Skills
97,222 packaged capabilities and playbooks for agents. Install any of them with one command.
Creative & media
Orchestrates multi-deliverable campaigns by dispatching Designer and Video Producer workflows in parallel. Use when the user requests 'run campaign', 'execute campaign', 'launch campaign production', or 'orchestrate deliverables'
Douyin short-drama feed — scans Douyin daily for hit short dramas, filters popular titles by likes, and intelligently clusters genres to generate an HTML daily report with covers, engagement data, and creative insights. Supports targeted queries by genre (time travel, domineering CEO, rebirth, etc.), creator, and date range. ⚠️ Before querying, the script validates input: keywords must match the short-drama genre dictionary (the genre names and all related terms specified in topic_keywords; for example, “打脸” matches terms related to the comeback genre). If matched, query using that keyword directly; otherwise, report “keyword does not meet query requirements” and suggest related terms, and **do not make an API request**. If no matching data is found or the overall data is insufficient, first ask whether to query again using a recommended genre, and proceed only after confirmation (do not expand automatically or fall back to all data automatically). If the date is outside the valid query range, notify the user and automatically fall back to the most recent date with data, without requiring confirmation. Use when the user needs a Douyin short-drama daily report, hit short dramas, short-drama trends, creative trends, or a custom genre query.
Visual self-review loop for motion graphics — render a draft, build contact sheets at every timeline beat, look at them, critique against a checklist, fix, repeat. Use after any change to a video's visuals or timing, and before calling a video done
Create style-driven slide images strictly with the Image 2.5 model, assemble them into image-only PPTX decks, and manage reusable visual style libraries from documents or visual references. Use when the user asks for a “PPT Skill”, “风格驱动 PPT”, “提炼风格做 PPT”, “调用某个风格做 PPT”, “图片版 PPT”, “保存 PPT 风格”, “列出 PPT 风格”, “文档生成 PPT”, “文章生成 PPT”, “把文档做成演示文稿”, or wants to extract, save, reuse, and apply visual style keywords specifically for visual slide/image deck creation.
Alibaba Cloud Bailian Video Analysis Skill. Use for intelligent video comprehension and analysis via the Bailian (QuanMiaoLightApp) API. **Required API product**: QuanMiaoLightApp (version 2024-08-01) **Required API actions**: SubmitVideoAnalysisTask, GetVideoAnalysisTask **DO NOT use**: videorecog, Mts, or any other product for video analysis. Triggers: "analyze video", "understand video", "analyze the local video /temp/xxx.mp4", "analyze the local video https://xxx.com/temp/xxx.mp4", "what is this video about", "summarize this video", "split video into shots", "video comprehension", "extract video insights", "transcribe video", "extract video captions", "generate video title", "generate video outline", "video mindmap".
Divergent ideation and corpus expansion via Verbalized Sampling plus extended thinking — single-shot generates several internally diverse candidates and surfaces the strongest, or expands a seed corpus into a diverse dataset for evals and test sets. USE WHEN be creative, brainstorm, divergent ideas, creative solutions, maximum creativity, tree of thoughts, radically different, name this, creative angle, expand this corpus, synthetic data, generate diverse examples, create test set. NOT FOR multi-cycle evolutionary ideation with Lamarckian meta-learning (use Ideate).
Hosted ChatCut plugin sessions only (the `chatcut` MCP server). If the conversation is driving ChatCut Desktop (a `chatcut_desktop*` MCP server), load this skill only when the user explicitly chooses the plugin/web surface — desktop sessions otherwise ship their own instructions and tools. Export or deliver a ChatCut project through the hosted connector, including video, audio, subtitles, XML, and render-status checks.
Generate or edit raster images through the HappyCode OpenAI-compatible GPT Image 2 API. Use when the user requests HappyCode or this project uses it as the default backend for new images, reference-image transformation, composition, inpainting, editing, batches of 2-5 images, or first-use configuration.
Provider-independent performance direction for generated video, avatar/spokesperson clips, animation, ads, film scenes, social content, and voice-led media. Use when an agent must cast or direct synthetic performers, avatars, animated characters, talking heads, AI video characters, or voice performances; plan acting beats, objective/obstacle/action, blocking, gesture, eye-line, facial expression, body language, lip-sync, voice alignment, rehearsal prompts, iteration, continuity, consent, and performance QA.
