设计与三维建模

Image toolkit: resize, compress, crop, watermark, convert, rotate, EXIF read/strip.

Checks a TLA+ model of something that moves between states (a checkout, an upload, a protocol, a job with retries) with the TLC model checker and draws it as a Mermaid statechart in the reader's words, marking each state that breaks a rule and the shortest path to it. Use when asked to 'draw the states', 'chart how this flow works', 'statechart', 'can this ever happen', 'check this flow keeps the rule', 'model the lifecycle', or to show a bug about order, duplicates, retries or late events as a path through states. Declines work with no states, such as a pure function. Not for architecture or sequence diagrams (vd:archify, vd:diagram-design).

教授智能体使用自主接力循环模式，通过 Stitch 迭代构建网站。

通过串联静态 HTML 提取、设计系统提取和文件上传，将前端代码（Vite、React、Angular、Vue 等）转换为 Stitch Design。当用户意图是将现有 Web 应用或 React/Angular/Vue 组件迁移到 Stitch 时（例如请求“保存”“迁移”或“上传”），**务必**使用此技能。即使是简单的“保存”操作也必须使用此技能，因为这是确保提取设计系统并正确关联资源的唯一方式。

直接从前端源代码中提取完整的设计系统（DESIGN.md）——React、Vue、Svelte、Angular、纯 HTML/CSS 或任何 Web 框架。分析组件文件、样式表、Tailwind 配置、主题定义和设计令牌，生成内容丰富、与 Stitch 兼容的设计系统文档。当用户希望从现有代码库逆向还原设计系统、审计项目的视觉语言、从源文件提取设计令牌，或理解前端仓库中的样式模式时，使用本技能——即使用户只是说“这个应用长什么样？”或“从这段代码中提取设计”。

通过内联 CSS 和图像，从已构建的 Web 应用或 React 组件中提取自包含的静态 HTML。每当你需要捕获特定 UI 状态、分享页面的静态版本，或为 Stitch 上传准备素材时使用此技能，即使用户只是要求“保存 HTML”或“模拟视图”。

通过 Stitch MCP 根据文字提示或图片生成新屏幕，用提示词与设计系统变量编辑已有屏幕，并生成设计变体。涵盖提示词增强流程、设计映射、专业 UI/UX 术语、设计变量和主题系统能力。

使用 MCP 工具管理 Stitch 中的设计系统，包括获取素材、创建设计系统或更新现有系统，并应用到屏幕。

将本地资源（图片、设计稿、提取的HTML、设计Markdown）上传到Stitch项目。当需要将视觉资源、HTML页面或设计文档上传到Stitch时，务必使用此技能，尤其是在直接调用MCP工具因base64令牌限制而失败或被截断时。

Design custom T-shirts, preview artwork, personalize team shirts and prepare a checkout link.

The director of Full Production Studio. Takes a raw video idea and runs it through the whole pipeline (idea → script → visual plan → motion → edit → captions → export), picking the right skill at every stage and keeping one project brief in sync. Use when the user wants to make a whole video end-to-end, says "make me a video/reel/short about…", "where do I start", "اعملي فيديو عن…", "عايز أعمل ريل", or is unsure which skill to use.

The director of Full Production Studio. Takes a raw video idea and runs it through the whole pipeline (idea → script → visual plan → motion → edit → captions → export), picking the right skill at every stage and keeping one project brief in sync. Use when the user wants to make a whole video end-to-end, says "make me a video/reel/short about…", "where do I start", "اعملي فيديو عن…", "عايز أعمل ريل", or is unsure which skill to use.

