技能路由图

AI 短剧五冠军的总控说明与统一交接契约。用户需要判断当前该用哪位冠军、五个冠军分别做什么、何时接手、交付什么，或需要在冠军之间交接项目、资产和连续性事实时使用；下级 Skill 只能在用户明确批准后调用。

AI 视频旧总路由兼容入口与审批门。用于识别五冠军正式链路之外的候选 Skill，并在调用任何下级 Skill 前向用户征询明确批准；不得自动分派或执行下级 Skill。

统一索引、审计和校验 AI 视频 Skill 路由。用于查询当前有效 Skill、冠军链路、上下游边界、使用证据、路由冲突、导出同步状态，或在修改 AI 视频路由前验证登记表。

为AI视频项目编写角色、场景、道具、首帧和故事板资产图提示词与资产清单。用于原创小说/剧本的资产前置；实际一键生成、下载和验收走ai-video-asset-production，有原片证据的转绘走mx-shortdrama路由；不直接生成生视频。

小说创作总控 Skill 的 AI 视频路由部署版 v1，完整保留来源 v5.5 总控正文。用于从零创建小说、生成或修改大纲、写单章或场景、续写、修订、红队审查、去 AI 味润色、状态维护、里程碑审查和整书完本审查；小说定稿后再由 ai-video-novel-to-script 接入 AI 短剧改编。

将小说、网文、大纲或人物设定改编为完整的AI短剧总纲、人物弧线、分集大纲、场次细纲、可拍剧本和资产需求。用于原创文本改编；不直接生成图片或视频。

将叙事文本转化为影视开发阶段可用的多用途读本。先做快速诊断判断开发价值，再按受众生成完整开发读本、演员读本、编剧改编读本、导演视听读本、制片摘要或 AI 剧本化提示包。适用于长会话记录、短篇小说、大纲、剧本初稿、互动叙事等输入。

多风格AI短剧的导演前期与关键资产节点。仅在剧本定稿后，用于完成十项剧本解析、情绪曲线、项目确认的视觉/色彩/声音基准，以及角色和关键道具母板；三维国风漫剧是可选类型，不覆盖真人实拍或其他已确认风格；输出创作基准和 asset_manifest，不做场景成图、分镜、视频提示词、生成或剪辑。

Use when creating, adapting, diagnosing, refining, or directly generating cinematic 中式天宫 imagery: credible Chinese timber palaces, cloud seas as spatial terrain, monumental scale, tiny human anchors, restrained vermilion–blue-gray–gold palettes, and physically plausible mythic light. Trigger for 中式天宫、东方仙宫、云海宫阙、天庭仙境、国风神话史诗、云上宫殿、仙侠电影概念图, coherent image series, or images that look generic, kitschy, structurally incoherent, or insufficiently Chinese, including built-in image generation and explicitly requested GPTX-compatible third-party API generation.

AI 短剧五冠军的提示质控与视频交付节点。已有 shotlist、资产和连续性事实，需要把表演微动作、镜头运动、空间物理与声音转为可靠图像/视频提示词，或诊断提示词和连续性问题时使用；目标为 Seedance 2.5 时，负责精确时间轴、多模态参考与最终交付编译；输出最终提示词与 continuity_ledger，不调用生成渠道。

小而美天宫漫剧：把完整小说蒸馏成上、中、下三集、每集约3.5分钟的原创可拍漫剧包。适用于“小说蒸馏漫剧”“三集天宫漫剧”“上中下三集”“小而美漫剧”“把小说压成精华完整剧情”等请求。 借鉴女娲式六维并行蒸馏，吸收叙事诊断、原作事实账本、可拍动作链、资产交接和国风视觉证据方法；不做连续长剧，不逐章改编，不直接提交生图/生视频。

Novel 转 Tiangong Manju：This skill should be used when adapting a Chinese fantasy, Honghuang, xianxia, or other novel into a high-quality Chinese celestial-palace 3D animated micro-drama. It preserves the source novel's characters, world, cultural names, main plot, and core relationships while improving screenability, short-drama pacing, episode hooks, visual storytelling, asset continuity, and Seedance 2.5 image-to-video delivery.

Relic Collector Ip 系统：Build and maintain the Relic Collector original-IP system for award-oriented AI narrative film work. Use when the user wants the Relic Collector world, protagonist, rules, emotional logic, key scenes, asset planning, or a reusable project memory layer around that IP.

AI 短剧五冠军的编剧节点。用户要从创意、小说、大纲或已有素材得到可拍剧本、人物弧线、分集结构、场景和对白，或需要修订故事结构与节奏时使用；开放式原创短剧启动时，基于知识卡筛选框先产出三套完整候选方案供用户选择，不先索要一句核心问题；输出剧本、故事圣经和 project_state，不做资产、分镜或视频提示词。

天宫漫剧项目专属冠军。用户要做中式天宫 3D 国漫漫剧（抖音 9:16）的整部制作时使用：从创意或小说素材出发，经知识筛选、剧本、3D 国漫角色与天宫场景资产、分镜走位、Seedance 2.5 图生视频提示词，到候选片段验收的完整链路。生图默认 Midjourney（用户 2026-08-24 拍板），OpenLux gpt-image-2-c 为备用兜底。仅服务于天宫漫剧项目，不改写其他项目的五冠军路由。

Seedance 2.5 旧入口兼容说明。精确计时、30 秒、延长、多人和多模态视频提示词编译已并入官方冠军 hell-grind；此入口只将旧调用转交给 hell-grind，不独立输出提示词。

Seedance2 A Contest Orchestrator：Orchestrate an all-AI Seedance2 plus Image2 contest-film workflow for award-oriented narrative entries. Use when the user wants an all-AI command-track film, a priority-A contest entry, or a routed workflow across research, asset building, shot execution, assembly, and submission prep.

AI 短剧五冠军的拆镜节点。剧本和关键资产已确认、用户需要把场次变成可生产的镜头表、走位、机位、镜头节奏和视频提示计划时使用；输出 shotlist 与 continuity_ledger 草案，不写剧本、不重新设计资产、不提交生成任务。

审计 Master Thread：Audit and advance multi-agent Chinese short-drama production from finalized script through accepted video clips and edited delivery. Use when resuming a project, reviewing a handoff, checking production readiness, locating the earliest failure, reconciling project state, or turning repeated failures into validated workflow improvements.

AI 短剧五冠军的知识入口。用户提供小说、视觉参考、导演资料、提示词包或控制台资料，需要在写剧本、做资产、拆镜或质控前提取可验证规则时使用；原创题材启动时先审计话题度、网感和爆款底层逻辑；只输出来源明确的 knowledge_brief，不创作剧本、资产或提示词。

Use when 用户需要把仙侠天宫、云海神殿、浮空宫殿、飞瀑、廊桥或星河古树等场景图转换为 Seedance 图生视频提示词，尤其是人物容易静止、云层缺少漂移、银河不流动、星光不闪烁、树叶不摆动，或需要在保持人物身份、建筑尺度和空间连续性的同时设计电影化运镜与自然动态。

为仙侠、天宫、云海神殿、浮空宫殿等宏大远景设计可复用的图片提示词与可选视频运镜。用于用户上传参考图或用文字描述后，要求拆解同款风格、构图、空间层次、镜头语言，生成可修改属性的提示词结构、完整案例、负面提示词、构图锁定句，或把静态构图扩展为缓推、侧移、跟拍、揭示等仙侠运镜时；同时区分 ChatGPT-image2 的结构化中文提示词与 Midjourney/MJ 的紧凑电影设定提示词，覆盖柱廊、天阶、瀑布、水庭、悬台、弧形长廊、月夜轴线栈道、圆月门框景、云海廊桥、神兽托城、洪荒能量瀑布、蓬莱海阁、沙海天宫等场景。

为原创小说、剧本和短剧项目建立并执行一键式AI视频资产包。用于将已确认的角色、场景、道具、首帧需求编译为资产清单，按依赖调用指定Image2文生图或图生图渠道，下载验收并登记可复用资产；不用于有原片证据的转绘项目。

用 GPT Image 2 生成高质量故事板：按人物图、关键信息图、首帧图、故事板的四阶段顺序锁一致性与减少 AI 感，并提供口播故事板模板与生视频衔接。

Img2threejs：Turn an object or character reference image into a quality-gated, animation-ready procedural Three.js model built in code. Use for image-to-3D reconstruction, detail-accurate object rebuilds, stylized/likeness-maximized human characters, sculpt specs, and staged code generation.

This skill should be used when the user asks to put local reference images and Seedance 2.5 prompts into the Windows desktop app 灵豆同学·Dola 创作版. It prepares the app in AI创作 → 视频 → Seedance 2.5 state, asks for usable account numbers when they are not provided, maps per-segment R01–R15 images, injects the prompt, and stops before sending or generating.

Mx Shortdrama 01 Frame 提取：Step 01 for Chinese short-drama redraw to Mexico: extract and enhance a production-complete evidence package from one episode video, including native-resolution reference frames without a fixed frame-count cap, 60-second chunk manifests for long episodes, shot-level start/mid/end frame supplements, TransNetV2 shot segmentation, Baidu OCR API smart hard-subtitle/text QA, audio/ASR handoff, and evidence-pack alignment before source timeline reconstruction and shot prompt production. Use when the user provides a domestic short-drama video and asks for frame extraction, key frames, shot frames, subtitle frames, source evidence frames, OCR, shot detection, or evidence before timeline/prompt reconstruction.