Export sprites to PNG, GIF, or spritesheet formats with JSON metadata for game engines. Use when the user wants to "export", "save", "output", "render", "generate", "create file", mentions file formats like "PNG", "GIF", "animated GIF", "spritesheet", "sprite sheet", "texture atlas", "tile sheet", or game engine integration with "Unity", "Godot", "Phaser", "Unreal", "GameMaker". Trigger on layout terms ("horizontal", "vertical", "grid", "packed", "strip"), scaling ("2x", "4x", "upscale", "pixel-perfect"), file operations ("save as", "export to", "output to"), metadata formats ("JSON", "XML", "metadata", "atlas data"), and delivery terms ("for web", "for game", "for Twitter", "for itch.io", "optimized").
Orchestrates translation of motion designer video specifications into working Remotion code by coordinating specialized agents. Acts as pipeline coordinator that delegates to remotion-scaffold, remotion-animation, remotion-composition, and remotion-component-gen. Use when you have a complete VIDEO_SPEC.md and need full Remotion implementation.
Automatically generate a Seedance 2.0 video on Higgsfield using Playwright browser automation. Use when the user provides an image and wants to create an AI video from it on Higgsfield. Triggers on requests like "generate a video", "create a seedance video", "make a video on higgsfield", "auto-generate video", or any request to automatically submit a video generation job to higgsfield.ai. Requires Playwright MCP tools.
Guide for editing speech-led videos where spoken delivery or conversation drives the cut — single-speaker talking-head / 口播, two- or multi-speaker interview / 访谈, video podcast, lecture, tutorial, course, and similar formats. Load before analyzing, planning, or performing non-trivial edits of those formats, including speech cleanup (剪口播 / 口播剪辑 / 去口癖 / clean up fillers / smooth speech), pause or repeated-take removal, motion graphics layered onto the footage (口播加 MG / 加动画), B-roll (加 B-roll / add B-roll), music, or captions. For motion graphics specifically, use this together with the active Motion Graphics skill/workflow available in the current ChatCut environment — this skill adds speech-specific guidance (rhythm-aware timing, frame-aware placement, subject/caption protection, placement verification).
Turn ONE topic into a finished Vox-style paper-collage explainer / ad video, end to end with Aliyun Bailian CLI + local ffmpeg — script, collage keyframes, motion, voice-over, music, captions, all automated. Use this whenever the user wants a "Vox style" video, a paper/torn-paper collage animation, a "motion collage", a narrated explainer or short ad built from AI-generated collage posters, a scrapbook-style tribute, or wants to turn a topic / product / person into a punchy narrated collage video — even if they don't say the word "Vox". Also use when reproducing Stav Zilber / rom1trs / Higgsfield-style collage ad workflows. Three input modalities: a topic (B-roll), a talking-head video (A-roll mode), or a single photo of a person/product anchored into the collage (C-roll mode). Triggers: "vox video", "collage video", "motion collage", "paper collage explainer", "make a collage ad", "turn this topic into a collage video", "turn my photo/this product shot into a collage video".
Use when the user wants to write or publish a WeChat Official Account article—they provide scattered ideas, drafts, or notes and ask you to polish them, generate a cover image and explanatory illustrations, or prepare them for upload to mp.weixin.qq.com. Triggers: “write a WeChat article,” “official account,” “polish,” “cover image,” “publish to official account,” “/wjs-publishing-wechat”.
Use when someone asks “does this sound AI?”, “check if this is AI-written”, “what gives this away as AI”, “run ai-check on this”, or “score this text”. Also use when reviewing a draft for AI tells before publishing, or when a piece of text reads as suspiciously polished, generic, or pattern-y and the user wants a forensic breakdown of why.
World-class expertise in TCG/CCG design, resource systems, and metagame balance. Use when “card game design, TCG design, CCG design, trading card game, collectible card game, mana curve, deck building, card templating, keyword mechanic, set design, draft format, limited format, card rarity, color pie, faction design, card balance, magic the gathering style, hearthstone style, deckbuilder, card text, rules text, card-game, tcg, ccg, trading-card, collectible-card, deck-building, mana-curve, draft, limited, constructed, set-design, card-templating, mechanics, keywords, rarity, balance” are mentioned.
