Skills
96,981 packaged capabilities and playbooks for agents. Install any of them with one command.
Creative & media
Generates article cover images with 5 dimensions (type, palette, rendering, text, mood) combining 10 color palettes and 7 rendering styles. Supports cinematic (2.35:1), widescreen (16:9), and square (1:1) aspects. Use when user asks to "generate cover image", "create article cover", or "make cover".
A Chinese content quality review tool and main entry point (gateway) for the writing suite. Comprehensively reviews a Chinese draft from four perspectives—editor, reader, author, and platform—across 16 criteria, then outputs a score, issue list, and routing recommendations. Does not edit; it only reviews and routes. Triggers: “review how this reads,” “does this have any issues,” “check this article,” “is this draft acceptable,” “review this for me,” “article quality review,” “Chinese writing diagnosis,” “can this be published,” “what needs changing in this draft,” “how should I revise it.” EN: “review my Chinese article,” “check this article quality,” “Chinese writing diagnosis,” “proofread this piece,” “is this article good enough”. Not for: direct rewriting (→ plainspoken-writing / editor-revisor); title, layout, or punctuation edits only (→ packaging-workshop); finding one’s voice (→ voice-dissolver); verifying the truth of facts, data, or quotations (→ L0 fact-review).
A complete workflow that turns descriptive text into comic pages. Supports two styles, sketch (black-and-white line art) and colored (color comics), reads text from direct input or files (.docx, .txt), allows editing individual pages after generation, and outputs the final result as a PDF document.
Inspect, capture, create, and refine graphics in an open Singular Composer session, including composition elements, control nodes, layouts, timeline animations, and composition scripts. Use when AI needs to control Composer, verify graphics and runtime behavior, or respond to generate improvement handoff with retrospective feedback on user corrections; the retrospective requires no paired session.
Transform long-form content into 15+ platform-ready assets. Use when repurposing a blog post, article, or guide into social media content. One piece becomes threads, carousels, audiograms, video scripts, quizzes, and discussion prompts across all major platforms.
Deconstruct viral web novels on benchmark platforms: learn the structure, don't copy the words. Use when the user wants to deconstruct a book, analyze why a hit novel went viral, benchmark against a popular title, extract the opening/hook/power-up/pacing formulas of a hit, or asks to "write my own following this hit's approach". Outputs reusable structural templates and strictly upholds the no-plagiarism line.
Dialogue and lines for Chinese web fiction. Use when the user wants to write or revise dialogue, design quotable lines for iconic scenes, write face-slapping or flexing lines, angsty or sweet-romance lines, complains that "the dialogue is too stiff / too formal / sounds like AI / every character talks the same", or wants each character's speech to be distinguishable. Produces lines that are colloquial, carry subtext, and match the character's voice.
The art and science of creating AI-powered digital presenters, avatars, and synthetic spokespersons. This skill covers HeyGen, Synthesia, D-ID, Tavus, and the emerging landscape of photorealistic AI humans that can speak any script in any language. Digital humans aren't replacing human presenters—they're enabling scale that humans can't achieve. A product demo in 50 languages. Personalized video messages for thousands of customers. 24/7 customer support with a friendly face. Training videos that can be updated without reshoots. The practitioners of this skill understand both the power and the responsibility. They know when digital humans enhance experiences and when they feel uncanny. They navigate the ethics of synthetic media thoughtfully. They create AI presenters that feel helpful, not deceptive. Use when "digital human, AI avatar, AI presenter, HeyGen, Synthesia, D-ID, Tavus, synthetic, talking head AI, AI spokesperson, personalized video, video at scale, multilingual video, AI actor, digital-humans, avatar, ai-presenter, synthesia, heygen, d-id, synthetic-media, personalization, multilingual" mentioned.
Designs plot structure and story arcs using McKee's value turns, Campbell's monomyth, and Sanderson's Promise/Progress/Payoff model to ensure causal integrity and emotional resonance. Use when constructing narratives, outlining novels, or diagnosing why a story feels broken.
Hook design for Chinese web fiction. Use when the user wants to write an opening/first line, a golden-three-chapters hook, a chapter-ending cliffhanger, a paywall-node suspense point, or complains that the opening fails to keep readers. Produces several candidate hook options and explains the underlying hook mechanics.
