AI & ML

shuotao Skill

專案品質檢核：圖紙編號一致性、詳圖元件完整性、視圖組織、參數正確性、MCP 系統健康度。觸發條件：使用者提到 QA、品質檢查、驗證、一致性、專案檢核、系統健康、quality check、review。工具：get_all_sheets、get_detail_components、get_all_views、query_elements_with_filter。

shuotao Skill

批次材質修改（複製原材質模式）：為每個 Type 的原材質建立複本，只修改複本的 Appearance diffuse color，保留 Graphics 顏色與其他屬性。適用於渲染白模（Enscape/V-Ray 看到白色，平面圖切割圖案維持原材質）。也支援復原（assign_existing_material）。觸發條件：使用者提到材質、material、白色模型、white model、渲染、render、批次換材質、統一材質、復原材質、restore material。工具：get_types_by_category、batch_set_material、assign_existing_material。

shuotao Skill

Wrap a standalone Revit IExternalCommand (.cs / DLL, WPF dialog allowed) into a revit-mcp MCP tool. Generates the C# command handler + switch case + TypeScript tool definition, builds, dual-path deploys, guides the restart, auto-verifies the tool is live, then commits (never pushes). Use when the user says "把這支 DLL/.cs 變成 MCP 工具", "封裝成 revit-mcp tool", "wrap this command as an MCP tool".

shuotao Skill

Switch a whole Revit project's display units in one tool call via the set_project_units MCP tool. Use when the user asks to change project units, "改專案單位 / 切公制 / 一鍵改單位", set Air Flow to m³/h for Taiwan §102, set plumbing units for 給排水 (L/min, mH2O, 1:ratio slope), or align units to a project standard. For Taiwan HVAC use mode='taiwan'; for Taiwan 給排水 use mode='taiwan-plumbing'.

shuotao Skill

Inventory and curate Revit's Manage → MEP Settings — pipe segment size catalogs, duct Rectangular/Oval/Round size tables, fitting angles, pipe slopes, fluids. Four-phase flow, catalog inventory → real model usage → standards reconciliation → add/remove. Use when the user says 盤點管徑 / 風管尺寸表 / segments and sizes / 這個尺寸有沒有人在用 / 刪掉沒用到的尺寸 / 增加管徑 / mechanical settings / CNS 對帳, names bare duct sizes like "2438 1371", or asks what sizes a model actually uses. Removal is guarded — never delete a size without running the usage scan first.

shuotao Bundle

Translates Revit-oriented BIM Skill intent and terminology into a guarded Archicad MCP workflow while preserving the original Domain method and Revit route. Use when applying BIM_MCP Skills to Archicad, selecting a BIM backend, querying Archicad elements, numbering Zones, producing Archicad quantity takeoffs, or handling Revit→Archicad 轉譯、名詞對照、GUID、Story、Zone、Layout、Tapir。

shuotao Skill

跨專案複製含 model views 的圖紙：Revit 原生「Insert Views from File」無法匯入含 model view 的 sheets，本 skill 用兩階段流程（讀取→確認衝突→執行）自動化重建。處理 view 四級分類（Tier 1 自動建立 / Tier 2 部分建立 / Tier 3 只匹配 / Tier 4 人工處理）、ScheduleSheetInstance 重建、DraftingView 內容跨文件複製、Sheet custom parameters（圖集/Discipline/繪圖人等）一併複製、衝突解決需使用者確認。內含補丁工具 sync_sheet_parameters_from_source 用於補齊舊 sheet 的 metadata。觸發條件：使用者提到跨專案複製圖紙、複製 sheets、Insert Views from File、轉移圖紙、source.rvt、import sheets、移植視圖、把另一個 rvt 的圖紙複製過來、cross-document、cross-project、補 sheet folder、修圖紙分組。工具：read_source_file_sheets、copy_sheets_from_file、sync_sheet_parameters_from_source。

goodchengwang Bundle

Extract reusable prompts or workflow specifications from user-provided URLs and create new source-traced Codex skills in the current project. Use when the user invokes $url-to-skills with a URL, especially for X posts, quoted prompt chains, image Alt prompts, or numbered prompt series; do not use for ordinary link summaries or for modifying existing skills.

