Vox (Voxyz_ai) recommends landing-page-design from elayadesign/ai-design-skills: an Agent Skill that handles layout, copy, fonts and spacing so Claude Code or Codex can build a landing page after one clarifying Q&A pass.
Comunidade de Agent Skills
Explore publicações revisadas no X, quem as compartilha e os projetos mencionados.
Publicações por dia
62 publicações coletadasA contagem inclui apenas publicações sobre Skills coletadas e revisadas. Um dia vazio indica que nada foi coletado aqui. Datas em UTC.
Publicações recentes
62 publicações coletadas
Yiwei Ho (1weiho) releases cinematic-shot: an Agent Skill that directs product UI like a film—hero shots with steep angles, thin focal planes, slow camera drift and self-acting UI—for launch videos, teasers and cinematic stills without After Effects.
Jakub Krehel (jakubkrehel) releases teardown: an Agent Skill that turns any physical object or how-it-works topic into a technical illustration with editable parts, colors and annotations, exportable as SVG or PNG.
Sanem Avcil (Sanemavcil) recommends hallmark: an open-source design skill for Claude Code, Cursor and Codex that steers agents away from generic AI UIs—build/audit/redesign/study modes, themes, and slop-test checks before returning output.
Baoyu (dotey / JimLiu) ships the BaoCut Skill: an Agent Skill that, with the BaoCut app, drives an editable AI video timeline—transcribe, subtitle, translate, dub, cut and animate—for archive/documentary-style layout videos.
meng shao (shao__meng) recommends dimensional-drawing: an Agent Skill that turns UI components into Apple-accessory-style monochrome dimension/architecture spec sheets (line weights, slots, variable formulas) and ~10s stroke-motion films—without painting real copy.
半吊子程序猿大铭 (CoderDaMing) recommends hallmark: Together AI’s anti-“obviously AI” design skill for coding agents—48 visual themes set the look, pages follow visitor intent, then a real-browser audit checks common AI design tells; /hallmark redesign can remake an existing page while keeping content.
丰收 (ai_fengshou) recommends hallmark: an anti-"obviously AI" design skill for coding agents—48 material themes set the look, then a real browser audit catches empty boxes, fake buttons, and contrast failures so pages feel more human.
Alchain (花叔) released huashu-design 3.0: builds high-fidelity HTML prototypes, slides, and motion—always drafts three design directions first, then executes with optional expert review.
牙仔 updated handraw-style with FE-068 Tangshan Shadow Puppet: bilingual prompt packs that render carved negative-space, backlit leather texture, and classic cinnabar/charcoal Chinese folk aesthetics for design and cultural creative work.
Creator released educlopez/diorama (skills/diorama/SKILL.md): 用无依赖 SVG 画可动的等距小场景(toy/soft/mono/line/print/night 主题),另有 diorama-worlds 与 diorama-pixel 变体
Creator released lucasmarkes/skills (skills/specimen/SKILL.md): 用一句话命名物体,生成路径追踪的 3D 产品图标(视频或可嵌入元素)
Manorgw recommends jakubkrehel/skills: Agent Skills that teach UI craft—details, typography, OKLCH color, layout, accessibility, and UX copy review—so generated pages look intentional, not just functional.
周览资源 recommends cathrynlavery/diagram-design: generate professional flowcharts, architecture diagrams, and dozens of other chart types as HTML/SVG from natural language, with brand color/font matching.
Creator released a-lav-dsgnr/ux-writing (skills/ux-writing/SKILL.md): 按 UX writing 最佳实践撰写或改写产品界面文案(按钮、报错、空状态等)并给出多稿理由
Leon Lin ships TasteSkill v2.1 with stronger dashboard taste and small-detail polish so agents build cleaner websites and admin dashboards.
Creator released hamzaabduljabbar/autoConst-p6-revit-4d (SKILL.md): 把 P6/MS Project 进度与 Revit 模型自动关联成 4D,校验施工顺序并写回 Revit
Creator released ashishkaloge/map-design (skills/map-design/SKILL.md): 用真实地理数据生成自包含 HTML/SVG/PNG 地图(分区着色、路线、城市海报等),避免模型凭记忆画边界
gleaming releases Monument, a Skill that reconstructs named historical buildings in interactive 3D phase by phase on real plans, with dated sources and a hard line between fact and AI guesswork.
Majid Manzarpour points to blender-game-skills for turning concept art into Blender CLI/Python-driven, game-ready 3D faces and assets measured against references.