HTML PPT refined presentation designer - turns user content into polished HTML presentation pages, with LLM-driven design as the priority and support for multiple visual styles, motion design, and a color cognition system. Trigger words: PPT, presentation, HTML slides, slide, presentation, make a PPT, create a presentation.
Removes traces of AI-generated text. Suitable for editing or reviewing text to make it sound more natural and human-written. Based on the comprehensive Wikipedia guide on "AI writing characteristics." Detects and fixes patterns such as: exaggerated symbolism, promotional language, superficial analysis ending in "-ing," vague attribution, overuse of dashes, the rule of three, AI-typical vocabulary, negative parallelism, and excessive transitional phrases.
Generate images via multiple mechanisms. Supports: (1) End-to-end AI generation via Gemini/ZenMux — given a prompt, directly output an image. (2) Code-based rendering — generate HTML/React single-file, render to PNG via Playwright. (3) Prompt engineering — generate optimized prompts for external models (nano-banana-pro, etc.). Trigger words: image, generate image, 生图, render, poster, 海报, banner, card, 卡片
End-to-end UGC video in one call. Provide EITHER a text description of the person, OR a portrait URL (R2-hosted), OR an uploaded image. The pipeline auto-generates the missing portrait, builds a character sheet, and produces a 5/10/15s vertical selfie video with native lip-synced audio of your script.
Generate and edit high-quality images using Google's Nano Banana Pro (Gemini 3 Pro Image) AI model. Use this skill when working with AI image generation or editing tasks including (1) Creating images from text prompts with professional quality, (2) Editing existing images with natural language instructions, (3) Multi-turn iterative image refinement, (4) Product photography and marketing visuals, (5) Creating infographics and diagrams with text, (6) Photorealistic and artistic image generation, (7) Batch image generation workflows, (8) Testing and troubleshooting Gemini image generation API. Supports up to 4K resolution, multiple aspect ratios, reference images, and advanced prompting techniques for cinema-quality results.
Use narova whenever the user names Narova or reel.config, or wants a deterministic scene-scripted video: narrated/captioned explainers, multi-host dialogue, prompt/script/README-to-video, product walkthroughs with real browser actions, or source-grounded videos from sites, articles, docs, and repositories. Supports optional local neural TTS, word-synced captions and reveals, music/SFX, platform export presets, SRT/VTT, 2D HTML/CSS/SVG, Three.js/WebGL 3D, mixed compositing, AI clips, and silent marker-driven work. It turns a scene script into MP4 through HyperFrames or the browserless Skia/FFmpeg provider; scene.threeModule is the raw 3D escape hatch. Use plain HyperFrames for unrelated silent motion graphics.
Use when designing interactive narratives, branching storylines, player agency systems, non-linear story structures, or act frameworks for games. Triggers: branching narrative, player agency, story arc, interactive story, dialogue.
Generate covers for Xiaohongshu and Bilibili AI tool walkthrough videos, with script mode and autonomous Agent mode, real video evidence, three aspect-ratio outputs, and an optional creator avatar. Trigger only when the user explicitly mentions oil-cover, $oil-cover, uses oil-cover, or names this Skill; do not auto-trigger for ordinary cover, thumbnail, or video cover requests.
The packaging workshop—the final step in the writing suite. Handles presentation after content is finalized: title polishing, opening hooks, layout optimization, punctuation cleanup (integrates with zan-punct-fix), image suggestions, and platform adaptation. Does not change the body text. Its six modules can be triggered independently. Triggers: “polish the title,” “this title isn’t good enough,” “how should I write the opening,” “check the layout,” “the paragraphs are too dense,” “check punctuation,” “review the images,” “which platform suits this article,” “I’m ready to publish—review the packaging,” “check the presentation,” “the content is fine but it’s tiring to read,” “pre-publish check.” EN: “polish the title,” “fix the opening hook,” “check layout,” “fix punctuation,” “review images,” “platform formatting,” “pre-publish checklist”. Not for: rewriting the body text (→ plainspoken-writing / editor-revisor); content quality diagnosis (→ chinese-write-checker); voice calibration (→ voice-dissolver).