Mx Shortdrama 02 Source Timeline：Step 02 for Chinese short-drama redraw to Mexico: rebuild the original episode source reference timeline from a validated Step 01 evidence package, including frame matching, PaddleOCR API smart selective OCR, ASR, TransNetV2/OpenCV shot evidence, and native-resolution visual review. Use for source timeline, episode pull analysis, original dialogue recovery, source plot reconstruction, continuity evidence after frame extraction, smart OCR QA, and QA reference before localized shot prompt production.

为短剧转绘建立证据驱动的人物资产图提示词与生产合同。用于人物定妆母图、角色三视图、角色设定卡、服装状态卡、妆造、发型、配饰、表情、人物一致性、角色图 QA；作为 mx-shortdrama-04-asset-prompts 的人物资产子 Skill，不处理场景、道具或生视频提示词。

Niannian 画布 Asset Import：Import one approved local image, video, or audio asset into the current Niannian AI production canvas without browser file chooser or clipboard automation. Use when an approved asset must become a durable project asset and a node in a default canvas group, especially after browser upload, cross-drive, or clipboard transport failures.

Seedance 2.5 国风 T2V / I2V 提示词模板。为天宫漫剧等国风题材编写可直接投喂的提示词：动作链、镜头逻辑、情绪板参考分离，正式正文不写秒数与时间戳。

Seedance 2.5 童装带货口播与镜头模板。按商品事实卡选择场景沉浸型、换装变装型或知识科普型主风格，集中一条连续妈妈口播，动作镜头只写画面执行，可交付给即梦 Seedance 2.5。

将用户输入的剧本/创意，按动作先后与镜头逻辑编译为即梦 Seedance 2.5 专用视频提示词（多模态 + 纯文字双版本）。支持单段、延长、超长、多模态参考、白模控制、迁移创意、去字幕去BGM纯净素材和多语种对白；内部可做时长校验，但最终提示词正文不输出秒数、时间戳或帧号。当用户需要 Seedance 2.5 视频提示词、剧本转2.5、2.5分镜脚本、延长提示词或超长视频脚本时使用。触发词：Seedance2.5提示词、剧本转Seedance2.5、2.5分镜脚本、视频延长提示词、超长视频脚本。

Seedance2 Narrative Shot 工作流：Execute Seedance2 narrative shots from approved Image2 assets. Use when the user wants shot-by-shot Seedance2 prompts for AI short films, contest videos, all-AI narrative pieces, or Mx-Shell-style workflows with first-frame locks, timing beats, audio rules, and reroll strategy.

可靠执行念念画布生图任务，覆盖节点提交、任务轮询、素材登记、结果回写、预览验收和失败修复。Use when a canvas image node生成后显示加载失败、节点消失、结果未回写，或需要验证生图链路是否真正完成。

Primary router for domestic Chinese short-drama redraw to Mexico through the redraw pipeline: Step 01 frame/audio extraction, Step 02 source reference timeline, Step 04 one-pass Mexico localization plus final asset/video prompt Word package, and Step 05 asset image execution. Step 03 is now an internal Step04 localization binding layer by default, exported only on explicit request. Use for 短剧转绘、墨西哥转绘、西语本土化、分集推进、skill 路由、资产提示词、人物图、场景图、道具图、生视频提示词、Word提示词包、redraw/remake, and batch redraw automation. The optional frame-anchor add-on is explicit-only.

Convert Chinese short-drama episode videos into localized bilingual remake/redraw production packages for AI video production. Use when the user provides a short-drama MP4 or episode folder and asks for 拉片, 视频转剧本, 转绘剧本, 外语版短剧, 本土化改编, AI转绘重制, 配音字幕脚本, shot-level first-frame/key-frame/last-frame prompts, all-purpose reference video prompts, or production-ready episode packages that preserve the original narrative while replacing Chinese visual/cultural elements.

Send Chinese-named short-drama redraw deliverables to WeChat through either Hermes Agent in WSL or Windows WeChat automation. Use when the user asks to send, forward, or batch-deliver 转绘产物, Step2, Step4, Word/docx, zip files, asset prompt packages, or short-drama episode deliverables to 微信、文件传输助手、微信群/群聊, especially Mexico redraw outputs and sleep-time automated delivery.

Agnes生视频 — 用 Agnes Video 2.5 / 2.5 Flash API 直接生成 AI 视频，支持文生视频、首尾帧控制、参考图生成，本地图片自动上传，任务轮询，成片自动下载。 触发词：「agnes生视频」「agnes video」「agnes出视频」「用agnes抽卡」「agnes flash」「用 agnes 生成视频」「agnes API 出片」。 关键事实：Flash 版限时免费（720P/4-12秒）；标准版 $0.025~0.055/秒；官方无上传端点，本地图片默认自动传免 Token 的 catbox 获得公网 URL；不必走 Pavo 画布，API 可直连。

Astorie Seedance2 渠道：Operate an already authenticated AStorie project canvas for the NianNian all-purpose Smini channel. Use for locked Seedance 2.0 Mini text-to-video or one-image-to-video tasks, live price readback, one Generate submission, media download, and delivery evidence.

Beecode Image2：Generate raster images through the BeeCode OpenAI-compatible relay at beecode.cc using Image2 / GPT-Image-2. Use when the user asks to create images with BeeCode, a Zhongzhuanzhan/API relay, image2, GPT-Image-2, OpenAI-compatible image generation, or this skill by name.

GPT 图片 2 Style 库：Choose GPT-Image2 / gpt-image-2 visual styles and industrial prompt templates from the awesome-gpt-image-2 style library. Use when an agent needs to create, rewrite, classify, or improve image-generation prompts with repository-backed templates, categories, style tags, scene tags, pitfalls, and example cases.

Ikun Image2：Generate or edit raster images through the user's Ikun / Monkey Tools NewAPI OpenAI-compatible image channel. Use when Codex should create Image2 / GPT-Image-2 images, test the api.monkey-tools.cn channel, list available relay models, or run text-to-image and image-to-image jobs through this global skill.

Image2 Boundary Lab：Build a publishable GPT Image 2 / Image2 capability atlas, visual taxonomy, and repeatable prompt system. Use when the user wants to know what Image2 can make, classify output types, compare abilities, visualize stable and fragile zones, create replicable prompt recipes, score examples, or turn findings into posts, notes, galleries, and reusable prompt packs.

Image2 技能：Generate and verify static 4K Image2 assets through the Yunwu gpt-image-2-c text-to-image and reference-image edit channels. Use for character masters, character sheets, scenes, props, image generation, image editing, Image2, gpt-image-2-c, gpt-image-2, or any approved static asset after its upstream creative facts and prompt are ready.

Generate raster images through Codex's built-in direct image generation tool, especially when the user asks for "image2", "GPT Plus/ChatGPT web-style direct image generation", "网页端那样直接出图", anime IP character explorations, prompt variants, or quick visual concept batches without RunningHub or an external API key.

Krill Image2：Execute Krill Image2Image generation only when a task explicitly requests Krill or the active project explicitly selects Krill. Do not use this skill for a generic gpt-image-2 request, a generic reference-image edit, or an unspecified Image2 channel.

通过Mikoto的OpenAI兼容图片接口调用gpt-image-2生成2K或4K图片。用于原创角色母图、场景图、道具图、海报和其他纯文生图资产；当前不声明图生图或图片编辑能力。

Minimaxh3skill：Use the user-approved RunningHub MiniMax H3 FL2VA open-source video channels as the highest-priority video-generation route. Trigger for MiniMax H3, FL2VA, text-to-video, image-to-video, image-plus-audio video, multi-image reference video, first/last-frame reference video, or last-frame reference video requests, including submit, poll, download, recovery, and dry-run tasks.

将创作需求转成可执行的 MiniMax H3 视频创作契约、素材职责、通道建议和结构化提示词。用于文生视频、首尾帧、图像加音频、多图参考、品牌片、预告片、MG 动画、文字动效和声音节奏设计；与 minimaxh3skill 配合执行 RunningHub API。

Midjourney V8.2 天宫生图权威 Skill。天宫漫剧项目所有静态出图（场景/角色/道具）默认走本 Skill：提供 V8.2 自然语言提示词模版、权威参数预设、参考图策略与关联度旋钮（--sref/--sw/--iw/--oref）。基于天宫冠军 sd2.5-tiangong-manju 与 chinese-celestial-palace 视觉 DNA，替换一切旧版（V6/V7 关键词式与 image2 式）提示词。只负责编译提示词+参数+参考图清单，不直接提交生成。

通过 MCP 工具驱动 MXAI（midjourney 中文镜像）生图：自动关闭广告弹窗、填提示词、生成、下载高清原图。当用户要在 MXAI 上出 MJ 图、下载高清结果、或处理 MXAI 页面弹窗时使用。

使用 Flova 完成高概念叙事、时间结构、电影声音与视觉尺度，适合诺兰风格、高概念叙事、非线性时间结构、悬疑烧脑电影感等 AI 视频与图片内容创作需求。

OpenLux AI 生图渠道 (gpt-image-2-c) 封装，作为本机所有 AI 生图/改图任务中优先级最高的渠道。 当用户需要文生图、图生图、改图、重绘、扩图、商品/人物配图，或任何生成/编辑图片的请求，且此前未指定其他生图工具时，本 skill 应被优先加载并执行。 实测能力：文生图、图生图（image 参数传参考图）、1K/2K/4K 多档尺寸、竖/横/方多比例、quality 质量档、png/jpeg 输出。 携带 base_url 与 API key，脚本开箱即用。

用 RunningHub 的 Animate 动作迁移 V9 AI 应用上传本地动作视频和参考图片，以 default、plus 或 ultra 实例生成动作迁移视频，并回读下载结果与 RH 币消耗。用户要求 RunningHub 动作迁移、儿童服装动态展示、参考视频驱动、网页上传或将此应用自动化时使用。

Runninghub Image2 Text：Generate images from text prompts with the RunningHub Image G 2.0 / GPT-Image-2 low-price text-to-image channel only. Use when Codex should submit, poll, and download RunningHub Image2 low-price text-to-image jobs using RUNNINGHUB_API_KEY for ad stills, IP character concepts, e-commerce visuals, posters, or prompt iteration.