stylabs Skill

Keep Hostaway as the PMS. Do not migrate operators off Hostaway or push Hostaway's website builder. Use hostaway-kit MCP tools for listings, calendar, inbox, and reports.

emskills-mcp Bundle

Measures d-spacings and fits atomic lattices from high-resolution STEM images using Bayesian inference, reporting posterior intervals for each lattice parameter.

manhvann Bundle

Pack repositories into AI-friendly files with Repomix (XML, Markdown, plain text). Use for codebase snapshots, LLM context preparation, security audits, third-party library analysis.

manhvann Bundle

Build 3D web apps with Three.js (WebGL/WebGPU). 556 searchable examples, 60 API classes, 20 use cases. Actions: create 3D scene, load model, add animation, implement physics, build VR/XR. Topics: GLTF loader, PBR materials, particle effects, shadows, post-processing, compute shaders, TSL. Integrations: WebGPU, physics engines, spatial audio.

manhvann Bundle

Bootstrap new projects with research, tech stack, design, planning, and implementation. Modes: full (interactive), auto (default), fast (skip research), parallel (multi-agent).

dga-sd Bundle

ถอดเสียงประชุมภาษาไทย (.m4a/.mp3/.wav และทุกฟอร์แมตที่ ffmpeg อ่านได้) เป็น transcript พร้อม timestamp ในเครื่องด้วย Typhoon ASR (scb10x/typhoon-asr-realtime) รองรับไฟล์ยาวระดับชั่วโมงโดยหั่นท่อนตามช่วงเงียบ มีโหมดแยกผู้พูด (--speakers, pyannote diarization) และโหมดใส่ชื่อจริงแทน speakerN โดยวิเคราะห์บริบทใน transcript ประกบรายชื่อผู้เข้าประชุม ข้อมูลเสียงไม่ออกจากเครื่อง Use this skill whenever the user provides a Thai meeting audio file and asks to transcribe it (ถอดเสียง / ถอดความ / transcribe), with or without speaker separation (แยกผู้พูด) or speaker naming (ใส่ชื่อผู้พูด / ใส่ชื่อ / map ชื่อ). Also use the naming workflow (section 4) when the user already has a .speakers.transcript.txt and asks to replace speakerN with real names. Do NOT use for non-Thai audio (the model is Thai-only) and do NOT use for summarizing minutes (that is the meeting-summary skill, downstream of this one).

vassiliylakhonin Skill

Additive Codex runtime overlay for the central-asia-caspian root skill. Use only after loading the root contract when work runs as a multi-step agent loop, writes files, or chains into validation.

manhvann Skill

Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.

vassiliylakhonin Skill

Additive Codex runtime overlay for the gulf-middle-east root skill. Use only after loading the root contract when work runs as a multi-step agent loop, emits evidence-packet JSON, writes files, or chains into validation.

the-biomechanist Bundle

Forensic debugger for messy Windows dev setups. Use when users report in casual language that installs "look done" but still fail, PATH/shell behavior is inconsistent, tools only work in one terminal/editor, or prior-agent changes cannot be verified.

manhvann Bundle

Build AI agents with Google ADK Python. Multi-agent systems, A2A protocol, MCP tools, workflow agents, state/memory, callbacks/plugins, Vertex AI deployment, evaluation.

manhvann Bundle

Check context usage limits, monitor time remaining, optimize token consumption, debug context failures. Use when asking about context percentage, rate limits, usage warnings, context optimization, agent architectures, memory systems.

rowland-dot Bundle

复用 SOP 小红书扫描浏览器、Tampermonkey 和本地脚本，按需同步、去重、分类并沉淀小红书收藏为 Obsidian 知识库。适用于主动触发收藏整理、收藏面板管理、知识卡片生成、资源索引和行动清单维护；运行时不依赖 Codex、Claude 或任何特定 AI Agent。

greatbigm Bundle

新 skill 生成构成规范：文件层级、构成约束、打包整理，生成新技能前必读。

pax-beehive Bundle

Use paxm as an active agent memory layer. Trigger when the user asks to recall prior context, search or inspect memory, remember working state or a durable fact, debug paxm history/metrics, or when a task would benefit from active memory recall before answering, especially repo, project, preference, architecture, or previous-decision questions.

