Skills
97,222 packaged capabilities and playbooks for agents. Install any of them with one command.
Creative & media
Produce and operate Amazon Nova Reel video-generation jobs through Amazon Bedrock. Use for Nova Reel text-to-video, image-conditioned animation, automated or manual multi-shot storyboards, async S3 delivery, cost and approval gates, production continuity, output custody, and AWS-specific safety, privacy, IAM, and provenance decisions.
Use Microsoft Azure Speech in Foundry Tools for media-production speech workflows: speech-to-text, fast and batch transcription, diarization, captions/subtitles, real-time transcription, text-to-speech neural and HD voices, SSML, custom/personal voices, text-to-speech avatars, speech/video translation, localization, quotas, regions, privacy, consent, rights, and production QA.
Create beautiful visual art in .png and .pdf documents using design philosophy. You should use this skill when the user asks to create a poster, piece of art, design, or other static piece. Create original visual designs, never copying existing artists' work to avoid copyright violations.
Turn photos, videos, and Live Photos into a black-and-white to color reveal clip (单彩渐变 + 动感模糊 + 丁达尔旋焦). The frame starts monochrome and spun, color sweeps in from the left under a horizontal motion blur, and the spin settles from the center outward into a softly glowing full-color frame. Use when users ask for 上色效果, 黑白变彩色, 丁达尔旋焦, 旋焦上色, 剪映上色特效, color reveal, colorize transition, or want this effect on images, iPhone Live Photos (HEIC/JPG + MOV), Android motion photos, or video clips. Outputs H.264 MP4.
Use for Deepgram speech and voice production workflows: speech-to-text transcription, live captions, diarization, audio intelligence, Aura text-to-speech, Flux and Voice Agent live audio, model selection, cost/limits/privacy checks, artifact custody, and production QA.
Create high-quality animated explainer visuals for essays and blog posts. Use when the user wants to visualize concepts, processes, data, or ideas with interactive web animations. Triggers on requests like "create a visual for", "animate this concept", "make an explainer", "visualize this idea", "diagram this process", "show this data", or when essay content would benefit from visual explanation. Handles abstract concepts (mental models, frameworks), technical processes (algorithms, systems), and data visualization (trends, comparisons). Outputs self-contained HTML/CSS/JS that embeds directly in web content.
Produce HeyGen avatar videos with Direct Video, Video Agent, Digital Twin, Avatar Realtime, and related avatar/voice/asset APIs. Use for provider routing, avatar and voice selection, script-to-video workflows, lip-sync from audio, backgrounds/scenes, callbacks and polling, consent and rights checks, localization handoffs, artifact custody, QA, and troubleshooting HeyGen avatar-video productions.
[L1 Voice authenticity] AI-tone detection and removal — only deals with "does it sound like a real person." Recognizes 19 AI-tone patterns and selects a strategy by severity. By default it only diagnoses and does not rewrite; rewriting requires user confirmation. Three rewriting modes: automatic replacement, question-driven, and hybrid. Can accept a voice-dissolver voice list as a protected zone. Triggers: AI tone, sounds machine-written, too formal, reads as fake, de-AI, too much AI feel, sounds like ChatGPT wrote it, too templated, tone too standard, lacks personality, AI accent, reads without warmth, sentence patterns too neat, concluding remarks, translationese, doesn't sound human, formulaic. Not for: content quality is insufficient but there is no AI tone (→ L0 chinese-write-checker), structural redundancy/wordiness (→ L2 editor-revisor), can't find your own voice (→ voice-dissolver), layout/punctuation/headings (→ L3 packaging-workshop).
Plan, generate, edit, reference, and quality-control still images with Kling AI's hosted IMAGE surfaces or Kuaishou's open Kolors checkpoints. Use for Kling IMAGE 3.0/3.0 Omni/O1/2.1 web, official Kling CLI/MCP or Open Platform integration, and local Kolors text-to-image, image-to-image, IP-Adapter, ControlNet, inpainting, or LoRA work. Do not use for Kling video, avatars, lip-sync, unofficial gateway APIs, or for assuming that current hosted Kling Image models are identical to the 2024 open Kolors weights.