Runninghub Image2 图片：Generate or edit images from one or more reference image URLs with the RunningHub Image G 2.0 / GPT-Image-2 low-price image-to-image channel only. Use when Codex should submit, poll, and download RunningHub Image2 low-price image-to-image jobs using RUNNINGHUB_API_KEY for restyling, product image edits, IP character refinement, or reference-guided ad visuals.

RunningHub 冠军入口：将 RunningHub/ComfyUI 工作流或 AI 应用定制、发布或调用为稳定 API，并统一专项 Skill 的素材上传、消费级 API Key、RH 币结算和网站回传路径。用于任何 RunningHub 请求，包括工作流、AI 应用、动作迁移、MiniMax H3、Image2、画布素材、消费级 Key 或念念智选服务端接入。

Generate images with a three-channel fallback: use RunningHub first, then the OpenAI-compatible Canvas channel, then Ikun/Monkey Tools when needed. Use for Codex image-generation threads that need stable routing across RunningHub, 画布渠道, and Ikun.

Runninghub 画布 媒体 Upload：Bind local image, video, and audio files to the correct RunningHub ComfyUI canvas nodes and verify their real use. Use for RunningHub workflow media uploads, multi-image reference workflows, image/video/audio node mapping, canvas-side upload recovery, and pre-run asset checks.

Sd2 视频 生成：Primary router and workflow for AI video generation with Seedance2 / SD2 on StoReel Canvas. Use whenever the user asks to generate, continue, extend, remake, batch, or finish AI videos; create image-to-video or reference-video clips; write SD2 video prompts; bind first frames, character references, product references, or previous video references; poll tasks, write results back to canvas, save local deliverables, or switch StoReel accounts when quota/points are exhausted.

Storeel Seedance2 画布：Low-level StoReel Canvas operations for canvas.storeel.vip: login, browser navigation, opening or creating free-canvas projects, inspecting project canvas data, uploading assets, and preserving project structure. For end-to-end AI video generation, SD2/Seedance2 prompting, task submission, polling, writeback, local archiving, or account rotation, use sd2-video-generation first as the router.

Tensor 画布 直连 节点：Automate Tensor.art canvas video workflows by connecting to a logged-in Chrome session through CDP, uploading or binding verified local assets, creating or locating video nodes, writing model/capability/ratio/resolution/duration directly into the Tensor business canvas store, verifying persistence, submitting paid generation only when approved, and recording account/output provenance. Use when working with Tensor.art canvas, canvas.tensor.art, Tensor login/account switching, video node setup, Happy Horse 1.1, Seedance 2.0 Mini, or when UI clicking is too slow or desyncs from saved canvas state.

Tmlab 视频 渠道：Operate TMLab AI video channel through canvas agent.tmlab.top and API preflight for real video generation. Use when user mentions TMLab, tmlab, agent.tmlab.top, api.tmlab.top, AI video main channel, canvas video generation, API video generation, upload/generate/poll/download, account quota, or provider evidence/ledger for TMLab.

统筹一项原片短剧转绘任务，从用户提交素材到可播放视频持续推进。用于新线程启动、跨步骤恢复、并发调度、渠道执行或现有转绘 Skill 路由出现缺口时；加载 mx-shortdrama-00-router 作为唯一专业路由，并将真正需要用户判断或需要补全路由的卡点压缩为最多三道选择题。

为原片转绘任务选择、生成和迭代可直接注入生图与生视频提示词的视觉风格配置。仅在用户明确要求风格，或当前任务配置标记需要风格时使用；支持先启用试运行基调，再依据实际产物和用户反馈迭代。

This skill should be used when creating or updating a Chinese celestial-palace moodboard from multiple reference images or video frames, especially when the task involves人物、远景巨构、中近景日常、素材道具分类, extracting visual style/cinematography/color-lighting/character/material responsibilities, or compiling reusable prompts for celestial-palace image and video generation. It coordinates the moodboard layer with the authoritative celestial-megastructure DNA and routes the result to MJ image prompts or Seedance video prompts without copying protected characters, dialogue, logos, watermarks, or exact frames.

故事板 Director：Build or run a storyboard-director workflow that reads a script, reviews the beat structure, plans professional storyboard shots, validates action timing, expression, and object state, critiques and scores the plan, and outputs detailed storyboard prompts or screenplay-style shot lists. Use when the user wants script-to-storyboard prompts, shot-by-shot visual planning, storyboard review, or a professional breakdown before video generation.

Mx Shortdrama Production Iteration：Record, evaluate, and verify evidence-based quality and throughput improvements for original-footage short-drama redraw jobs. Use after provider outputs, user visual feedback, channel failures, timing measurements, or task completion; also use when updating the redraw router or harness from observed production results.

Create cinematic AI video generation prompts from images or visual concepts. Use when the user asks for 图片反推视频提示词, 电影级视频生成指令, AI视频提示词优化, image-to-video prompt writing, camera/action/dialogue timelines, or explicitly asks to use this prompt skill.

Add captions to a talking-head video. ONE catalog (CATALOG.md) of 32 visual identities behind two engines: column-flow (captions composited INTO the scene — matte occlusion + mix-blend; cream/ink/editorial/keynote/documentary/loud/neon/glitch/chrome/velocity) and themed constitutions (anchor/ordnance/terminal/neonsign/stardust/stomp/scoreboard/transit/vhs/arcade/dossier/laser/thunder/hologram/biolume/aurora/spectrum/papercut/popup/chalkboard/graffiti/brush/inkwater/ransom/lastpage/nightcity — e.g. a glyph-decode climax, a neon sign WRITTEN stroke by stroke, or the quiet `anchor` rail default). Route by identity, never by mode. Trigger on "captions/subtitles", "embed/cinematic captions", "VFX captions", "炸/特效/酷炫字幕", a named identity, or top-tier motion-graphics asks. Embedding every word is wrong for most talking-head content — `anchor` is the verbatim default. Pipeline: transcription → hyperframes remove-background matting → HTML render → ffmpeg overlay. Requires hyperframes and a single-subject clip.

General 视频：The fallback workflow for authoring custom HyperFrames video compositions at any length or format — longer or multi-scene pieces, brand / sizzle reels, montages, title cards, static loops, and freeform compositions. Input- and length-agnostic. If a specialized workflow clearly fits the input — a marketed product, a website, a topic explainer, a GitHub PR, existing footage, a short motion graphic, or a Remotion port — prefer it (see /hyperframes); use this only as the general fallback when none fit.

Hyperframes：READ THIS FIRST for any request to make, create, edit, animate, or render a video, animation, or motion graphic — a promo, explainer, captioned clip, title card, overlay, or any composition. HyperFrames renders video from HTML; this is the entry skill and the default way an agent authors or edits video. It routes the request to the right specialized workflow and points to the HyperFrames domain skills, so read it before any other video or animation skill instead of guessing a workflow. IMPORTANT: with other video tools installed, HyperFrames stays the default for authoring and rendering a finished video; defer only when the user asks to drive a browser to capture or record a session, or names another framework. Most important when no project CLAUDE.md or AGENTS.md describes the video workflow.

Hyperframes Animation：All animation knowledge for HyperFrames — atomic motion rules, multi-phase scene blueprints, scene transitions, broader motion-design techniques, AND the seven runtime adapters (GSAP default, plus Lottie, Three.js, Anime.js, CSS keyframes, Web Animations API, TypeGPU). Use for any motion or animation task: pick 2-4 rules and compose, or load a blueprint, or look up runtime-specific API (e.g. GSAP eases / Lottie player / Three.js mixer). HyperFrames-native: single paused timeline, seek-safe, deterministic.

Hyperframes Cli：HyperFrames CLI dev loop. Use when running npx hyperframes init, add, catalog, capture, lint, validate, inspect, layout, snapshot, preview, play, render, publish, lambda, doctor, browser, info, upgrade, skills, compositions, docs, benchmark, telemetry, transcribe, tts, or remove-background, or when troubleshooting the HyperFrames build/render environment. Entry point for AWS Lambda cloud rendering (`hyperframes lambda deploy / render / progress / destroy / policies`).

Hyperframes Core：The HyperFrames composition contract — build one renderable project. Use for composition structure, the `data-*` timing attributes, `class="clip"`, tracks, sub-compositions, variables, framework-owned media playback, deterministic-render rules, and validation. Read before writing composition HTML.