greatbigm Bundle

skill 审查：构成审查 + 发布前审查双清单，发布前必读。

studiogo Bundle

Neutralizuje zakotwiczenie (anchoring) w wycenach i szacunkach AI. Gdy w kontekście jest podsunięta liczba (cena klienta, oferta konkurencji, budżet, czyjaś sugestia), model dryfuje za nią i ta sama praca bywa wyceniana wielokrotnie różnie. PriceGuard izoluje kotwicę i daje wycenę niezależną. Użyj, gdy pytasz o wycenę, stawkę lub szacunek liczbowy, a w prompcie albo kontekście pojawia się czyjaś liczba (cena, budżet, stawka, oferta) — np. „ile to wycenić", „ile wziąć za to", „czy ta cena jest ok", „klient proponuje X, ile policzyć", „ile to realnie warte".

emskills-mcp Bundle

Reference-free 2D classification of cryo-EM particle micrographs. Includes motion correction, contrast transfer estimation, and clustering with configurable binning and angular sampling. Exposes a typed MCP toolset for batch runs against a session directory.

manhvann Bundle

Orchestrate Agent Teams for parallel multi-session collaboration. Use for research, implementation, review, and debug workflows requiring independent teammates.

manhvann Skill

AI agent orchestration board for task visualization and team coordination.

topprismdata Bundle

Use HuggingFace's hf-agents to run local AI coding assistants. Translates natural language requests into setup, model selection, launch, monitoring, and troubleshooting actions for local LLM-based coding agents. Covers hardware detection, model recommendation, llama-server management, and Pi coding agent integration.

yongkyu4803 Bundle

Installs skills or MCP servers from a GitHub repo, and moves skills between Claude Code, chat, and Cowork. Use for install requests, or when a skill is missing from one of them.

ricardodantas Bundle

Learn and explain an unfamiliar codebase by first discovering which agent skills are installed, then applying the ones relevant to this repo's stack to map its architecture, modules, data flow, conventions, and how to build/run/test it. Produces an in-conversation briefing and a saved docs/CODEBASE_OVERVIEW.md. Use when the user asks to "understand/learn/explain this repo", "onboard me to this codebase", "give me an overview of the project", "how does this codebase work", or when starting work in a repo you have not seen before.

lexler Bundle

Iterative code refactoring through progressive lenses via a worker-reviewer agent team.

ricardodantas Bundle

Reconcile a repository's documentation with its current state — README, docs/, guides, CHANGELOG/release notes, agent docs (AGENTS.md, CLAUDE.md, .claude/, skill files), API/reference docs, and inline comments — and wire them together, fixing broken links and orphaned pages so every doc is reachable. Orchestrates learn-codebase to model the repo, writing-for-agents to edit agent-facing docs, and find-docs (Context7) for version-specific library/framework/CLI details. Use when the user asks to "update the docs", "sync docs with the code", "docs are out of date", "refresh the README/AGENTS.md", "document recent changes", "link the docs together", "fix broken doc links", "nothing links to this doc", or after features land before a PR or release.

luffy666code Skill

Use when: the user asks to make or render charts with flint-chart, visualize tabular data, generate a ChartAssemblyInput, validate/render through MCP, or add Flint to a JS/TS project. Author the semantic spec, transform data before Flint when needed, install/import Flint only when executable code is needed, and reserve backend-specific style tweaks for after compiling from Flint.

mahsan07 Bundle

Control a running TouchDesigner instance through MCP to create operators, set parameters, wire networks, and execute Python. Use when the user explicitly requests this capability or a closely related task.

mikepans1013 Bundle

Install, configure, verify, and troubleshoot the official GitHub MCP server for an OpenClaw bot, including auth identity checks, permissions, tool-loading verification, and write-action guardrails.