Procedure Guide: Welding Methods and Consumables Based on Material and Wall Thickness
Pipe Welding Process Guidance (@user_a8c636dd/pipe-welding, v1.0.1, Community Source, Professional Category) is a skill designed for industrial pipe welding process guidance. Based on pipe material, wall thickness, pipe class, and service conditions, it generates welding methods (GTAW/SMAW/GMAW/SAW), matching welding consumables, preheating temperatures, PWHT parameters, NACE MR0175 hardness control requirements, and WPS/PQR qualification requirements. It covers nine material systems, including carbon steel, low alloy steel, Cr-Mo steel, stainless steel, duplex stainless steel, and P91, while providing guidance on welding consumable selection for dissimilar metal welding. It supports scenario recognition; inputs such as "20# DN200 wall thickness 10mm, which electrode to choose?" or "What preheating temperature is required for 15CrMoG?" will automatically match welding solutions. Based on the universal SKILL.md format, it requires no API Key. This article introduces its positioning, core functions, installation and activation methods, typical trigger phrases, and applicable scope, noting that before formal engineering use, it must be reviewed according to applicable standards and welding procedure qualifications must be completed.
Read Moredsh-background-compaction: Background context summarization before threshold, zero-call swap-in for the next turn
Introducing OZONGO's DSH plugin dsh-background-compaction: it generates background summaries before reaching the proportional threshold defined by thresholdTokens, then, in the next turn's pre-step, it submits the checkpoint using the held summaries, thereby avoiding stalls caused by synchronous summarization in dsh-compaction-basic. It covers installation (pnpm and bundle), the /bcompact command, the settings page and config.json, the threshold math, acceptance steps, and the relationship with the native engine fallback.
Read MoreSoft Furnishings Styling Assistant: Clarifying Color Proportions, Furniture Dimensions, and Layered Lighting All at Once
The "Soft Decor Matching Assistant" is a general-purpose Skill in the SKILL.md format released by the community, located at @user_af163eaa/z-soft-decor. Version 1.0.1, published by user_af163eaa, it provides recommendations on 60-30-10 color ratios, furniture dimensions and traffic flow, and a three-layer lighting scheme (ambient, task, and atmosphere) based on style preferences such as Modern, Nordic, Japanese, Mid-century Modern, and Wabi-sabi, while requiring adherence to actual spatial dimensions. Tagged with requires_api_key=false, it does not rely on external services requiring an API key. No official installation command is provided in the documentation, and the installation count is 0.
Read More