Hyperframes Registry：Install and wire registry blocks and components into HyperFrames compositions. Use when running hyperframes add, installing a block or component, wiring an installed item into index.html, or working with hyperframes.json. Covers the add command, install locations, block sub-composition wiring, component snippet merging, registry discovery, and authoring a new block or component to contribute upstream (idea → scaffold → validate → PR).

Hyperframes 创意：Non-animation creative direction for HyperFrames videos. Use for design spec (frame.md / design.md) handling, palettes, typography, narration, beat planning, audio-reactive visuals, composition patterns, and brand / style decisions. For atomic motion patterns and scene blueprints, use `hyperframes-animation`.

Hyperframes 媒体：Audio and media assets for HyperFrames compositions, produced by one shared audio engine (`scripts/audio.mjs`) — multi-provider TTS (HeyGen / ElevenLabs / Kokoro local), background music + sound effects (HeyGen audio-library retrieval by default, with local Lyria / MusicGen BGM generation and a bundled SFX library as the no-credential fallback), Whisper transcription, background removal, and caption authoring. Use for voiceover / TTS, BGM, SFX / sound effects, transcription, captions / subtitles / lyrics / karaoke / per-word styling, voice + provider selection, and music-mood prompting.

Pexoai 智能体：Use this skill when the user wants to produce a short video (5–120 seconds). Supports any video type: product ads, TikTok/Instagram/YouTube content, brand videos, explainers, social clips. USE FOR: video production, AI video, make a video, product video, brand video, promotional clip, explainer video, short video.

Pr 转 视频：turn a GitHub pull request (a PR URL like github.com/<owner>/<repo>/pull/<N>, an <owner>/<repo>#<N> ref, or 'this PR' in a checked-out repo) into a code-change explainer video, up to ~3 min (sweet spot 30-90s) — changelog, feature reveal, fix, or refactor walkthrough, rendered from the diff / commits / files. The input is a CODE CHANGE read via the gh CLI; there is no website capture. Use this skill for a GitHub PR. Do not use it for a product launch/promo (use /product-launch-video), a tour of a real website (use /website-to-video), a topic explainer with no PR (use /faceless-explainer), captions on existing footage (use /embedded-captions), or a short unnarrated motion graphic (use /motion-graphics). If the intent is unclear, route through /hyperframes first.

Remotion 转 Hyperframes：Port an existing Remotion (React) composition to HyperFrames HTML. Use ONLY when the user explicitly asks to port/convert/migrate/translate a Remotion source. Do NOT use: (a) authoring a new HyperFrames composition; (b) Remotion mentioned in passing; (c) Remotion code shared as reference only; (d) "same video as my Remotion one" without explicit migrate request — treat as fresh build. Doubt → `/general-video`. One-way, Remotion-only: no reverse export (HyperFrames→Remotion or any framework), no non-Remotion source (After Effects, Framer Motion, plain React/CSS) → out of scope, re-create via `/general-video`. Flags unsupported patterns (useState, useEffect, async calculateMetadata, third-party React libs, `@remotion/lambda`) and recommends runtime interop over lossy translation. Unsure whether to port vs. build fresh, or only a passing Remotion mention? → /hyperframes.

Create TikTok/CapCut-style burned-in English video captions with short phrase-level timing, selected keyword highlights, ASS/SRT outputs, and ffmpeg verification. Use when Codex needs to add or revise subtitles for vertical social videos, commerce clips, TikTok/Reels/Shorts edits, or when the user asks for "TikTok same style captions", "重点词高亮字幕", "不要逐词高亮", or non-overlapping highlighted English captions.

产品发布视频：turn a product or marketing URL, pasted script, or brief into a product launch video, including SaaS promos, feature reveals, app launches, company promos, and product marketing videos. Use this skill when the user wants to market, launch, promote, or reveal a product. Do not use it for general non-launch website tours, non-product topic explainers, GitHub pull requests, captioning existing footage, or short unnarrated motion graphics. If the intent is unclear, route through /hyperframes first. This is the new shot-sequence architecture: every visual frame is authored as a time-coded shot sequence picked from a menu of golden blueprints, so frames develop over their full duration instead of freezing after entrance.

动效图形：Use when the user wants a short, design-led motion graphic where motion is the message: kinetic typography, stat or number count-up, chart/data-viz hit, logo sting, brand lockup, lower-third, callout, social overlay, animated headline/tweet/news item, motion poster, or quick captured-page highlight. Usually under 10s and up to ~30s, with no narration arc, voice-over, or live-action subject. Can render to MP4 or transparent overlay. Not for longer, multi-scene, narrated, or brand-reel pieces (use general-video), narrated website videos (website-to-video), topic explainers (faceless-explainer), product promos (product-launch-video), PR videos (pr-to-video), or captions on existing footage (embedded-captions). When unsure whether it's a quick motion-first piece or a longer / narrated treatment, see /hyperframes.

口播视频重剪：Package an existing talking-head / interview / podcast video by layering timed, designed GRAPHIC OVERLAY cards onto the playing video — titles, lower-thirds, data callouts, quotes, side panels, picture-in-picture — synced to the transcript. The source video plays in full; the agent designs and writes each card's HTML in conversation, then renders to MP4 via hyperframes. Use when the user asks for graphic overlays, on-screen graphics / lower-thirds / data callouts / kinetic titles on a video, "package / dress up my video", "add overlay cards / graphic cards", or AI-composed graphic packaging of an existing video. NOT for plain subtitles (→ embedded-captions) or building a video from scratch (→ the creation workflows); when unsure overlays-vs-captions, see /hyperframes.

幻灯片：Author a HyperFrames slideshow composition — a presentation, pitch deck, or interactive deck with discrete slides, fragment reveals, branching sequences, and hotspot navigation. Use as an intent check when the user asks for a presentation, pitch deck, slide deck, interactive deck, or page-to-deck conversion that might be a slideshow; if the user did not explicitly ask for a slideshow / slide show, confirm before authoring.

无脸解说视频：turn arbitrary text — an article, notes, a topic, a brief — into a faceless explainer video, up to ~3 min (sweet spot 30-90s), where every visual is invented (typography, abstract graphics, diagrams, data-viz) rather than captured. There is no URL, no website capture, and no real assets. Use this skill for topic explainers, concept breakdowns, how-tos, listicles, and narrative explainers. Do not use it for a product launch/promo (use /product-launch-video), a tour of a real website (use /website-to-video), a GitHub PR (use /pr-to-video), captions on existing footage (use /embedded-captions), or a short unnarrated motion graphic (use /motion-graphics). If the intent is unclear, route through /hyperframes first.

网站转视频：Capture a general website/URL and turn it into a HyperFrames video (site tour, showcase, or social clip from the site's own visuals). Uses headless Chrome screenshots + brand assets. Use when intent is general — portfolio/blog/landing-page showcase or social clip from the site. NOT for: product/SaaS launch or promo (→ /product-launch-video, even from a URL); topic explainer with no site (→ /faceless-explainer); GitHub PR (→ /pr-to-video); adding captions to existing video (→ /embedded-captions); short unnarrated page-highlight motion graphic (→ /motion-graphics). Unclear launch-vs-general-site? Ask one question or start at /hyperframes.

音乐 转 视频：Use when the user has a music track (an audio file, or a video to pull audio from) and wants a beat-synced HyperFrames video, calm to hard-hitting. The music drives everything: one analyzer reads it once, the orchestrator lays out the frames and fills a per-frame plan, and one sub-agent builds each frame. Typography and templates are the floor — a complete video needs zero assets — but any images or videos the user supplies are cut into the frames on the same beat grid (beat-cut / ken-burns). The genre (lyric video, slideshow, kinetic promo) falls out of the per-frame choices; the pipeline never branches on it.

Douyin Caption Cover：Create Douyin publishing packages including titles, captions, hooks, hashtags, cover text, cover frame direction, and variant options. Use when the user needs copywriting, description polishing, topic tags, cover concepts, or final post packaging for a Douyin video.

Plan and operate a Douyin fruit-selling content account, including persona positioning, content pillars, video formats, posting cadence, conversion paths, and iteration metrics. Use when the user asks how to grow a fruit commerce account, mentions 水果带货, 果园产地, 果农, 美女主播吃水果跳舞, 朴素老太太推销, fruit livestream-style short videos, or wants a repeatable Douyin fruit content strategy.

Douyin 工作流 Orchestrator：Coordinate an end-to-end Douyin content workflow from choosing a video or topic, planning production, preparing captions/covers, and operating the Creator Center upload page. Use when the user asks for a full Douyin publishing workflow, wants Codex to manage multiple Douyin skills, or says to select/make/upload/publish a Douyin video.

Use when selecting, reselecting, scoring, or explaining topics for the user's long-term AI information-gap IP and digital-human videos, especially when evaluating GitHub/open-source AI projects, AI tools,热点,候选选题卡,视频方向, or whether a topic is worth producing before Window B.

Kuaishou Publish Packager：Package a finished video for Kuaishou publishing. Use when Codex needs to create or improve titles, captions, hashtags, cover text, cover-frame selection, first comments, publish timing, category/account fit, and final pre-upload copy for Kuaishou short videos.

Kuaishou 内容 Pipeline：End-to-end Kuaishou short-video operations workflow. Use when Codex needs to coordinate the full domestic Kuaishou process from topic/material selection, video planning or production, cover/title/caption packaging, browser/proxy preparation, upload, publish confirmation, and post-publish iteration.