当多个独立视角可能改变决策时使用：多领域设计、原因不明、架构权衡，或需要新证据的阻塞问题。分配不同的证据角度，由责任负责人决定。对于明确任务，使用单智能体直接处理；对于无需额外视角的恢复工作，使用解除阻塞。

使用actor在Swift中实现线程安全的数据持久化——内存缓存配合文件存储，从设计上消除数据竞争。适用于在Swift中持久化数据并需要从设计上消除数据竞争或线程安全问题的情况。

使用actor在Swift中实现线程安全的数据持久化——内存缓存配合文件存储，从设计上消除数据竞争。适用于在Swift中持久化数据并需要从设计上消除数据竞争或线程安全问题的情况。

使用 actor 在 Swift 中实现线程安全的数据持久化——内存缓存配合文件存储，从设计上消除数据竞争。

Polished system design diagrams (SVG/PNG) from a tiny YAML spec, ~18x fewer tokens than raw SVG.

Design scalable systems: requirements, architecture, data models, trade-offs, and scaling strategies. Use for architecture and interviews.

每当出现故障时，在提出或实施任何修复之前，你必须使用本技能，包括快速修复以及用户赶时间的情况：缺陷、失败或红色的测试、构建或 CI 失败、崩溃、性能问题、异常行为。它分四个阶段（调查、对比、验证单一假设、先写失败测试再修复）找出根本原因，而不是猜测；在三次修复失败后停止并质疑设计。不适用于编写新功能。

使用 Tailwind CSS 构建可用于生产的设计系统，涵盖设计变量、组件变体、响应式模式和无障碍支持。

Tailwind CSS v4 原则。CSS-first 配置、容器查询、现代模式、设计令牌架构。

Use when building flex or grid layouts, positioning elements (absolute/relative/fixed/sticky), managing z-index/inset, or @container queries.

Use when setting font family/size/weight, aligning or wrapping text (balance/pretty), applying text-shadow or decoration, or letter/line spacing.

Find AI design tools, compare catalog records, and inspect stored evidence and limitations.

使用 Markdown 复选框执行设计文档中的实现任务。适用于：(1) 实现 feature-design-assistant 输出的功能；(2) 恢复中断的工作；(3) 批量执行任务。由“start implementation”“run tasks”和“resume”触发。

面向 angelcore / cloud-trance / hyperpop 系音乐视频和短视频剪辑的创意指导层——包含具名流派的美学词汇、情绪+色彩+光线系统、节拍同步剪辑语法，以及串联 ECC 视频技能（从 b-roll 生成到分发）的流水线。当视频必须显得有意图而非仅仅功能可用时使用——音乐视频、粉丝剪辑、情绪板驱动的短视频，或为 AI-generated b-roll 赋予连贯的视觉方向。

基于蒸馏风格包生成新视频并剪辑为成片——依据参考素材的剪辑节奏规划镜头，在 fal 上生成，用风格包实测 LUT 调色，按实测节奏剪辑，融入现有素材，合成叠加图层，铸造 3D 道具，并对结果进行数值验证。适用于用户想以捕获的风格制作视频、补充现有素材，或将生成的片段组装成真实剪辑。

面向 Google Stitch 的语义设计系统技能。生成对智能体友好的 DESIGN.md 文件，强制执行高端、反通用化的 UI 标准——严格排版、校准色彩、非对称布局、持续微动效以及硬件加速性能。

用于文件驱动的多模态图像、视频和 3D 资产发现；品味访谈；提炼或应用工作流；风格包验证；可编辑的 EDL/FCPXML 导出；来源审计；以及在提供商生成前必须失败关闭的离线规划。

Define a multi-agent team for ANY task on ANY system as an explicit deployment topology - pick the shape from the task's dominant risk, specify the runtime/communication/trust layers, place model capability by lane, present it as a diagram + table + trust-boundary note + open choices, and deploy ONLY after the user agrees to the proposed shape. Use when the user asks to "define/design/lay out the team", "what topology are we deploying", "how should the agents be arranged", "design the team for <task>", "what formation for <task>", or wants a team expressed as a topology (nodes, edges, models, trust boundaries) rather than spawned ad hoc. Substrate-agnostic - Cotal mesh, harness subagents, Workflow stages, containers, or any orchestration system.