Generate images and videos through the currently installed Media.io CLI. Use for text-to-image, image-to-image, text-to-video, image-to-video, reference-to-video and published workflows. Select the model from the bundled static catalog (`references/model-catalog.md`), which is generated from the production registry; only fall back to `mediaio model list` on the triggers listed in the discovery guardrail. Copy every `job_type` byte for byte — some contain a literal space, and display names often do not match the identifier. Effects may be used only when their parameters have been independently verified because the current CLI exposes `effect list` but not `effect get`. Always confirm the parameter schema with `model get` before submission. Submitting spends the user's credits, but the CLI stays quiet about the amount unless asked; surface the cost with `--show-credit`, and get an explicit approval first, only when the user is cost-sensitive or has raised credits, price or balance. When the balance runs short, or the user wants their task history or their uploaded files, hand over the product-page link printed by `mediaio link get` — never write or edit a media.io URL yourself. Use the human-readable discovery output, upload local files before generation, submit with `generate create`, wait with the separate `generate wait` command, and retrieve result files with `generate download` instead of reproducing signed result URLs. On hosts that sandbox local command networking, the first networked `mediaio` or `curl` Shell/Bash tool call must request host approval before process launch. Never probe network availability by running it in the default sandbox.
Writing coach that extracts educational content from your daily experiences and turns it into publish-ready newsletter drafts. Use when brainstorming newsletter ideas, writing content for The Little Blue Report, or when you want help turning experiences into educational articles.
Arrange opening positions for every scene in an AI short drama: first identify objects in the scene image that can be sat or stood on (projected onto a top-down plane), then assign each character in every scene to a side of an object, sitting or standing, facing direction, and whether extras are present. The program writes an opening blocking sentence for each scene; copy it into the first segment of that scene's storyboard. By default, deliver a draggable HTML tool report for adjusting positions; it can export staging.json in the page. JSON is also saved as working data, and Markdown can be exported as needed. All 8 quality gates are checked deterministically by script (number of identified objects, objects present in the image, exact cast count, every present character assigned to an identified object, no overlapping characters on the same side, specified extra positions, and more). Pre-fills the skeleton from novel-script's script.json and novel-art's art.json. Zero dependencies, zero API keys. Use when asked about blocking, staging, who sits where, seating arrangements, or staging for an AI short drama.
Select, run, and govern NVIDIA Cosmos world-video generation across Cosmos 3 Generator, Predict2.5, Transfer2.5, downloadable checkpoints, self-hosted NIMs, and hosted preview surfaces. Use for text/image/video-to-world, controlled world transfer, multiview or action-conditioned physical-AI video, local checkpoint pinning, hardware planning, inference approval, and output validation; do not use for Cosmos Reason/Embed-only tasks.
Toolkit for creating animated GIFs optimized for Slack, with validators for size constraints and composable animation primitives. This skill applies when users request animated GIFs or emoji animations for Slack from descriptions like "make me a GIF for Slack of X doing Y".
Cut a long webinar, demo, or recorded call into several short highlight clips optimized for social feeds. The core judgment call is finding which moments actually deserve to become clips: a hook, a concrete outcome, a striking stat, a sharp Q&A answer, decided from the real transcript rather than guesses about wall-clock position, before any cutting happens. Use when the user says "cut this hour-long webinar into five short clips for LinkedIn", "turn this demo recording into social clips", "pull the highlights out of this call", "make some short clips from this long video", or "give me a few clippable moments from this recording".
A comprehensive suite of protocols, meta-prompts, and orchestration tools designed to streamline software development workflows, project management, and team collaboration. Includes the VibeCode Protocol Suite with web interface, user manuals, and reusable templates for efficient development processes
Comprehensive patterns for AI-powered audio generation including text-to-music, voice synthesis, text-to-speech, sound effects, and audio manipulation using MusicGen, Bark, ElevenLabs, and more. Use when "music generation, text to music, AI music, voice cloning, text to speech, TTS API, ElevenLabs, MusicGen, Bark, audio synthesis, sound effects generation, voice synthesis, AudioCraft, " mentioned.
Produce and review production image-generation and image-editing workflows with Amazon Nova Canvas on Amazon Bedrock, including native InvokeModel payloads, safe authentication, validation, retries, cost controls, provenance, and lifecycle migration checks. Use when a task names Nova Canvas, amazon.nova-canvas-v1:0, Bedrock image generation, Canvas inpainting/outpainting/background removal, image conditioning, color guidance, image variation, virtual try-on, or Canvas fine-tuning.
Use Amazon Transcribe for AWS-based speech-to-text production: batch S3 transcription, real-time streaming, captions/subtitles, diarization, channel identification, custom vocabularies, vocabulary filters, language identification, PII/PHI handling, toxicity detection, Call Analytics, Medical, and secure S3/IAM/KMS workflows.
Produce high-fidelity design mockups using HTML as the medium (landing pages, slides, interactive prototypes, animations, design systems, infographics, mobile mockups, etc.). Act as a professional designer rather than a generalist frontend engineer, emphasizing asking before doing, referencing existing design systems, offering multiple variants, and avoiding AI-style clichés. Triggers: /axi-front-design, "做个落地页", "设计一张海报/封面", "做张幻灯片/PPT", "做个交互原型", "做个动画", "design a landing page", "make a deck", "prototype this UI", "design system", "hi-fi mockup". Not applicable: plain text content/logic code, needs real image generation (use an image generation model instead), needs native PPTX/Figma files (HTML preferred).