Execute Kuaishou creator-platform publishing tasks. Use when Codex needs to open or troubleshoot Kuaishou/快手 creator pages, configure Clash proxy routing for Kuaishou domains, log in with user assistance, upload video files, set title/caption/cover/tags, verify publish settings, and stop for explicit confirmation before final publish.

Create, rewrite, replicate, or quality-control realistic natural commerce short-video scripts, shot lists, and AI image/video prompts for products, especially food, fruit, agricultural goods, local specialties, handmade goods, and live-commerce assets. Use when the user asks to copy/复刻/reference a high-performing commerce video or image, produce product short videos, Douyin/Kuaishou/Xiaohongshu commerce visuals, Image2/Seedance/Runway/Jimeng prompts, product first frames, branded package scenes, or complains that generated images look fake, staged, plastic, generic, over-designed, lack hook, lack proof, lack viral selling point, or lack natural light.

Generate fruit commerce videos with the user's RunningHub custom workflows, especially Wan2.2 Animate action transfer workflow 2056752570487623681 and LTX2.3 ecommerce digital human workflow 2057025848015937537. Use when the user mentions RunningHub, 动作迁移, 数字人口播, LTX2.3高清超自然电商数字人, Wan2.2 Animate动作迁移V8, fruit sales videos, 美女主播吃水果跳舞, 果园产地, 果农, or wants Codex to run and summarize these video workflows into reusable skills.

Create Seedance2/即梦-ready commerce IP videos for short-form selling, including persona positioning, 15-second hard/soft ad scripts, first-frame planning, reference-image upload order, image-to-video prompts, camera/lighting constraints, and iteration guidance. Use when the user mentions Seedance2, 即梦, 首帧图, 中间帧, 尾帧, 参考图, 带货视频, 女装带货, 水果带货, 快手/抖音/小红书种草, AI主播IP, or asks to turn product/model images into commerce video prompts.

Self 媒体 技能 Route：Use this route when turning a project, article, video, screenshot, product, or technical result into publishable Chinese social-media content for Xiaohongshu or Douyin. It coordinates research, content ingestion, visual planning, script planning, compliance checks, draft publishing, and performance review through the locally deployed projects in this route.

Xianyu AI Demand Radar：Discover, validate, score, and productize AI-related Goofish/Xianyu demand. Use when the user wants to mine AI information-gap products, find hot AI services, turn accumulated AI workflows/assets into sellable listings, rank opportunities, create listing angles from trends, or build a repeatable demand research workflow for Goofish/Xianyu.

Xianyu Product 发布器：End-to-end Goofish/Xianyu listing workflow. Use when the user gives a Xianyu/Goofish competitor product link and wants Codex to read the listing, reconstruct stronger Chinese sales copy, generate original product images, prepare a publish pack, automate the Goofish publish page with Playwright/CDP, upload images, and publish only after explicit user confirmation.

小红书相关操作，覆盖账号定位、选题研究、内容生产、发布执行与复盘修复的小红书全链路运营技能。凡是小红书的浏览/搜索/发布/评论任务，默认必须使用 OpenClaw 内置浏览器流程并指定 profile=\"openclaw\"；除非用户明确要求，否则不要使用系统 open 或外部浏览器。

自媒体主控·运营层。老大创作内容+生产视频完成后，我处理：发布包、合规、数据、日历。T4只负责发布执行。触发词：自媒体、四账号、念念漫剧社、念念带货屋、念念AI社、念念制片厂、发抖音、选题、排期、运营、数据。

Create complete Xiaohongshu image-and-text posts from a topic, article, document, or rough idea, including research, fact checking, post copy, pagination, storyboards, a 3:4 cover, 9:16 content pages, image generation or fully local SVG-to-PNG fallback, and final quality review. Use for 小红书图文、小红书封面、知识卡片、漫画科普、文章转图片、AI/科技科普图组、逐页生图提示词, or revisions to an existing Xiaohongshu series. Prefer an available image-generation model; use the bundled local renderer only when no image-generation tool is available. Apply an SVG text patch only after the user explicitly reports incorrect or unreadable text in an image.

小红书原创图片发布：Publish a verified original-image Xiaohongshu post through AitoEarn in Edge. Use when a user asks to prepare, check, or publish an existing Xiaohongshu image post from locally verified source images, especially when images must remain unchanged and an explicit final publishing confirmation is required.

小红书流量审美护栏：Guardrail for Xiaohongshu traffic-first image notes. Use when creating, reviewing, or revising Xiaohongshu covers, image cards, note copy, or image2 prompts where click-through, platform-native feel, and anti-AI aesthetic quality matter. Trigger this before publishing any Xiaohongshu image note, especially when the draft risks turning into PPT-style info cards, AI-flavored self-explaining text, public boundary disclaimers, or mismatched counts that do not match the visual content.

小红书笔记素材创作技能。当用户需要创建小红书笔记素材时使用这个技能。技能包含：根据用户的需求和提供的资料，撰写小红书笔记内容（标题+正文），生成图片卡片（封面+正文卡片），以及发布小红书笔记。

快手视频制作：Convert a selected Kuaishou topic or raw material into a short-video production plan. Use when Codex needs to write hooks, scripts, voiceover, shot lists, editing instructions, image/video generation prompts, subtitles, QA checks, or a final cut checklist before publishing on Kuaishou.

快手视频选题：Select and prioritize Kuaishou video topics, source clips, remake candidates, and publish opportunities. Use when the user needs help choosing which video to make or post, comparing existing footage, mining ideas, evaluating trend fit, or building a Kuaishou content queue before production.

抖音发布执行：Operate Douyin Creator Center upload and publishing preparation with browser assistance, including opening the upload page, handling file upload, filling title/caption/tags/cover settings, checking platform prompts, and stopping for final user confirmation. Use when the user asks Codex to upload, prepare, schedule, or publish a Douyin video.

抖音视频制作：Turn a selected Douyin topic, raw footage, or draft video into a practical production and editing plan with hook, beat sheet, subtitles, asset list, quality checks, and export guidance. Use when the user needs to make, revise, polish, or brief a Douyin short video before upload.

抖音视频选题：Select and prioritize Douyin video ideas, source videos, product angles, hooks, and reusable footage from user materials or market research. Use when the user asks which video to make, which draft to publish, how to choose a topic, how to score raw clips, or how to find a Douyin content angle before production.

I Have Adhd：Shape output for a reader with ADHD: lead with the next action, number multi-step work, restate state across turns, suppress tangents, give specific time estimates, make wins visible. Invoke with /i-have-adhd; stays on until "stop adhd mode".

This skill should be used when creating, auditing, revising, or producing content for the “念念旅行团” original Chinese realistic-fantasy travel IP, including its worldbuilding, slow-burn serial scripts, first-person travel Vlog scenes, 3D character continuity, celestial-palace daily life, and production handoff.

Stop Slop：Remove AI writing patterns from prose. Use when drafting, editing, or reviewing text to eliminate predictable AI tells and produce more human-like writing.

粘贴B站视频链接，一键解析返回无水印视频下载链接

抖音每日最热作品榜查询工具，输出单日点赞TOP50榜单，支持赛道筛选、30天回溯与订阅推送。

输入抖音号即可拉取该账号全部作品，解析无水印下载直链，支持批量一键下载视频与图片到本地

Jina 搜索：Search the web and read web pages with Jina AI Search/Reader. Use when Codex needs a lightweight web search, page extraction, or source gathering workflow through Jina using JINA_API_KEY, especially for Chinese/English research notes, prompt research, or summarizing URLs.

Extract and preserve WeChat public-account articles from mp.weixin.qq.com links into Markdown for AI research and evidence workflows. Use when Codex needs to read a Chinese WeChat/微信公众号 article, when Jina/ReadGZH returns captcha, 401/402/429/503, "环境异常", or when OpenCLI Browser Bridge should be used to download公众号文章正文, save evidence paths, and continue upstream research without bypassing verification.

粘贴视频号分享链接，一键解析视频标题、封面图和高清无水印下载地址

媒体 Use：Agent Media OS — resolve any media need (BGM, SFX, image, icon) into a frozen local file + ledger record. One verb (`resolve`) handles the full cascade — project cache, global cache, HeyGen catalog search, freeze, register. Keeps search noise on disk, hands the agent a path. Use when a composition needs background music, sound effects, images, or icons.

Figma：Use the Figma MCP server to fetch design context, screenshots, variables, and assets from Figma, and to translate Figma nodes into production code. Trigger when a task involves Figma URLs, node IDs, design-to-code implementation, or Figma MCP setup and troubleshooting.

Figma Use：**MANDATORY prerequisite** — you MUST invoke this skill BEFORE every `use_figma` tool call. NEVER call `use_figma` directly without loading this skill first. Skipping it causes common, hard-to-debug failures. Trigger whenever the user wants to perform a write action or a unique read action that requires JavaScript execution in the Figma file context — e.g. create/edit/delete nodes, set up variables or tokens, build components and variants, modify auto-layout or fills, bind variables to properties, or inspect file structure programmatically.

