Image

Build a live, physically driven 3D simulator of a real machine (engine, robot, linkage, gearbox, landing gear, turbine, excavator, watch, any mechanism) from its real published dimensions, with a dashboard whose every gauge is computed from the same simulation state, plus a clipping-plane cutaway. Use when the user asks to build, simulate, rebuild, animate, tear down, cut away or "show how it works" for a machine, engine, vehicle, robot or mechanism, especially when they name a real model or att

Make animated Axolotl Launcher axolotl stickers (美西螈表情包) with the AxoStickers engine — a voxel axolotl in water under big pink 3D Chinese text, 400×400 looping GIF ≤ 1 MB whose first frame is the cover. Use whenever someone asks for new axolotl / 美西螈 stickers, 表情包, 动图, a sticker for a phrase, a batch of stickers, a change to an existing sticker (text, motion, face, prop, colour), or the water-tornado sticker — even if they only give the phrase.

Full workflow for turning a user-supplied PBR texture set (zips of maps, scan or procedural) into the standard deliverables - ball + plane stills, .blend scenes, optional GLB, turntable, texture sheet - with a coordinator and specialist subagents. Use for any "build the scene as usual / new texture project / assemble the maps" request.

Call Zettos Photos APIs for photo search, timeline buckets, user albums, similar-photo tasks, photo map, and AI recognized albums.

Batch-prepare real project photos for ads and social posts. Corrects white balance, levels, converging verticals and vibrance, exports every platform size (Google Ads, Business Profile, LinkedIn) from one master with metadata stripped, builds before/after review sheets, and renders branded cards. Deterministic script; the model only reviews reports and writes small JSON overrides.

Create Instagram carousel posts (1080×1350) for a tech/AI creator page in one of five house styles — Editorial, Sticker, Dark tech-news, Minimal notebook, Bold pastel — plus an English and Hinglish caption. Use when the user gives a topic, script or link and asks for a carousel, slides or an Instagram post.

Renders architectural sketches, plans and 3D exports as photoreal DSLR images via Gemini, 1K first, with a sheet per pass. Use when a drawing should become a photoreal image, or to revise or 4K one.

Make a Fringe Lines image carousel (Instagram/LinkedIn, 1080x1350 PNG slides + PDF) — photo slides with big condensed caps headlines and accent words, large-type text-only slides, chart/screenshot slides, explainer-graphic slides and a follow CTA. Use when asked for a carousel, slide post, or "swipe" post from an episode or a topic.

Creates complete Instagram carousels: slide-by-slide script, caption, hashtags and ready-to-post 1080×1350 PNG slides (rendered from HTML). Covers five archetypes (educational/listicle, portfolio/photo dump, continuous/panoramic, series template and cinematic with photos and expressive type) and slide-to-slide continuity techniques (continuous background, elements crossing the seam). Use whenever the user asks for a carousel, a multi-slide post, 'Instagram slides', a photo dump, a series of posters (matches, events, products), or wants to turn a topic, article or set of photos into a swipeable post, even without saying 'carousel'.

Route greybox creation, gameplay revision and art work through the shared Python CLI in WorkBuddy, including with DeepSeek models without image generation tools.

Generates optimized prompts for AI tools. Activates only when the user explicitly asks to write, fix, improve, or adapt a prompt for a specific AI tool (LLM, Cursor, Midjourney, image AI, video AI, coding agents, etc.). Does not activate for general conversation, coding tasks, document writing, or other non-prompt-engineering work.

Rebuild slide images, screenshots, scanned/image-based PPT or PPTX files, and PDF decks into object-level editable PowerPoint (.pptx) — real text boxes, native shapes, native tables, and icons/illustrations re-traced as SVG — preserving page order and speaker notes. Runs fully inside Claude Code with no image-generation API. Use this whenever the user wants to "make this slide editable", convert a slide screenshot/PNG/JPG/PDF/flattened PPTX into a PPT they can edit, recover text from an image-only deck, or reconstruct an infographic as PowerPoint objects, even if they just say "turn this image into a PPT" or "I can't edit the text in this deck". Not for authoring new presentations from an outline, article, or idea.

Write, edit, debug, compile and package Typst (.typ) documents — papers, reports, theses, CVs, letters, slides, posters, math-heavy notes, data-driven docs. Covers markup/math/code syntax, set/show rules, layout, tables, figures, bibliographies, introspection (counters/state/query), data loading, PDF/HTML/PNG/SVG export, PDF/A and PDF/UA accessibility, CLI (compile/watch/eval/init/fonts), templates, Typst Universe packages, and LaTeX/Markdown conversion. Use whenever the task involves Typst, typst.toml, .typ files, "typst compile", or turning content into a typeset document and Typst is an option.