Build and render a papercut episode from a storyboard with the library. Scaffolds the project, builds scenes with parallel agents, assets and audio, renders the mp4, then harvests reusables. Use for "build the reel", "scaffold an episode", "rebuild scene 4", "render it"
Decide what a papercut reel should be. Picks the topic, one familiar format, the audience level, the comment-keyword funnel and caption, studies competitor reels, and runs the post-publish autopsy. Use before any hook or script, or when reviewing how a posted reel did
Tired of hunting through endless music libraries just to find a track that fits your video's mood? The pixabay-music skill connects your video workflow directly to Pixabay's vast royalty-free music catalog, letting you search, preview, and apply background music without ever leaving your editing session. Find tracks by genre, mood, or tempo, then layer them onto your mp4, mov, avi, webm, or mkv files instantly. Perfect for content creators, marketers, and educators who need great audio fast.
Build and revise Rive Luau scripts across Node, Layout, Converter, Path Effect, Transition Condition, Listener Action, Util, and Test protocols. Use when the user asks to write or modify Rive scripts, choose protocols, wire script inputs or data binding, debug runtime behavior, or plan unit tests for script logic.
The period feel comes from camera defects and one tiny everyday incident. Write the camcorder's autofocus hunting, exposure shifts and handheld shake as an explicit list, keep the story small, and the clip reads as a real old tape. Use when the user wants a Seedance 2.5 / 2.0 video prompt for this kind of clip (Early-2000s DV home video; 中文触发词: 早年 DV 家庭录像, DV, 家庭录像, VHS, 复古, 年代感), or asks how published cases of this type were prompted. Read references/cases.md before drafting.
Generate videos with ByteDance Seedance 2.0 via inference.sh CLI. Unified model for text-to-video, image-to-video, and reference-to-video with synchronized audio, up to 1080p, 4-15s duration. Pro and Fast variants. Studio variants with private asset library for portrait consistency. Use for: social media videos, music videos, product demos, animated content, AI video with sound. Triggers: seedance, seedance 2, bytedance video, seedance t2v, seedance i2v, seedance r2v, video with audio, seedance 2.0, bytedance seedance, seedance studio
Use when writing any short-form video script, hook, opening line, or scroll-stopping content (TikTok, Reels, Shorts) for any niche or business — when the first 2 seconds must grab attention. Also the shared psychology core (six storytelling levers) that the other storytelling-suite skills reference
Build and optimize Three.js post-processing with EffectComposer, RenderPass, bloom, anti-aliasing, SSAO, depth of field, outlines, color correction, custom ShaderPasses, render targets, resize handling, and frame-budget controls. Use when adding or fixing screen-space effects, composer pass order, render target sizing, or visual-effect cost. Triggers on: Three.js postprocessing, EffectComposer, RenderPass, UnrealBloomPass, FXAA, SMAA, SSAO, depth of field, bloom, ShaderPass, render target, screen-space effect.
Analyze the photograph uploaded by the user, deconstruct its visual evidence, abstract and distill the essential relationships, then reconstruct them as a new nonliteral lower-panel composition guided by 19 bundled references. Use for CLEAN upper-photo/lower-abstraction editorial artworks with a number, date, and short phrase in the lower panel. This is abstraction and reconstruction, never style conversion or a miniature redraw. Preserve the exact user-uploaded photograph unchanged above.
Judges whether a user's input suits the Vibe Creating style of video-prompt writing, and when it does, distills single-scene prompts, multi-shot descriptions, emotional imagery, or mixed input into prompts that are easier for a video model to generate from — while preserving any user-specified dialogue, voiceover, music, sound effects, and other hard constraints. Use when a user wants to turn an idea, story, feeling, or rough/over-specified prompt into a strong text-to-video prompt (Seedance, Sora, Kling, Veo, Runway, etc.), or asks to "rewrite", "improve", "clean up", or "vibe-ify" a video prompt. Do NOT use for long narrative films that need precise word-for-word dialogue sync, industrial shot lists meant to be executed verbatim, or functional/UI demos and step-by-step tutorials.
Processes videos to identify engaging moments, generate transcripts, and create highlight clips with artistic titles and custom cover images. Use when user needs to: extract highlights from long videos or livestreams, clip or cut best moments from videos, cut video highlights, process Bilibili/YouTube URLs or local video files, generate transcripts via Whisper, analyze content for engaging moments, create short-form clips with styled titles and covers, adjust cover text position and colors, find and export memorable scenes from recordings, burn subtitles into clips (with optional translation), guide clip selection with user intent, or identify speakers in multi-person conversations.