Figma 创建 New File：Create a new blank Figma file. Use when the user wants to create a new Figma design or FigJam file, or when you need a new file before calling use_figma. Handles plan resolution via whoami if needed. Usage — /figma-create-new-file [editorType] [fileName] (e.g. /figma-create-new-file figjam My Whiteboard)

Figma 创建 设计 系统 Rules：Generates custom design system rules for the user's codebase. Use when user says "create design system rules", "generate rules for my project", "set up design rules", "customize design system guidelines", or wants to establish project-specific conventions for Figma-to-code workflows. Requires Figma MCP server connection.

Figma 实现 设计：Translates Figma designs into production-ready application code with 1:1 visual fidelity. Use when implementing UI code from Figma files, when user mentions "implement design", "generate code", "implement component", provides Figma URLs, or asks to build components matching Figma specs. For Figma canvas writes via `use_figma`, use `figma-use`.

Figma 生成 库：Build or update a professional-grade design system in Figma from a codebase. Use when the user wants to create variables/tokens, build component libraries, set up theming (light/dark modes), document foundations, or reconcile gaps between code and Figma. This skill teaches WHAT to build and in WHAT ORDER — it complements the `figma-use` skill which teaches HOW to call the Plugin API. Both skills should be loaded together.

Figma 生成 设计：Use this skill alongside figma-use when the task involves translating an application page, view, or multi-section layout into Figma. Triggers: 'write to Figma', 'create in Figma from code', 'push page to Figma', 'take this app/page and build it in Figma', 'create a screen', 'build a landing page in Figma', 'update the Figma screen to match code'. This is the preferred workflow skill whenever the user wants to build or update a full page, screen, or view in Figma from code or a description. Discovers design system components, variables, and styles via search_design_system, imports them, and assembles screens incrementally section-by-section using design system tokens instead of hardcoded values.

GPT Tasteskill：Elite UX/UI & Advanced GSAP Motion Engineer. Enforces Python-driven true randomization for layout variance, strict AIDA page structure, wide editorial typography (bans 6-line wraps), gapless bento grids, strict GSAP ScrollTriggers (pinning, stacking, scrubbing), inline micro-images, and massive section spacing.

Hallmark：Anti-AI-slop design skill for greenfield pages, audits, redesigns, and design extraction from URLs or screenshots. Use when the user asks to build a new app or landing page, wants to redesign something, invokes Hallmark by name, or uses audit/redesign/study.

Icon Resource 路由：Choose verified icon resources for product UI and web experiences. Use when a task needs AI/LLM brand logos, model/provider selectors, integration directories, 3D visual icons, empty states, feature cards, or marketing illustrations, especially in AI, creator, content-production, and commerce products.

Impeccable：Use when the user wants to design, redesign, shape, critique, audit, polish, clarify, distill, harden, optimize, adapt, animate, colorize, extract, or otherwise improve a frontend interface. Covers websites, landing pages, dashboards, product UI, app shells, components, forms, settings, onboarding, and empty states. Handles UX review, visual hierarchy, information architecture, cognitive load, accessibility, performance, responsive behavior, theming, anti-patterns, typography, fonts, spacing, layout, alignment, color, motion, micro-interactions, UX copy, error states, edge cases, i18n, and reusable design systems or tokens. Also use for bland designs that need to become bolder or more delightful, loud designs that should become quieter, live browser iteration on UI elements, or ambitious visual effects that should feel technically extraordinary. Not for backend-only or non-UI tasks.

Use for any 念念 AI website, app, favicon, header, navigation, brand, or logo work. Provides the single user-approved authoritative logo asset and prevents substituting favicon, pink candidates, or deprecated logo variants.

Programmatic SEO：When the user wants to create SEO-driven pages at scale using templates and data. Also use when the user mentions "programmatic SEO," "template pages," "pages at scale," "directory pages," "location pages," "[keyword] + [city] pages," "comparison pages," "integration pages," "building many pages for SEO," "pSEO," "generate 100 pages," "data-driven pages," or "templated landing pages." Use this whenever someone wants to create many similar pages targeting different keywords or locations. For auditing existing SEO issues, see seo-audit. For content strategy planning, see content-strategy.

Redesign 技能：Upgrades existing websites and apps to premium quality. Audits current design, identifies generic AI patterns, and applies high-end design standards without breaking functionality. Works with any CSS framework or vanilla CSS.

Refactoring UI：Audit and fix visual hierarchy, spacing, color, and depth in web UIs. Use when the user mentions "my UI looks off", "fix the design", "Tailwind styling", "color palette", "visual hierarchy", "design system", "spacing scale", or "component styling". Also trigger when building consistent design tokens, creating dark mode themes, improving data visualization clarity, or polishing UI details before launch. Covers grayscale-first workflow, constrained design scales, shadows, and component styling. For typeface selection, see web-typography. For usability audits, see ux-heuristics.

Site Architecture：When the user wants to plan, map, or restructure their website's page hierarchy, navigation, URL structure, or internal linking. Also use when the user mentions "sitemap," "site map," "visual sitemap," "site structure," "page hierarchy," "information architecture," "IA," "navigation design," "URL structure," "breadcrumbs," "internal linking strategy," "website planning," "what pages do I need," "how should I organize my site," or "site navigation." Use this whenever someone is planning what pages a website should have and how they connect. NOT for XML sitemaps (that's technical SEO — see seo-audit). For SEO audits, see seo-audit. For structured data, see schema.

Taste 技能：Anti-slop frontend skill for landing pages, portfolios, and redesigns. The agent reads the brief, infers the right design direction, and ships interfaces that do not look templated. Real design systems when applicable, audit-first on redesigns, strict pre-flight check.

UX 启发式评估：Evaluate and improve interface usability using heuristic analysis. Use when the user mentions "usability audit", "UX review", "users are confused", "heuristic evaluation", "form usability", "navigation problems", "Nielsen heuristics", "cognitive walkthrough", or "usability testing". Also trigger when reviewing a design for usability issues, improving form completion rates, or evaluating information architecture and navigation. Covers Nielsens 10 heuristics, severity ratings, and information architecture. For visual design fixes, see refactoring-ui. For conversion-focused audits, see cro-methodology.

UX 文案：Create user-centered, accessible interface copy (microcopy) for digital products including buttons, labels, error messages, notifications, forms, onboarding, empty states, success messages, and help text. Use when writing or editing any text that appears in apps, websites, or software interfaces, designing conversational flows, establishing voice and tone guidelines, auditing product content for consistency and usability, reviewing UI strings, or improving existing interface copy. Applies UX writing best practices based on four quality standards — purposeful, concise, conversational, and clear. Includes accessibility guidelines, research-backed benchmarks (sentence length, comprehension rates, reading levels), expanded error patterns, tone adaptation frameworks, and comprehensive reference materials.

前端 设计：Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, artifacts, posters, or applications (examples include websites, landing pages, dashboards, React components, HTML/CSS layouts, or when styling/beautifying any web UI). Generates creative, polished code and UI design that avoids generic AI aesthetics.

商业 网页 Acceleration：Measure, diagnose, and improve commerce website delivery for product images, private media, preview video, playback, seeking, and downloads. Use when a storefront or content-creation workspace feels slow, media fails after a storage/server migration, CDN/cache changes are proposed, image/video traffic costs rise, or real multi-network performance must be verified without weakening authorization.

对网站公开页面发布、链接分享预览、SEO 元数据、canonical、robots/noindex、Open Graph、Twitter Card、图标、manifest 或 JSON-LD 做证据驱动的只读验收。仅在准备发布或分享真实页面、排查搜索收录/分享卡片、审查公开与私有页面索引边界时使用；不要为普通前端开发、纯视觉修改或尚无发布意图的页面自动触发。

网站产品路由：Website-specialized production router for Framer, Next.js, static sites, existing-site redesign, Figma-to-code, source-repo reuse, Image2 bounded assets, shadcn/ui boundaries, database/backend/admin, SEO, accessibility, performance, analytics, CRO, Playwright verification, deployment checks, website-quality-router handoff. Use when users ask build, redesign, launch, verify, iterate website, landing page, portfolio, SaaS site, dashboard, docs site, or web app.

网站质量路由：Website quality vertical router for auditing and improving real websites, landing pages, web apps, dashboards, portfolios, existing sites. Use when users ask to find/fix website aesthetic, frontend, UX, SEO, accessibility, performance, analytics, CRO, responsive, Playwright verification, launch-readiness, "no visible change" problems; routes each vertical to one champion skill plus at most one complement and requires browser evidence.

Top-level router for website and WeChat mini program product projects: intake, positioning, architecture, DESIGN.md, PROJECT_MANIFEST.json, design routing, website quality routing, implementation routing, launch, analytics, iteration. Use when users ask 做网站, 做小程序, 网站小程序总路由, 网站项目, 小程序项目, miniapp, landing page, SaaS site, portfolio, web app, or need end-to-end production workflow.

从 MotionSites、React Bits、Uiverse、Anime.js 与 Aceternity UI 发现、筛选、授权核对、复制或重构、集成并验收高质量网站动效。用于用户要求网站动效、Hero 动画、滚动效果、交互动效、动态背景、React/Next.js 动画组件、复制高质量动效源码、从动效案例选型，或需要判断某个动效是否适合商业网站时；也用于修复动效卡顿、移动端失控、重复动画引擎和视觉效果与业务不匹配的问题。

规划、审计和指导网站的视觉层级、GSAP 动效、互动、素材、性能和开发交接。用户提到网页视觉、网页动效、GSAP、ScrollTrigger、滚动叙事、Hero 动画、网站风格、动效方案、页面转场、3D/视频网页、动效很乱/很丑/很卡、或需要把网站需求交给开发时使用。本 Skill 是网站视觉与动效决策的权威来源；实际代码交给 website-product-router 和对应前端实现技能。

网页字体排版：Select, pair, and implement typefaces for web projects. Use when the user mentions "font pairing", "which typeface", "line height", "responsive typography", "web font loading", "type hierarchy", "variable fonts", "FOUT/FOIT", or "typographic scale". Also trigger when choosing between system fonts and web fonts, optimizing font loading performance, or designing readable long-form content layouts. Covers readability evaluation, CSS implementation, and performance optimization. For overall UI design systems, see refactoring-ui. For dramatic typographic experiences, see top-design.

