Image

Write or refine Midjourney image-generation prompts from an idea, visual brief, or reference image. Use when the user asks for a Midjourney or MJ prompt, including Chinese requests such as “写 Midjourney 提示词” or “优化 MJ 提示词”.

按主题完成资料查证、知识文章写作、Image 配图和 Obsidian 入库；也用于制作讲清项目用途、产出形式、安装和使用的图文 GitHub README。用户要图文学习笔记、一步到位学懂主题、品牌封面加图文项目介绍或修改上述内容时使用。简短知识问答不启动整套流程。

Consistent product, brand and marketplace-listing pictures with FLUX 3 Image Klein (Black Forest Labs, open weights) on a Google Colab GPU that is started for the job and stopped right after, via the fluxx CLI. Use when someone wants product shots, catalog or listing images, lifestyle scenes, social posts or ad visuals that look like the same brand every time - the same product, the same light, the same feel - without paying per image to an image API. Also makes marketplace listing packs for Amazon, Etsy, eBay and Shopify: a pure-white main image to Amazon's rules, the standard shot set per site, and drafted titles, bullets and tags within each site's limits. Each business gets a brand profile (what they sell, their look, what to avoid, reference photos of the product) that learns from their good/bad feedback. Renders several candidates per shot and an on-GPU judge keeps the best; outputs a high-definition PNG (~4 MP) and a medium JPEG (1024 px) per shot; keeps the GPU warm a few minutes for follow-ups. Al...

Turn any subject into a ready-to-post animated LinkedIn GIF in the popular infographic style, with icon nodes connected by flowing dotted lines (like "RAG vs CAG" or "How X works" posts). The user only gives a topic; Claude asks a few quick personalization questions (name, LinkedIn link, profile photo, light or dark background, language, color palette), then researches the subject, designs the flow, renders a looping 1080x1350 GIF with a Like / Comment / Repost / Follow call to action, and writes the caption. Supports English, French, Arabic (right-to-left) and other Latin-script languages. Use this skill whenever the user asks for an animated diagram, animated infographic, moving-lines GIF, LinkedIn visual, pipeline/workflow/process animation, "X vs Y" comparison graphic, or says something like "make a LinkedIn post about [topic]" with a visual, even if they don't say GIF or animation.

Design original logos, wordmarks, monograms and brand marks as clean, scalable SVG. Use whenever the user asks for a logo, brand mark, emblem, icon for a business or product, favicon, wordmark, monogram, or a visual identity starter kit — including requests to redesign, simplify, or produce variations of an existing logo they own. Produces several distinct concepts, refines the chosen one, and delivers SVG files plus PNG previews at multiple sizes on light and dark backgrounds.

"Context engineering is the delicate art and science of filling the context window with just the right information for the next step." — Andrej Karpathy. A frontier, first-principles handbook inspired by Karpathy and 3Blue1Brown for moving beyond prompt engineering to the wider discipline of context design, orchestration, and optimization.

生成一套风格统一的功能图标，可以交付悬停时会形变的动态版（SVG + JS），也可以导出静态 SVG 文件，沿用 Lumno 书签文件夹图标的构造：后片、内容、前片分层，外描边，所有颜色从一个主色推导，两组关键帧的悬停动效，可切换色板。用户要做图标、功能图标、动态图标、静态 SVG 图标、icon set、带 hover 动效的图标，或要往这套图标里加新图标时使用。

Find which LEGO or Duplo sets a mixed pile of pieces came from, using photos of the distinctive pieces, then build a tick-off checklist page and list the missing parts per set. Use when someone wants to sort mixed or second-hand LEGO back into its sets, identify sets from pieces or figures, or check what a set is missing.

SVG asset pipeline — Satori (HTML/JSX + CSS → SVG, then PNG via resvg or sharp) for title cards, thumbnails, OG images, stills; and SVGO v4 (optimize/minify SVGs, CLI + config) without breaking SVGs that will be animated. Use when generating images from markup, making video thumbnails/stills, or cleaning SVGs before animating them with GSAP / Anime.js / Remotion.

Explain a deep-learning concept with a beautiful, intuition-first static TikZ/pgfplots figure (compiled to PDF and PNG) plus a short walkthrough of how to read it, aimed at intermediate deep-learning students. Made for slides and decks (PowerPoint, Keynote, Beamer), papers, theses, handouts and lecture-note PDFs. Uses a consistent visual language (colour = role, Canziani NYU-DL book style) and standard DL notation. Use this whenever someone asks to visualize, illustrate, diagram, draw, sketch or "show me" any ML/DL idea (attention, convolution, backprop, normalization, residuals, RNN/LSTM, VAEs, diffusion, GANs, contrastive learning, optimizers, loss landscapes, activations, embeddings, energy-based models...), asks for a TikZ or LaTeX figure of a neural network or architecture, wants a figure for lecture notes, slides, homework, a thesis or a paper, or says "help me understand X visually", even if they never say "TikZ". For a figure the reader should drag, slide or hover, use tikz-live; for a whole illustrat

Produces finished social images with any image model (Google Gemini image, "Nano Banana", as the worked example) - feed statics, carousel slides, short-form video covers, YouTube and article thumbnails, local listing cards and animated statics. Bakes the headline into the scene instead of pasting text on top, keeps text inside the profile-grid safe zone, locks the likeness of real people to their reference photos, hits exact platform sizes, writes alt text and re-judges every re-roll on the filed file. Use when someone says "make an image for this post", "create a thumbnail", "make a cover from my video", "design a carousel", "animate this static", "the image model won't do 4:5", "my headline is getting cut off in the grid", or when a content pack needs its visuals.