Universal AI image generation supporting OpenAI DALL·E / gpt-image, Google Gemini Image / Imagen, Replicate (Flux / SDXL / any model), Stability AI, FAL, Ark (Seedream 4.5), Bailian (qwen-image / wanx), and SiliconFlow. Use this skill whenever the user asks to generate, create, draw, illustrate, render, or synthesize images from text prompts or reference images. Typical phrases include "draw a ...", "generate an image of ...", "画一张 ...", "给我来张图", "make a poster of ...", "create an illustration ...", or any mention of image-generation model families like DALL·E, gpt-image, Flux, SDXL, Seedream, Imagen, Gemini image, Kolors, or Wanx. Always use this skill even if the user does not name a specific model — pick a provider based on their EXTEND.md defaults or available API keys in the environment. Do NOT use this skill when the user explicitly mentions 即梦 / Dreamina / Jimeng — those go to happy-dreamina instead.
Analyzes events through poetic lens using close reading, metaphor analysis, imagery, rhythm, form analysis, and attention to language's emotional and aesthetic dimensions. Provides insights on emotional truth, symbolic meaning, human experience, aesthetic impact, and expressive depth. Use when: Understanding emotional dimensions, symbolic meaning, communication impact, cultural resonance, human experience. Evaluates: Imagery, metaphor, rhythm, emotional truth, symbolic depth, aesthetic power, resonance, ambiguity.
Use when optimizing 3D models for web delivery: reducing file size, compressing meshes, optimizing textures, or generating LOD. Provides a complete GLTF optimization pipeline using Draco, KTX2, mesh simplification, and gltf-transform. Keywords: optimize model, reduce file size, Draco compression, KTX2, gltf-transform, mesh simplification, LOD, GLTF optimize, bundle size, model too big, slow loading, compress 3D model.
Generate images and text through the reverse-engineered Gemini Web API. Supports text generation, image generation from prompts, visual input from reference images, and multi-turn conversations. Use when another skill needs an image-generation backend or the user asks to “generate an image with Gemini” or for “Gemini text generation”.
Hosted ChatCut plugin sessions only (the `chatcut` MCP server). If the conversation is driving ChatCut Desktop (a `chatcut_desktop*` MCP server), load this skill only when the user explicitly chooses the plugin/web surface — desktop sessions otherwise ship their own instructions and tools. Verify that ChatCut plugin edits are reflected in project structure and visible timeline output before reporting success.
Creative writing skill for capturing story brainstorming. Use when the user is exploring narrative ideas, discussing characters, planning episodes, or thinking through story possibilities. Creates minimal working notes that preserve creative freedom by recording only what was stated and marking sources.
Live-action talking-head video topic decision lead writer. Generate opinion/tutorial candidates from current signals in AIHOT and installed OpenCLI, verify the promise to beginner audiences, alignment with Douyin semantics, and PRIMARY original sources, then choose one topic that can be produced and reviewed. Use when the user says "抓热点" "AI热点" "找选题" "选题" "今天做什么" "redian", or bianpai routes to ①; for Douyin search alone, use laohan-douyinsousuo instead.
Use when designing motion concepts, animation systems, choreography, timing, easing, reduced-motion behavior, or deciding which motion stack to use for web UI, video, GIF, Rive, Lottie, GSAP, Motion, Remotion, or Three.js work
Generate videos via Venice. Covers the async /video/quote + /video/queue + /video/retrieve + /video/complete loop, text-to-video, image-to-video, first/last-frame transitions, video-to-video and upscale, audio input, reference images / videos / audio (R2V) and their per-model caps, Kling elements and scene images, Seedance source-matched edit/extend and omni_reference_task_type, H3 Max Multi-Angle camera_trajectory, and replayed async failures. POST /video/transcriptions is retired (410) — use chat completions with video_url.
Turn one approved master still into a character's whole animated sprite set - one image-to-video clip per action (idle, walk, run, attack, jump, hurt, cast and so on), all from that still, generated through generate2dmedia route_media.py (the xAI API when a key is configured, else the local Grok CLI) - then gate each take, key it with a soft matte, register it to the master, pick loops or retime actions, finish (HD by default, pixel on request) and package fixed-canvas frames, alpha WebM, packed MP4 for iPhone and a PNG fallback with engine metadata and a JS runtime. Also processes a supplied clip. Use for character and creature animation, fluid motion and any video-to-sprite work. Not for FX, icon or prop sheets (generate2dsprite), maps or plate loops (generate2dmap), code-drawn animation (codeart2d, only on request) or calling a video API by itself (generate2dmedia).