顶级网页设计：Create award-winning, immersive web experiences at the level of Awwwards-featured agencies. Use when the user mentions "premium website", "portfolio site", "scroll animations", "Awwwards quality", "brand experience", "cinematic web design", "parallax storytelling", or "agency-quality site". Also trigger when building landing pages that need to impress, designing creative portfolios, or elevating a standard website to a memorable digital experience. Covers dramatic typography, purposeful motion, scroll-based composition, and performance-optimized animation. For foundational UI, see refactoring-ui. For type selection, see web-typography.

Apiskill：Build and package media-processing APIs that accept images, prompts, audio, and video, enqueue long-running jobs, run Playwright browser automation, and return durable output-video URLs. Use for API wrappers around web applications, RPA workflows, video-generation pipelines, uploads, API-key authentication, job status, callbacks, Docker deployment, or future tasks that need reusable FastAPI + Redis/Celery + Playwright + MinIO/S3 + PostgreSQL patterns.

GitHub Selected Projects：Manage and run the local GitHub selected-project bundle containing WorldMonitor, DeepTutor, code-review-graph, OfficeCLI, and Archscribe. Use when the user asks to inspect, install, start, stop, check, index, render, or use any of these projects locally, or asks to add another GitHub project to this local bundle.

Niannian AI 画布：Build, evaluate, integrate, and quality-check an AI media infinite canvas for Niannian AI. Use when the task involves node-based image or video generation, storyboard canvases, RunningHub/TapNow/ComfyUI-like workflows, Mayi Canvas, Daxiong Canvas, media nodes, provider execution, canvas persistence, or visual workflow integration into ai.cauai.fun.

Build, improve, brand, deploy, and verify 念念智剪, the OpenChatCut-based AI video editor at edit.cauai.fun. Use for automatic rough cuts, MiMo ASR or TTS, Qwen forced alignment, captions, transitions, exports, editor-brand changes, public deployment, and integration with Niannian AI.

Niannian 商业 Release Integrity：Prevent mixed website assets, stale cache releases, duplicate runtime ownership, workbench regressions, false completion, and unsafe promotion for Niannian AI websites. Use before any live ai.cauai.fun or dh.cauai.fun code, CSS, JavaScript, SSR, route, asset, authentication, media, or deployment change; also use when a change appears deployed but users still see old or broken behavior.

Niannian 商业 网站：Build, repair, and evolve the Niannian AI websites around their real user paths. Use for ai.cauai.fun main-site workbench, studio, Nomi canvas, projects, brand, and approved surfaces; use for dh.cauai.fun login, TZB, projects, materials, templates, six-step workbench, video delivery, pricing, and billing. Keep the two product boundaries separate.

Niannian 网页 开发：Run NianNian AI web changes from GitHub Issue through isolated branch, focused implementation, real verification, pull-request handoff, release candidate, and explicitly approved production deployment. Use for NianNian Studio, canvas, API, provider, asset-library, homepage, deployment, rollback, beta-readiness, CI, observability, data-recovery, or two-person collaboration work.

Playwright：Use when the task requires automating a real browser from the terminal (navigation, form filling, snapshots, screenshots, data extraction, UI-flow debugging) via `playwright-cli` or the bundled wrapper script.

Playwright 交互：Persistent browser and Electron interaction through `js_repl` for fast iterative UI debugging.

Procedural Gen：Generate game content procedurally — seeded deterministic RNG, value/Perlin/ Simplex noise for terrain and heightmaps, grid dungeon generation (rooms + corridors, BSP, random walk), and weighted loot/drop tables. Engine-neutral algorithms. Use when the user mentions procedural generation, perlin/simplex noise, random seed, dungeon generator, heightmap/terrain, or loot tables.

Prototype Fast：Build a playable prototype in about an hour to answer one question — is it fun? — with greybox primitives, a hard timebox, and explicit keep/kill criteria. Use when prototyping a mechanic, making a vertical slice or MVP, greyboxing/blockout, or judging throwaway vs keep.

Save Systems：Design save/load for game state — choosing what to serialize, file formats, save slots, atomic crash-safe writes, schema versioning and migration, and autosave. Engine-neutral. Use when the user mentions save system, save/load, game state persistence, save slots, autosave, save file corruption, or migrating old saves to a new version.

Screenshot：Use when the user explicitly asks for a desktop or system screenshot (full screen, specific app or window, or a pixel region), or when tool-specific capture capabilities are unavailable and an OS-level capture is needed.

Coordinate the user's Windows, Mac, and Tencent Cloud server Codex peers through verified threads, AppServer stdio, approved SSH aliases, Tailscale, and restricted relay channels. Use when the user asks Codex peers to communicate, inspect or send work across Windows/Mac/server Codex, run a one-shot exchange with reply writeback, control fixed Mac Codex employees, recover a cross-device job, or mentions 三端 Codex, Codex 点对点, Windows/Mac bridge, 腾讯云 Codex, server-main, Tailscale, SSH relay, AppServer, or 远程调度 Codex.

Wizstar Bitbrowser：Automate Wizstar in BitBrowser fingerprint browser windows with Playwright CDP. Use when the user asks to open a numbered BitBrowser window, log into wizstar.com with a stored local Google account, switch or reuse local account credentials, or navigate to Wizstar tools such as /tools/generate_video.

安全 Best Practices：Perform language and framework specific security best-practice reviews and suggest improvements. Trigger only when the user explicitly requests security best practices guidance, a security review/report, or secure-by-default coding help. Trigger only for supported languages (python, javascript/typescript, go). Do not trigger for general code review, debugging, or non-security tasks.

安全 威胁 Model：Repository-grounded threat modeling that enumerates trust boundaries, assets, attacker capabilities, abuse paths, and mitigations, and writes a concise Markdown threat model. Trigger only when the user explicitly asks to threat model a codebase or path, enumerate threats/abuse paths, or perform AppSec threat modeling. Do not trigger for general architecture summaries, code review, or non-security design work.

Safely identify, connect to, inventory, and operate the user's Windows, Linux, and Mac server fleet through recorded SSH aliases and host-specific boundaries. Use when the user asks about a server, VPS, 云服务器, SSH, deployment host, migration, backup, capacity, service, port, domain-to-host route, or which machine should receive work.

Doc：Use when the task involves reading, creating, or editing `.docx` documents, especially when formatting or layout fidelity matters; prefer `python-docx` plus the bundled `scripts/render_docx.py` for visual checks.

PDF：Use when tasks involve reading, creating, or reviewing PDF files where rendering and layout matter; prefer visual checks by rendering pages (Poppler) and use Python tools such as `reportlab`, `pdfplumber`, and `pypdf` for generation and extraction.

AI-driven multi-format SVG content generation system. Converts source documents (PDF/DOCX/URL/Markdown) into high-quality SVG pages and exports to PPTX through multi-role collaboration. Use when user asks to "create PPT", "make presentation", "生成PPT", "做PPT", "制作演示文稿", or mentions "ppt-master".

协调本机 Codex 与服务器 Hermes 的 AI Native 协作。用于涉及项目开发、线上发布、持续巡检、定时任务、服务器运行、跨项目上下文或需要在两者之间交接任务与结果的请求。

一人公司通用自动化执行 Skill：把 CSV、XLSX、JSON 或按目录组织的提示词任务编译为可审计的浏览器/桌面自动化队列，优先用 Playwright 完成网页输入与上传，并保留截图、日志、任务状态和人工确认门。 当用户要求批量录入提示词与参数、自动填写网页、上传素材、重复点击生成、读取回执、定时执行机械流程，或要为新的网页自动化建立可复用 Skill 时使用。涉及登录态、付费提交、生成、发布、删除或不可逆操作时必须停在确认门。 首个适配器为用户授权的 MXAI MJ 生图页面；不绕过验证码、审核、权限、付费或平台限制。

飞书文档证据驱动知识卡系统。适用于读取已登录飞书私有文档、截图、图片、视频和附件，提炼课程或案例内容，区分原文事实/操作方法/文章案例/整理者推导，生成单文件 Markdown 知识卡、批量索引和质量审计，并将已验证方法转为 ANS 可复用流程；涉及飞书资料总结、知识卡沉淀、课程索引、批量整理、截图文档审计或飞书内容转 Skill 时使用。

Route Feishu/Lark tasks to the narrowest workflow for account and session identification, document or wiki creation and editing, visual organization of AI instructions, permissions and sharing, drive capacity, and local-file upload or collaboration. Use when the user mentions 飞书、Lark、文档、知识库、云盘、空间、账号、权限、分享、上传或需要把内容沉淀到飞书。

