Creation Skills创作技能
Browse creation skills in the Orkas marketplace — install any of them in the Orkas desktop app.在 Orkas 市场浏览创作技能——可在 Orkas 桌面应用中安装。
- image-processDeterministically process local image files: resize, crop, convert format, compress, set quality, strip EXIF, batch-normalize, and preserve originals. Use for existing image files when no semantic AI generation/editing is needed.
- content-writerUnified long-form content writing, rewriting, and audit skill for research-backed articles, blogs, newsletters, tutorials, opinion pieces, outlines, citation organization, voice preservation, natural rewrites, and pre-publish fact/link checks.
- image-designPlan image generation and editing work: creative brief, structured prompts, aspect ratio, style, reference-image strategy, model/tool routing, edit workflow, and acceptance criteria. Use when the user needs an image plan, prompt, provider choice, reference consistency, or edit plan before execution.
- social-writerUnified social content skill for Chinese platform drafts, global single posts, existing-draft optimization, content calendars, and comment/complaint/crisis response drafts across Xiaohongshu, WeChat, Zhihu, Douyin/Kuaishou, Weibo, X, LinkedIn, Instagram, and TikTok.
- dreaminaExecute image or video generation through an already logged-in Jimeng / Dreamina CLI account, including credit checks, async submit/query, task history, text-to-image, image-to-image, text-to-video, image-to-video, multi-frame or multimodal video, and first/last-frame video workflows.
- image-reviewReview generated or edited images against the original request: prompt adherence, technical quality, composition, fitness for purpose, text, reference consistency, safety, and copyright risks, with regeneration or edit recommendations.
- nano-bananaExecute Nano Banana / Gemini / Imagen-style image generation or editing through available providers, with provider selection, API-key checks, reference-image upload consent, cost notes, output download, and failure fallback.
- social-publishUse a visible browser to place prepared content into a social platform editor, save a draft, or stop at the pre-publish confirmation page. Supports title, body, tags, topics, images, videos, cover, category, and similar fields. Never clicks final publish automatically.
- image-process对已有本地图像做确定性处理：改尺寸、裁剪、转格式、压缩、设置质量、去 EXIF、批量规范化，并保留原图；适合“把这些图转成 webp 并压缩”“裁成 1080x1080”“批量去掉图片元数据”；触发词：改尺寸、裁剪、转格式、webp、压缩、去 EXIF、批量处理、图片后处理
- content-writer写作、改写和审核长文内容的统一入口：支持研究型文章、博客、newsletter、教程、观点文、提纲、引用整理、作者声音保留、去 AI 味润色和发布前事实/链接核验；适合“写一篇有引用的文章”“改得更自然”“发布前检查引用”；触发词：写文章、长文写作、研究写作、文章提纲、引用、改写、去 AI 味、润色、事实核查、发布前检查
- image-design将图像生成或编辑需求转成创意 brief、结构化 prompt、画幅/风格/参考图方案、模型或工具选择、编辑步骤和验收标准；适合“帮我设计一张海报怎么生成”“给这组产品图写 prompt”“保持参考图风格做系列图”“这张图该生成还是编辑”；触发词：生图方案、图片设计、图像 prompt、参考图、同风格、角色一致、模型选择、编辑方案
- social-writer社媒内容创作的统一入口：支持中文平台成稿、国际平台单帖、已有稿优化、内容日历、评论/差评/危机回应草案；覆盖小红书、公众号、知乎、抖音/快手、微博、X、LinkedIn、Instagram、TikTok；适合“写小红书”“优化公众号”“排内容日历”“这条差评怎么回”；触发词：社媒写作、小红书文案、公众号文章、知乎回答、抖音脚本、X 帖子、LinkedIn post、内容日历、社媒优化、危机回应
- dreamina通过已登录的即梦 / Dreamina CLI 执行图片或视频生成、查询积分、提交/查询异步任务和查看任务历史，处理文生图、图生图、文生视频、图生视频、多帧/多模态视频和首尾帧视频；适合“用即梦生成图片/视频”“调用 Dreamina CLI”“查询即梦积分”“用多张图生成故事视频”；触发词：即梦、Jimeng、Dreamina、生图、生视频、文生图、文生视频、图生视频、多帧视频、多模态视频、Seedance、Seedream
- image-review评审生成图或编辑图是否符合原需求，检查 prompt 一致性、技术质量、构图、用途适配、文字、参考图一致性、安全和版权风险，并给出重生成或局部编辑建议；适合“帮我看看这张图能不能发”“这张图和 prompt 匹配吗”“哪里需要重做”；触发词：图片评审、质量分析、图像 review、提示词一致性、构图、技术瑕疵、重生成、改进建议
- nano-banana通过 Nano Banana / Gemini / Imagen 相关 provider 执行图像生成或编辑，处理 provider 选择、API key 检查、参考图上传确认、成本提示、输出下载和失败回退；适合“用 Nano Banana 生成一张图”“用 Gemini 图像模型编辑这张图”“对比 Atlas 和 Google 的生成成本”；触发词：Nano Banana、Gemini 图像、Imagen、Google 图像生成、Atlas Cloud、参考图编辑、图像 API
- social-publish使用可见浏览器把已准备好的内容填入社媒平台后台，保存草稿或停在发布前确认页；支持标题、正文、标签、话题、图片、视频、封面、分类等字段；禁止自动点击最终发布；适合“保存成公众号草稿”“填到小红书草稿”“放进 X 发帖框但不要发布”；触发词：保存草稿、发稿、社媒发布、准备发布、公众号草稿、小红书草稿、发帖框、不要发布、agent-browser