Automatically collect trending AI content, write official account articles, generate accompanying images, and publish them to the WeChat Official Account draft folder. Trigger when the user says “发布AI热点,” “写公众号文章,” “采集AI内容,” or “publish AI news.”
Export from Blender 5.2 with verified presets - Roblox (FBX/glTF, cages, R15, SurfaceAppearance), Unity (FBX, Humanoid), Unreal (FBX skeletal, Nanite, UCX), Godot/web (GLB, meshopt/Draco, WebP), USD. Scale/axes, smoothing, tangents, leaf bones, clips, batch export, post-export verification. Use whenever an asset leaves Blender.
Generate, refine, and validate Character Card V3 cards and lorebooks. Use when creating new characters from a concept, converting or upgrading existing cards to CCv3, improving weak character behavior consistency, or building lorebook entries that steer scenario outcomes while staying token-efficient.
A guide to good Chinese writing: improve Chinese writing and revise existing articles. Based on the writing insights of Paul Graham, Yu Guangzhong, and Wang Dingjun. Use when the user says “polish,” “rewrite,” or “make the text flow better”; asks to “improve the writing,” “write it better,” or produce “high-quality Chinese”; provides an article for editing or polishing; or needs high-quality Chinese content. For technical documents (README, API documentation, CLI instructions), use conservative mode (see mode three below) to avoid changing structured lists and terminology sequences. Triggers: /good-writing, /润稿, good writing, 润稿, 改写, rewrite, 文笔, 让文字更顺, 技术文件润饰, conservative mode
Based on the real-person AI voiceover script and Jeffrey identity reference image, use the 秋芝03KV visual system by default, select the scene template that best matches the semantics, and generate only three full-size covers for the same concept in 3:4, 4:3, and 16:9. Use when the user says “generate cover prompts,” “make a cover,” “cover,” “cover image,” or “cover wording,” or when the workflow reaches ⑥.
Make clean, high-taste motion graphics videos (explainers, product promos, tutorials, social reels) with Remotion. Use when the user asks for a motion graphics video, an animated explainer, a promo or launch video, a reel/short with kinetic text, or wants to copy the style of a reference video. Covers research, script, voiceover timing, original music and sound effects, a ready Remotion template with a locked style, automatic overlap QA, and final render.
Neta capability index and routing skill—helps choose the right Neta-related skill (neta-space / neta-creative / neta-adventure / neta-community / neta-suggest). Use this skill when you need to understand Neta's overall capabilities, decide which skill to use for the current task, or migrate from older documentation.
Designs Amazon-compliant main product images and the 7 secondary image types (Infographic / Multi-angle / Detail Shot / Lifestyle / Variants / What's in Box / Size Reference) — with platform-mandatory main-image rules, conversion-rate-tuned secondary types per product category, and multi-image consistency anchored to the main image URL. Use when the user says "Amazon main image", "亚马逊主图", "product main image", "white background product image", "Amazon listing image", "亚马逊副图", "secondary images", "product carousel images", or any request for the 1:1 product image suite that lives on the Amazon detail page carousel. For A+ Content / detail-page modules (21:9 Hero Banner and 3:2 module layouts), see the `product-shots-detail-page` skill.
Writing a Succession-style ensemble series (继承之战群像剧写法) — distilled from Jesse Armstrong's complete shooting scripts for all four seasons, 39 episodes (Faber 2023), with his introductions and footnotes, plus Frank Rich's and Lucy Prebble's essays on the writers' room. Covers the invisible-act three-test for scripts with no ACT marks, container episodes built on a ceremony's running order, pressure chambers, one core question per episode, finales with fewer and longer scenes, the absent protagonist at the climax, humiliation passed downward, stage directions carrying subtext and "maybe", reversals turned on one word, writing long, alts, the mega-chart, and engineering an ending once the plot engine loses pressure. Use when the user wants a dynastic family drama, a media-power ensemble, a wedding/funeral/summit set piece, a streaming hour with no act breaks, a finale that redraws alliances, or asks how to "Succession-ize" a script or kill a lead mid-season.
Generate a detailed AI image prompt for creating an attractive female UGC ad character. Use when the user wants to create a “hot girl” character image for UGC ads, influencer-style content, product testimonials, or social media ad creatives. Triggers on requests like “create a UGC girl,” “hot girl for ads,” “female influencer image,” “UGC character,” “testimonial girl,” or any request for a female character for ad/UGC content.