将本地项目的剧本、章节、shotlist、角色、场景、道具和资产图片同步到本地 Jellyfish 工作台。用户要求把项目内容填入 Jellyfish、同步项目资料、导入剧本/分镜/资产时使用。依赖 jellyfish-mcp 提供的 jf_* 工具。

当要把一个项目的产出物、文档与交接材料整理后归档到项目资料库（WorkBuddy 原生资料库 / Project Drive / netdrive MCP 通道），供团队员工接手继续推进时使用。触发词：归档到资料库、上传资料库、交接文档、项目资料归档、团队交接、把手头的活交给员工、防止线程搞错。覆盖完整流程：线程状态锚点核对 → 交接文档编写 → 写操作确认门 → 上传三件套（file_upload → curl PUT → file_upload_complete）→ 落库验收 → 作废批次管理。同时内置两个高频坑的解法：本机 Python 解释器选择（托管 Python 精简版缺标准库）与中文路径 curl 上传。通用模板，可套用于任何含产出物的项目（短剧/网站/文档/资产图）。与「归档skill」（未完成事项审计交接）互补：本 skill 侧重把已产出物与资料完整归档，归档skill 侧重把未完成事项审计成交接清单。

在 Windows 上跑 teamai init/push 把本地 skills 同步到 GitHub 团队仓库的完整流程，含三个 Windows 特有坑及绕过方法

Use when Codex should run a task with Agent Team role separation: classifying task size, creating a numbered task spec, implementing from that spec, independently reviewing with evidence, or managing issue-based rework. Trigger when the user mentions Agent Team, 三窗口, 多窗口, task-spec, worker-report, review-report, issue 闭环, independent review, or wants complex Codex work split into requirements, implementation, and verification roles.

Use when Codex should use or maintain the local codex-agent-mem MCP memory layer for project continuity, thread bootstrap, scope guardrails, durable decisions, open-work checks, completion checks, snapshots, policies, or when the user mentions codex-agent-mem, memory, 记忆层, 上下文包, continuity, mem_context_pack, mem_open_work, or wants the workflow applied to every thread.

Computer Accelerator：Find process-level root causes and safely improve Windows computer performance, including CPU, memory pressure, disk I/O, temperatures, background processes, startup load, and stale development tasks. Use when the user says the computer is slow, freezes, becomes slower over time, has memory problems, needs health monitoring, or asks to install and apply System Informer, LibreHardwareMonitor, Mem Reduct, or Windows maintenance tools.

Deepseek Vision：Use whenever the task contains an attached image, pasted screenshot, local image path, or network image URL and image understanding is needed. This is the mandatory global image-understanding route: call deepseek-vision before OCR, Read, file metadata, or any other local image processing.

本地磁盘重要资料备份库。用于把一台电脑上有价值的内容按 P0/P1/P2 分类、只读复制到异地/异盘备份根、生成索引清单与完整性校验，并明确“备份优先于清理”的安全边界。触发词：备份电脑、打包资料、迁移文件、整理磁盘、防止丢失、做个备份、把重要文件拷到别的盘。

当用户提到"自定义模型遇到图片报错""纯文本模型不支持 image_url""给模型补视觉能力" "relatedModels.vision" "models.json 视觉路由" 时使用。 解决的问题：WorkBuddy/CodeBuddy 的自定义模型若 lacks supportsImages:true，遇到图片输入会直接报错； 通过 relatedModels.vision 字段把图片任务路由到已支持视觉的模型。

女娲造人：输入人名/主题/甚至只是模糊需求，自动深度调研→思维框架提炼→生成可运行的人物Skill。 两种入口：(1)明确人名→直接蒸馏 (2)模糊需求→诊断推荐→再蒸馏。 触发词：「造skill」「蒸馏XX」「女娲」「造人」「XX的思维方式」「做个XX视角」「更新XX的skill」。 模糊需求也触发：「我想提升决策质量」「有没有一种思维方式能帮我...」「我需要一个思维顾问」。 English triggers: "distill [person]", "nuwa", "create a [person] perspective skill", "how does [person] think", "I need a thinking advisor".

Qingli 技能：Analyze, compare, preview, and safely clean Windows disk space. Use when a user asks to inspect a full C drive, find large files, compare disk growth, remove temporary files or caches, clean Windows update leftovers, migrate growing application data, or schedule low-risk maintenance. This unified skill routes the request through four bundled community backends and keeps deletion separate from read-only analysis.

查询 tokenrhythm.studio（中转渠道）账号的剩余额度 / 余额 / 调用明细，以及批量接入多个账号到 WorkBuddy/CodeBuddy 自定义模型配置。 触发词：「查额度」「tokenrhythm 额度」「sk_tr 余额」「中转渠道还剩多少」「tokenrhythm 余额」「接入 tokenrhythm」「把 key 接入」「批量查额度」「多账号接入」。 关键事实：API Key（sk_tr_ 开头）本身查不到账户余额，必须改用网站登录态 sess_ 令牌当 Bearer 鉴权。

技能优化器：Diagnose and optimize Agent Skills (SKILL.md) with real session data and research-backed static analysis. Use when auditing skill quality, checking trigger/undertrigger problems, or reviewing SKILL.md structure. Works with WorkBuddy, Claude Code, Codex, and any Agent Skills-compatible agent.

Select, combine, score, and improve Codex skills across projects. Use when starting or resuming a project, choosing or auditing skills, recording skill performance, distilling verified repairs into reusable rules, or when the user asks to 沉淀方法论 / 自进化 / 更新全局 skill / 把第二遍修好的经验沉淀下来 / 不要只留在聊天里 / 把根因写回对应 Skill / 下次不要再犯 / 把修复变成可复用规则 / 第一遍没做好但第二遍找到了根因.

提示词管理库。当用户需要搜索、分类、索引、记录使用次数、查看统计或管理提示词时使用。触发词：提示词管理、搜索提示词、找提示词、提示词统计、prompt vault、记录提示词使用、提示词用了几次。

提示词技能路由：Route prompt-heavy tasks to the strongest local skill stack before drafting, revising, approving, or executing prompts. For AI video, compile authoritative route facts into a verified prompt-routing contract before Image2, first-frame, storyboard, video-task-spec, or channel work. Use for image/video prompts, character boards, storyboards, short-drama assets, first frames, commerce video, prompt QA, failure diagnosis, reusable prompt templates, or choosing a prompt method.

识别对话和真实任务中可复用的成功方法、失败根因、业务判断、质量标准与稳定用户偏好，用一段自然语言主动询问是否沉淀，并仅在用户明确许可后把底层逻辑路由到最窄的权威 Skill 或机器质量门；同时在本 Skill 自身发生漏问、乱问、Owner 误判、表达低效或沉淀效果不佳并获得真实反馈时，主动询问是否自进化本 Skill。用于用户表示“以后都这样”“这个有用/没用”、真实修复得到验证、同类步骤反复返工、某条逻辑能持续改善未来同类任务，或用户要求沉淀、自进化、复利、更新 Skill 路由、修正沉淀机制时；也用于用户说“统一上线”“单一发布者”“并行排查”“汇总上线”“不要互相覆盖”或“多线程查、一个线程发布”时。

创始人（老大/lsb）视角 Skill：基于 37,000+ 条真实指令与 485 个 git 提交蒸馏的决策操作系统。 用途：以创始人的思维框架分析问题、审阅方案、预判他的判断、代拟他视角的汇报。 触发词：「用老大的视角」「创始人会怎么判断」「以老板视角审一下」「laoda perspective」。

Pricing：When the user wants help with pricing decisions, packaging, or monetization strategy. Also use when the user mentions 'pricing,' 'pricing tiers,' 'freemium,' 'free trial,' 'packaging,' 'price increase,' 'value metric,' 'Van Westendorp,' 'willingness to pay,' 'monetization,' 'how much should I charge,' 'my pricing is wrong,' 'pricing page,' 'annual vs monthly,' 'per seat pricing,' or 'should I offer a free plan.' Use this whenever someone is figuring out what to charge or how to structure their plans. For in-app upgrade screens, see paywalls.

Product Marketing：When the user wants to create or update their product marketing context document. Also use when the user mentions 'product context,' 'marketing context,' 'set up context,' 'positioning,' 'who is my target audience,' 'describe my product,' 'ICP,' 'ideal customer profile,' or wants to avoid repeating foundational information across marketing tasks. Use this at the start of any new project before using other marketing skills — it creates `.agents/product-marketing.md` that all other skills reference for product, audience, and positioning context.

刺猬星球提示词与 AI 视觉创作方法知识库。用于学习、整理、检索和应用刺猬星球课程内容，尤其适用于提示词拆解、人物真实感、场景身份、行为逻辑、视觉语言翻译、反向控制、镜头与光线设计等任务；当用户提供课程截图、课程节次、课程笔记，或要求按刺猬星球方法写/审提示词时使用。该 Skill 是独立课程知识库，不隶属于童装带货或 Seedance 专用 Skill。

当用户想清理 WorkBuddy 线程、做项目交接审计、把"未完成且有价值的工作"归档并共享给团队成员时使用。触发词：归档、线程大扫除、项目交接、审计未完成事项、交接文档、团队资产库归档、把手头的活交给员工。该 Skill 会审计当前线程/工作区对应项目的未完成有价值工作，规范化分类索引，并写入项目共享资产库（资料库空间 / Project Drive），供接手员工直接执行。通用模板，可套用于任何项目线程。