Build and operate rights-aware Bria FIBO and FIBO Lite image-generation workflows with structured prompts, reference images, asynchronous status handling, webhooks, cost gates, and safe artifact downloads. Use when a user asks for Bria/FIBO generation, refinement, inspiration, reproducibility, hosted API integration, or a licensed-data image workflow; do not use for FIBO Edit, video, generic image editing, or third-party FIBO gateways.
Use when users want creative inspiration, storyboards or replication prompts for viral videos, analysis of visual elements and keyframes in trending content, or insight into the creative approach behind breakout videos in an industry. Manually trigger with: /inspiration
Generate music using Volcengine Imagination API. Supports vocal songs, instrumental BGM, and lyrics generation. Use when the user wants to create songs, background music, soundtracks, write lyrics, or mentions "music generation", "BGM", or "songwriting".
This skill should be used when the user asks to "create a character", "update a character", "add a character", "build a family tree", "character relationships", "character timeline", "character arc", "character profile", or needs to manage characters in a story project.
Use D-ID to plan, generate, stream, localize, and QA avatar/talking-head videos, including V2 Photo Avatar Talks, V3 Pro/Instant Avatar Clips, V4 Expressive Avatar Scenes, D-ID Agents, voices, consent, lifecycle, safety, and artifact custody.
Use this skill for local FFmpeg/FFprobe media inspection, remuxing, transcoding, filtering, evidence-bounded video review, transcript-assisted editorial plans, edit decision lists, podcast/audio cleanup, rendering, and output acceptance. It emphasizes explicit stream selection, source preservation, build-aware commands, bounded evidence, and verified new outputs. Do not use it for libav API programming, opaque whole-video understanding, automatic publishing, rights clearance, DRM circumvention, professional broadcast/color certification, or HyperFrames-authored compositions; route those tasks to their owning capabilities.
Use this skill when a media-production agent needs Google Cloud speech and voice services for transcription, captions/subtitles, long-form audio analysis, live caption planning, Text-to-Speech narration, Chirp 3 / Gemini-TTS voice selection, consented custom voice workflows, dubbing/localization planning, pricing/quotas/region checks, or speech-related safety and data-governance decisions.
Call the GPT imagegen skill to generate an "image-format PPT document" — one slide image per page with high information density, complex layout, and a consistent style (default: tech business/high-density infographic; also supports minimalist/cartoon/ink-wash/tech and any other style), and composite them into a .pptx with one full-page image per slide. Generates image-based slide decks (one rendered image per slide) from a topic or content. Use when the user wants to "make a PPT / generate an image-based PPT / create AI-generated slides / turn content into PPT images". This skill **only produces image-based PPTs** (one .png per page + a composited .pptx); if an editable pptx is also needed, then use GordenImage2PPTX.
Use Luma AI's first-party Photon and Photon Flash image API for text-to-image, image/style/character references, image modification, asynchronous generation, secure artifact handling, production prompting, cost control, and policy-aware delivery. Trigger for Luma Photon API integration or production work; do not use for Ray video, UNI image models, the Luma consumer app, or third-party Photon gateways.
Produce video with Moonvalley's Marey model family (Marey Realism v1.5) — a filmmaker-oriented, 1080p/24fps generative video model marketed as trained exclusively on licensed data. Use this skill when a task asks for Marey or Moonvalley specifically, when a brief demands brand-safe / legally reviewed AI video for commercial or studio work, or when the request needs Marey's director-style controls (camera trajectory, motion transfer, pose transfer, keyframing, reference conditioning). It covers what "commercially safe" really means versus the actual terms of service, access routes (Moonvalley web / Voyager, the fal API, ComfyUI), pricing, capability limits, prompt and reference strategy, where Marey fits production versus where other models win, iteration, and quality review. Do not use it for audio generation, for photoreal talking-head dialogue, or as a generic "best video model" default.
Use NVIDIA Speech NIM microservices for speech production and voice workflows, including self-hosted ASR/STT, TTS, text translation, speech-to-speech pipeline design, Riva Python client integration, deployment planning, GPU/runtime sizing, licensing/privacy review, and production QA for transcription, captioning, voice agents, localization, and synthetic voice output.