Turn an approved script, provided talking-head/voiceover media, and gathered assets into a polished, creator-grade vertical (9:16) short rendered with HyperFrames. Encodes a non-negotiable init→preview→lint→render→verify workflow; a reusable editorial house look (mixed serif/sans headlines, paper-stage and dark-grid grounds, stacked splits with seam captions, highlight-bar supers, name pills, headline cards, UI-card stacks); engineered hooks; A-roll/B-roll spine; non-metronomic cuts down to 0.2s; split-screen layouts with deliberate full-frame transitions; eased camera moves (zoom/push) with explicit anti-jitter rules; kinetic captions/keyword emphasis; themed motion graphics (number/spec/poster/callout cards); logo animation; layered SFX mixed under voice; visual theming derived from a style pack (styles/<style>/FRAME.md); asset integration via a manifest; and required verification (preview screenshots + ffprobe + frame spot-checks + a render gate script). Adopts hyperframes-helper conventions (GSAP power-eases, glow captions, lint gotchas) and extends them with an editorial look, a camera-motion system, style theming, SFX synthesis, and a verification gate. Use when you have a script and media and need a finished short.
Wang Jianshuo's thinking framework and style of expression. Distilled from 7 in-depth research studies (about 1 million English blog words + about 1.09 million Chinese blog characters, 2002–2022, all primary sources): 7 core mental models, 10 decision heuristics, and a complete bilingual expression DNA. Use to write, respond, and think in Wang Jianshuo's identity and voice—plainspoken, sincere, curious, fond of everyday metaphors and coined terms, repeatedly building bridges between the concrete and the abstract, and never writing about something not personally verified. Trigger when users mention “from Wang Jianshuo's perspective,” “how would Wang Jianshuo see it,” “Wang Jianshuo mode,” “write like Wang Jianshu,” “Jian Shuo Wang perspective,” or “switch to Wang Jianshu.” Also trigger for requests like “help me think about this from Wang Jianshu's perspective” or “how would Wang Jianshu write this.” Once activated, maintain Wang Jianshu's identity in all subsequent replies in the conversation until the user explicitly says “exit”—no need to name it again each turn. Not applicable when users want an objective introduction or factual information about Wang Jianshuo (“Who is Wang Jianshuo?” “What year did he start his business?”); answer normally without role-playing.
Generate, refine, and validate Character Card V3 cards and lorebooks. Use when creating new characters from a concept, converting or upgrading existing cards to CCv3, improving weak character behavior consistency, or building lorebook entries that steer scenario outcomes while staying token-efficient.
Complete audio encoding and normalization system. PROACTIVELY activate for: (1) Audio codec selection (AAC, MP3, Opus, FLAC), (2) Loudness normalization (EBU R128, loudnorm), (3) Audio extraction from video, (4) Format conversion, (5) Volume adjustment and dynamics, (6) Noise reduction and EQ, (7) Channel operations (stereo/mono/surround), (8) Sample rate and bit depth conversion, (9) Audio fade in/out and crossfades, (10) Podcast and broadcast processing chains. Provides: Codec comparison tables, loudness standards reference, two-pass normalization scripts, professional mastering chains. Ensures: Broadcast-compliant audio with proper loudness and quality
Balance game difficulty, resources, rewards, probability, progression, economies, and dominant strategies. Use when tuning combat, loot, upgrades, scoring, level curves, AI difficulty, risk-reward choices, or any system where numbers shape player decisions.
Turn any text or information into magazine-style HTML info cards, and automatically capture and save them as images. Suitable for sharing on platforms such as X (Twitter), WeChat, and Xiaohongshu. Trigger phrases: “生成信息卡”, “做张信息卡”, “把这段内容做成卡片”, “信息卡片”, “make info card”, “generate card”. Card features: large text, strong
Provider-independent lighting direction for AI-generated images, video clips, ads, film scenes, avatars, product shots, interviews, beauty work, and social content. Use when an agent must specify, translate, maintain, troubleshoot, or QA motivated lighting, key/fill/back/rim/practical lights, hard or soft light, color temperature, exposure, contrast, time-of-day, continuity, reference strategy, or safety/accessibility risks in generated media.
Alibaba Cloud IMS (Intelligent Media Services) based video translation Skill. Supports subtitle extraction (ASR/OCR), translation, and speech synthesis translation modes. Trigger words: "视频翻译", "translate video", "翻译视频", "字幕翻译", "video translation"
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.