Turns a 3D model (GLB/glTF/OBJ/STL/PLY/OFF) or any 2D image (photo, render, drawing, logo) into a technical blueprint sheet or desktop wallpaper - blue/graphite/paper style with hidden-line views, dimension lines, numbered callouts, title block, description, spec table and design remarks. Use this whenever the user asks for a blueprint, technical drawing, "tavola tecnica", "progetto tecnico", schematic, engineering-style sheet, cutaway-style poster or technical wallpaper of ANY object, vehicle, creature, character, building, product or image - even if they only say "make it look like a technical drawing" or "sfondo stile blueprint". Also use it when they upload a .glb/.gltf/.obj/.stl and want it as a drawing, or an image and want it as line art with dimensions and callouts.

Add film grain (noise) to CSS gradients and backgrounds so they look printed instead of plastic, and remove colour banding. Works on one background or a whole site or app, where it scans the codebase, lists every background and lets the person pick. It measures each background, picks the blend and amount itself, shows a before/after on the real page, and writes CSS that adapts to the visitor's screen (HDR/OLED, high contrast, print). Use when someone says a background or gradient looks cheap, flat, plastic, too smooth, digital or not premium, shows stripes or banding, asks for grain, noise, texture or a film or poster look, asks to check their site's gradients, or mentions the feTurbulence noise trick.

Turn a supplied photograph into a refined 3:4 split poster pairing the photo with an isometric miniature diorama, preserving recognizable subjects, spatial relationships, and authentic materials across photographic scenes. Use for Photo to Isometric Diorama Poster requests and photo-based miniature scene posters.

设计"海报风格"的演讲/展示类单页 HTML，并基于用户投喂的图文设计图沉淀可复用配色与构图规则。使用时机：(1) 用户投喂一张或多张图文设计参考图，要求分析配色、构图、字体、装饰并沉淀为可复用设计规则；(2) 用户给出需求（主题内容、情绪调性、指定页面比例），要求生成海报风格演讲页、金句页、观点页、图文排版页等单页 HTML，比例为任意值（16:9、9:16、4:3、1:1、自定义等）。生成 HTML 以本技能自带的 baseline.md 与 checklist.md 为准；环境若装有 html skill 可额外读取其设计规范。

基于 AI 智能分析的文档配图生成工具。无需依赖特定格式，自动理解内容并生成专业配图。

Make or fix pixel art - sprites, sprite sheets, tiles and tilesets, icons, game effects such as fire, smoke, water or sparks, and short frame animations - and save them as PNG, GIF, APNG or frames for a video. Use when someone asks to draw, animate, fix or export pixel art or game sprites. Pixelbolt JSON recipes, local, no account, no API key.

Plan, design, and export seamless 4:5 Instagram carousels from user-provided photos and video. Use for swipe-through photo spreads, framed photo journals, and mixed photo/video carousel posts when the user wants to preserve the original framing.

Generate article or newsletter thumbnail candidates using the Gemini API from inside Claude Code. Claude reads article copy, proposes composition concepts, writes image generation prompts incorporating brand specs, calls Gemini to generate the images, evaluates the results via computer vision, and returns ranked candidates with rationale. Use when asked to create thumbnails, generate cover images, or produce visual candidates for an article or newsletter.

为 X/Twitter 主页生成与左下角真实圆形头像自然连接的 3:1 横幅；通过截图反推头像位置，只改横幅不改头像或界面，并进行预览、偏移校准。适用于生成个人主页 header、banner、avatar-overlap 背景与融合光效设计。

Use when a user wants a graduation or college-admission destination map from images, TXT, Excel/CSV, PDF/Word, chat messages, web tables, or mixed rosters, or wants to correct an existing class map. Not for tracking people's live locations.

Write a structured JSON image prompt in English for Google NanoBanana Pro or NanoBanana 2. Trigger when the user asks for an image prompt, a Nano Banana prompt, or a visual/photo idea (also Turkish requests like "görsel promptu yaz", "prompt hazırla", "nano banana").

Finds open-licence illustrations and stock photos, recolors art to the tokens and wires it in. Also adds icons drawn like the few a product already has, from open sets, with their licences.

Two entry points, one standard of output: a watertight STL, a parametric script