Edit any video into a captioned showcase — transcribe (any language, defaults to large-v3), present a transcript_review.txt for the user to fix mishears BEFORE rendering, then build a HyperFrames composition with liquid-glass caption pills, liquid blob background, liquid morph wipes, optional behind-subject text via background removal, and render the final video. Use whenever the user provides a video file and asks to edit it, caption it, add subtitles, fix existing captions, make a reel/promo/captioned tutorial, or "do the same" pattern as a prior captioned video. Supports English, Hebrew, and any Whisper-supported language. **Renders both 16:9 (YouTube / horizontal) and 9:16 (TikTok / Instagram Reels / YouTube Shorts) from the SAME 16:9 source** — vertical mode uses a centered footage strip with a blurred backdrop + liquid blobs and a vertical-tuned caption pill, no need to re-shoot. THE PIPELINE PAUSES FOR USER APPROVAL on the transcript before final render — this is the support mechanism for getting captions perfect (especially Hebrew). Pairs with hyperframes, hyperframes-cli, hyperframes-registry, and yuv-design-system skills.
Provides a vlog editing workflow: analyze videos with analyze_video.py (OpenVINO GenAI), generate a storyboard script (storyboard.json), and compose the final video. Use when the user provides a video directory and asks to make a vlog, generate a storyboard or editing script, or mentions vlog editing, video editing script, video storyboard, or editing script, or needs to render a final cut from storyboard.json.
Voice calibration before writing: clarify “which real me is hidden in this text” before rewriting. By default, outputs a voice anchor card (which can be passed directly to plainspoken-writing / editor-revisor as a protected zone); deep mode outputs a complete voice map and follow-up questions. Does not rewrite. Triggers: “dissolve the AI taste,” “real voice,” “find my voice again,” “doesn’t sound like me,” “root of the AI tone,” “clarify my voice,” “who am I behind this passage,” “what do I want readers to feel,” “preserve my edge,” “preserve my uncertainty,” “who is this passage really meant for,” “it feels like I’m not in the writing.” EN: “dissolve the AI taste,” “find my real voice,” “what did I actually want to say here,” “get the real me back in the text,” “AI smoothed away my edge,” “doesn't sound like me”. Not for: direct rewriting (→ plainspoken-writing / editor-revisor); content review (→ chinese-write-checker); title, layout, or punctuation polishing (→ packaging-workshop).
On an Apple Silicon Mac with a matching version of Jianying installed, generate editable drafts headlessly, modify an independent copy of an existing multi-track draft, or export native MP4 when explicitly requested. The Jianying Headless core project must be checked out separately; transcription alone does not use this Skill.
H3 prompt writing for stylized 3D animated shorts: story beats, character/scene locks, a six-column shot table and per-second instructions, producing shot-by-shot prompts that can be fed to MiniMax H3. Use for narrative animation storyboard prompts. Not for single-image retouching or photorealistic live-action scripts.
Research supported advertising promises from a specific buyer's context, current product facts, customer language and observed ads. Turn them into static, video or copy ideas; match video ideas to actual recipes and save a product-scoped angle bank for write-video-ad-script. Product-led discovery can use verified facts without reviews. Reference persistence and reach are observations, not proof of performance.
Analyze text for AI-style signals, false-positive risk, and directions for naturalizing rewrites. Supports text or URL input. Trigger words: AI-flavored / de-AI-flavor / make it read like a human wrote it / does this read like AI / Zhuque detection. Analyze text for AI-style signals, humanize wording, assess detector risk. Not for bypassing detectors.
AITuber API skill for AI video creation. Generate videos with AI narration, visuals, and captions for YouTube Shorts, TikTok, Instagram Reels, and long-form content. Supports AI-generated images, video clips, stock footage, and viral templates like skeleton and character styles. Handles the full pipeline: pick a voice, generate a video, poll for completion, export to MP4 and download, or publish to connected YouTube, TikTok, Instagram, Facebook, Threads, and X channels. TRIGGER when the user wants to: create an AI video, generate a video from a script or idea, list or browse AI voices, export a video to MP4, download a rendered video, list connected channels, publish or schedule a video to social media, check their AITuber subscription or credit balance, or automate video creation. DO NOT TRIGGER for: editing existing video files, uploading user-provided video footage, live streaming, video transcription or captioning of external files, image generation without video context, or anything unrelated to the AITuber platform.
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.