Define a multi-agent team for ANY task on ANY system as an explicit deployment topology - pick the shape from the task's dominant risk, specify the runtime/communication/trust layers, place model capability by lane, present it as a diagram + table + trust-boundary note + open choices, and deploy ONLY after the user agrees to the proposed shape. Use when the user asks to "define/design/lay out the team", "what topology are we deploying", "how should the agents be arranged", "design the team for <task>", "what formation for <task>", or wants a team expressed as a topology (nodes, edges, models, trust boundaries) rather than spawned ad hoc. Substrate-agnostic - Cotal mesh, harness subagents, Workflow stages, containers, or any orchestration system.

Your project's design system for coding agents: tokens, components, contrast checks, guidelines.

実装前に呼ぶ。High/Medium リスク変更のテストオラクル（期待値の根拠）とファルシフィケーション項目を定義する。

Draw reliable, well-aligned text/ASCII architecture diagrams with nested boxes, arrows, and connectors. Use when user asks to draw/create text diagrams, ASCII art diagrams, box diagrams, or architecture diagrams in plain text. For rendered SVG or HTML diagrams, use vd:diagram-design instead.

为幻灯片、文档、报告、HTML 落地页等产物设置主题样式的工具包。提供 10 个预设颜色/字体主题，可应用到任何已创建产物，也可即时生成新主题。

在规划或做出架构、设计、战略决策前，运行虚拟智库，即结构化的多角色辩论。用户准备规划系统、选择技术、评估取舍、决定方案，或任何可借助多视角改善结果的决策时使用。用户说“think tank”“debate this”“perspectives on”“trade-offs”“should I use X or Y”“help me decide”“before we plan”，或要求比较不同方案优缺点时也触发。必须在任何实现规划开始前运行，产出结构化分析以改进后续计划。

Works out what could go wrong in a system's design before it is built — who would attack it, how, and which defences are worth having. Use this whenever the user is designing a system that handles sensitive data, money, or authentication, asks what the security risks are, mentions threat modeling or STRIDE, or is starting a feature with a meaningful attack surface. This is design-time and systematic; for finding vulnerabilities in code that exists, use security-analysis.

Turn text or an image into an animation-ready 3D model (GLB): generate, rig, animate, retexture.

Free text/image → 3D: generate, rig, avatar-ify, and refine GLB models. No auth, no payment.

Text-to-speech, speech-to-text, audio-to-face lipsync, and motion-capture clips for 3D agents.

Drive a local Blender headlessly: inspect, convert, render and script 3D files.

Extract validated image data URIs from Solana SPL Memos and return them inline to MCP clients.

Image understanding for AI agents — analyze and describe any image via the three.ws pipeline.

Free text/image to 3D for Grok Bot, plus your three.ws agents once signed in. Never spends.

Convert a confirmed design/requirements doc into a persisted markdown task list with per-task status checkboxes, for projects that have no issue tracker (private/local-only repos). Use when a design doc has just been approved and the project has no GitHub issues (e.g. local-only git repos holding sensitive data), instead of to-issues/to-prd.

Aide's skills, the per-project .aide/config keys, how a project's own test/lint/build commands are detected, and where specs are stored. Use when: running a project's test, lint or build command; looking up which Aide skill does what; reading or writing .aide/config; deciding where a spec belongs. Do NOT use for: how to run tests correctly (that is the testing rule), the spec files' own layout (use the spec-structure skill).

将音频文件转写为文本，可选说话人分离和已知说话人提示。当用户要求转写音频/视频中的语音、从录音中提取文本，或标注访谈或会议中的说话人时使用。

将音频文件转写为文本，可选说话人分离和已知说话人提示。当用户要求转写音频/视频中的语音、从录音中提取文本，或标注访谈或会议中的说话人时使用。

Transform raw podcast transcripts into polished, readable documents. This skill removes filler words, fixes grammar, adds structure, and preserves authentic voice. Use when processing podcast transcripts, interview recordings, video captions, or any spoken content that needs cleanup while maintaining 100% fidelity to meaning.

处理自然语言、计算机视觉、音频或多模态任务的预训练 transformer 模型时使用。用于文本生成、分类、问答、翻译、摘要、图像分类、目标检测、语音识别，以及在自定义数据集上微调模型。