Plan and scope production work with Odyssey's real-time interactive video world models (odyssey.ml — Odyssey-1/Odyssey-2/Starchild-1/Agora-1). Use when a request involves generating video that a user steers live with keyboard, text, speech, or controller input (streamed frames that respond in the moment), or when someone must decide between Odyssey-style interactive generation and offline clip generators (Sora/Veo/Kling) or downloadable 3D / game engines. Covers what is verified versus announced, current access routes, latency and coherence limits, feasible-now versus speculative use cases, pricing/access terms, and content/rights considerations. This is a fast-moving research-stage product; the skill teaches honest scoping, not a promise of production reliability.
Produce audio-first podcast episodes with generative tools — design the show and episode format (interview, narrative, news brief, two-host conversational), write scripts for the ear, decide between fully-synthetic and hybrid (recorded human + synthetic) production and meet the disclosure duty each triggers, cast and direct multi-voice TTS for consistency and chemistry, build episode structure (cold open, intro/outro, segments, ad slots), assemble and edit, clear music and SFX rights, hit podcast loudness and delivery standards, ship metadata/chapters/transcripts through RSS, and pass platform AI-content policy and QA before publishing. Use this skill whenever the deliverable is a podcast episode or a podcast show bible, or when an agent must make production, rights, disclosure, or delivery decisions for spoken-word audio. Not for music tracks, single-voice notification prompts, or video where picture leads (route audio-for-video to a video skill).
Give an existing video a narration makeover - rewrite the script in a new tone (tighter, warmer, more formal, more energetic), pick a fitting voice, regenerate the voiceover, and re-align the visuals to the new timing. Use when the user says "make the narration less robotic", "rewrite the voiceover to sound friendlier", "this video sounds too stiff, revoice it", "change the tone of this video", or "redo the narration in a different style".
Splits a full sales demo into a chaptered, async-reviewable format built around buyer decision points (Integrations, Security, Pricing tiers shown, the feature that solves their specific problem) rather than a generic feature-by-feature or course-style structure. Adds an opening "what's in this demo" card so a prospect can decide where to jump before watching, and makes every section label and recap stand on its own with zero live-rep context, since no one is present to narrate or answer questions. Use when the user says "make this demo watchable without me in the room", "split this demo so the prospect can jump to what matters", "turn this into an async-shareable demo", "let the champion forward this and have stakeholders skip to their part", or "make my demo self-navigable".
Get a video under a hard time limit by deciding what to cut and what to compress, while protecting a must-keep list and the core message. Use when the user says "get this under 2 minutes", "shorten this to 60 seconds", "it needs to fit in 90 seconds for the campaign", "cut this down but keep the demo", or gives a video plus a time budget it currently blows through.
Create vertical 9:16 video clips from full music videos for Instagram Reels, TikTok, and YouTube Shorts. Takes full-length music video and extracts key moments, re-frames for mobile, and optimizes for social platforms. Use when creating short-form social content from PsalMix music videos or any video content.
Turn a behavioral or compliance training topic - giving feedback, de-escalation, harassment policy, security awareness - into a story-driven animated scenario video: illustrated characters, setup, tension, response, a narrated takeaway after each beat, and a closing what-to-do checklist. Use when the user says "make a soft skills training video", "scenario-based training on X", "compliance training video", "turn this policy into a training video", or "behavioral training for managers".
Convert a PowerPoint standard operating procedure deck into a narrated training video plus a written job aid (a condensed, scannable quick-reference companion authored as the project's article). Preserves compliance-critical exact wording instead of paraphrasing it. Use when the user says "turn this SOP deck into a training video", "convert this PowerPoint procedure into a video and job aid", "make a training video from this SOP presentation", or hands over a procedure deck and asks for both a video and a reference sheet.
Produce two artifacts from one procedure in a single pass: a narrated SOP walkthrough video and a step-by-step written SOP with an annotated screenshot per step, with identical step numbering across both. Use when the user says "make an SOP video and doc", "I need this procedure as a video and a document", "create the SOP for this process", "standard operating procedure with screenshots", or "document this workflow for the rollout".
Dim everything on screen except the area being explained, so viewers can't look at the wrong thing while a step is narrated. Use when the user says "spotlight the important area", "dim the background", "focus attention on this part of the screen", "darken everything except the button", or "the screen is too busy - isolate what matters".
Turn numbers into a 30-60 second animated data story - one stat per scene with animated counters and bar reveals, building to a stacked payoff scene. Use when the user says "stats video", "infographic video", "animate these numbers", "turn this data/report into a video", "make a video from these survey results", "data story video", or pastes metrics, benchmarks, or ROI figures they want presented with motion and narration.
Change a video's narration voice - accent, gender, energy - without touching the script, offering a shortlist of three fitting voices first. Use when the user says "change the voice", "use a different narrator", "make it a British accent", "I want a warmer/more energetic voice", "swap the voiceover voice", or wants the same words spoken by someone else.
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.