Convierte transcripciones de reuniones de refinamiento, sprint planning o seguimiento en tareas para Azure Boards (título, descripción y exactamente un criterio de aceptación, responsable y esfuerzo inferido) y en documentación estratégica por proyecto (objetivo del sprint, hoja de ruta, acuerdos y registro de reconciliación). Reconstruye el estado final de las decisiones aunque se corrijan más adelante en la conversación o en reuniones posteriores, separa proyectos y distingue compromisos de ideas. Úsala cuando el usuario entregue una transcripción o notas de reunión, pida tareas para Azure Boards o DevOps, quiera actualizar el estado de un proyecto con una nueva reunión o necesite el objetivo del sprint y la hoja de ruta.

Design high-click YouTube thumbnails and vertical covers (Shorts, Reels, TikTok) with whatever image provider the user has, compositing headline text locally so it is always crisp. Plans the concept, generates 2–3 variants, keeps faces and logos faithful via reference images, and checks readability at small size. Use when the user asks for a YouTube thumbnail, video thumbnail, Shorts/Reels cover, podcast episode cover, or "make my thumbnail pop". Not for general images (creator-image) or channel branding (creator-brandkit).

Criar e refinar posters, matchdays, vitórias, anúncios de jogadores e campanhas de esports com fotomontagem editorial, recortes fotográficos, tipografia expressiva e acabamento texturado. Usar para peças gráficas de equipas e criadores, incluindo a direção visual dos posters do Light. Não aplicar a interfaces, landing pages ou edição de montagens de gameplay.

Make 1200×630 link-preview images (Open Graph / social cards, PNG) for blog posts, docs or newsletters in one editorial look - a serif headline set as real text, a byline and a generated block composition - from Markdown front matter or a list, and wire them into the site's meta tags. Use when someone wants OG images, social cards, link previews or Twitter/X cards for their posts.

MCP server for OHDSI OMOP medical vocabularies. Search, map, and navigate 11M+ concepts (SNOMED, ICD-10, RxNorm, LOINC) from any AI agent.

Turn one or more user-provided reference photos into a complete publishable WeChat sticker album: plan and generate exactly 20 separate identity-consistent Q-version stickers with varied Chinese reactions, preserve untouched originals and high-resolution generation sources, create current-spec WeChat banner/cover/chat-icon assets and submission metadata, validate every deliverable, and build upload-only plus full-archive packages. Use when the user wants to create, rebuild, package, or prepare a WeChat sticker pack from reference images.

生成 Bornforthis 个人 IP 风格的中文正文配图。用于用户要求为中文文章、帖子、博客、Notion 文档、知识讲解、工作流、方法论、流程、结构、状态、隐喻或观点制作“Bornforthis”“个人 IP”“蜡笔手绘”“正文配图”“文章插图”“配图建议”“shot list”“去标题”或“改图”时；默认使用 Bornforthis 青年男性角色、纯白背景、蜡笔或马克笔斜线纹理、深蓝粗描边、少量蓝红中文批注和温暖而略荒诞的绘本风格。

使用千问 Qwen-Image-3.0 模型生成或编辑图片。V3 核心升级：三段式令牌链条，新增提示词优化审核环节的代码级门禁--拆成 optimize（按官方范本审核优化提示词、登记「优化前后对比+改动说明」防敷衍、返回令牌A，绝不调 API）、prepare（消费令牌A、原样呈现对比供用户确认、返回令牌B、销毁令牌A，绝不调 API）与 generate（检查令牌B通过后才调 API、成功后销毁令牌B）三阶段。令牌接力少一张门票门就不开，模型无法从原始需求直接跳到生成。支持文生图（T2I）与图生图/图像编辑（I2I，1-3张参考图）。默认标准版 qwen-image-3.0，可选 Pro 版 qwen-image-3.0-pro。工作流：意图识别与澄清 -> optimize 审核优化 -> 回显确认 -> prepare 确认 -> generate 生成与迭代。Use when the user wants to generate, create, draw, or edit images, illustrations, posters, logos, wallpapers, or any visual artwork.

Build and run a local production Docker image for a Next.js app safely. Use this skill when the user wants to build a Docker image, run a container with an env file, check whether an image version already exists locally, or deploy a Next.js app locally with Docker without hard-coding the project name.

Add SEO tags to Lynx layouts and manage SEO settings. Use when writing a layout's head, turning SEO fields on for a template, or checking a page's title, description, share image, canonical link or robots rule.

Лист персонажа (character sheet) по бесплатному гайду GenHQ для Higgsfield: одно фото лица → образ с 5 ракурсами (спереди и сбоку без головы, сзади, крупный портрет, сверху), линейкой роста и 8 макро-деталями ткани и аксессуаров; лист потом идёт референсом в новые кадры (GPT Image 2.5, Nano Banana 2.1) и в видео с переносом движения (Higgsfield Genjutsu Motion transfer, Seedance 2.5). Используй для ИИ-инфлюенсеров, героев рилсов и рекламы, UGC, бренд-персонажей, переодевания героя, когда лицо или одежда «плывут» между генерациями.