threerocks/sweety-skills

sweety-image-art-direction

仅在用户主动选择或明确指定本技能时使用。独立完成人像、现实场景、实物、海报、信息图和手绘的美术指导、参考图运用、提示词与逐轮修正，减少稀碎感、塑料感、假光影和模板化处理。保留指定画风与主体特征，按交付要求检查成图；只要提示词时不生图。

Idioma do texto original: chinês

sweety-liulei-voice

把一句话、一段话改写成 liulei 本人口吻的句段级语音指纹 skill。触发词：用我的口吻、改成我的说法、像我一样说这句/这段、liulei 腔、把这段话转成我的风格、帮我把这个意思说成我的话、课程稿口吻、用我的话回复。输入是现成的一句话/一段话/几个要点，输出是 liulei 会怎么说它；也可按给定意思代拟单句或单段。不适用于：写/改整篇文章（用 sweety-liulei-writer）、起标题、搭结构、通过 AI 检测（做不到，见"检测边界"）。

Idioma do texto original: chinês

sweety-liulei-writer

用 liulei 的写作风格写、改写、续写中文文章时使用。触发词：用 liulei 风格写、帮我写篇、续写、扩写、改成 liulei 的口吻、像我这样写、按我的风格写、写成公众号文章/技术观点稿/复盘/水族新手向。适用于公众号深文、技术与 AI 观点、职场或项目复盘、水族养护经验、个人随笔。当用户只说"帮我把这个写成文章""用我的风格写一下"，且语境是个人表达而非客观资料时，也应触发。不适用于：纯资料整理与翻译、要求严谨第三方口径的客观报道、与 liulei 无关的平台平均科普或营销稿、纯标题或摘要生成。

Idioma do texto original: chinês

sweety-post-to-wechat

发布内容到微信公众号，支持 API 或 Chrome CDP 方式。支持文章发表（HTML、Markdown 或纯文本输入）和贴图发表（多图）。Markdown 文章模式默认将外部链接转为底部引用。触发词："发布公众号"、"post to wechat"、"微信公众号"、"贴图/图文/文章"。

Idioma do texto original: Vários idiomas

sweety-video-gen

通用 AI 视频生成。对接实现 /v1/media 开放协议的视频生成服务（如 seedance 2.0 中转源），用户只需配置服务地址和 API Key。支持文生视频、参考图生视频（图生视频）、时长与宽高比设置、任务查询、失败重试、自动轮询和结果下载。当用户要求生成视频、图生视频、文生视频、AI 视频时使用。

Idioma do texto original: chinês

sweety-image-reprocess

对已生成的图片做像素层重处理（注入真实高斯噪声、非均匀锐化、非整数重采样、JPEG 重编码、抹元数据），扰乱扩散/GAN 模型在频域与噪声层留下的统计指纹，降低"半真实"内容被 AI 检测器误判的概率。适用于"图片重处理""去 AI 频域指纹""降低 AI 检测概率""real grain/翻拍替代""图片重压破指纹"等请求。本 skill 只处理像素，不写提示词、不改提示词层 AI 味。

Idioma do texto original: chinês

sweety-four-layer-review

文章定稿前的四层自检终审。用于成稿后、发布前的最后把关，按"事实逻辑 → AI 腔 → 节奏 → 像不像作者本人"四层逐一检查并出质检报告。触发词：四层自检、终审、定稿前检查、过一遍、把把关、能不能发、查一下这篇。是 liulei 写作流水线的最后一棒（作者材料/初稿 → shuorenhua 边界审稿 → humanizer-zh 轻扫 → 本 skill 终审）。本 skill 只做检查与判定、把问题指回上游修复，不重写、不发明新规则。不适用于：还在选题/搭结构阶段的半成品（那时该用写作 skill）。

Idioma do texto original: chinês

sweety-trustworthy-title

中文自媒体可信标题生成技能。用于生成、改写、冷审公众号、小红书、滑读卡、图文、视频、封面字和作品标题，要求标题在 3 秒内让读者看懂对象、问题、情绪和点击理由；吸收问句、互动、痛点、故事、反常识、题文呼应、热点和投稿等开头范式中适合标题的部分，但核心目标始终是产出标题，避免震惊体、恐吓体、绑架体、悬空体、碰瓷体、许诺体和对立体标题党。

Idioma do texto original: chinês

sweety-viral-opening

中文公众号和自媒体爆款开头写作技能。用于生成、改写、冷审文章开头前三句话、首段、公众号黄金 3 秒开场、标题后的阅读承接和口播开头；以问句、对话、互动、自我表达、题文呼应、名言解构、热点和投稿等 8 种开头模板为核心，完成情绪引爆、价值承诺和阅读牵引，同时避免虚假投稿、硬蹭热点、悬空好奇、恐吓式开场和正文无法兑现的承诺。

Idioma do texto original: chinês

sweety-broad-hotspot-fetch

Use when the user asks to get today's broad hotspots, 泛热点, 热榜, 借势热点, 今日热点, or to judge whether current general trends can be naturally used by Sweety content projects. Do not use as a vertical search or Xiaohongshu/Douyin topic-research substitute.

Idioma do texto original: chinês

sweety-chinese-typography-prompt

中文字体与标题字生图提示词技能。用于生成或改写包含中文大字、封面字、海报字、banner 字、缩略图标题、Logo 感字形、手写字、书法字、涂鸦字、像素字、儿童绘本字等静态图像 prompt；要求把文字内容、字体类型、笔画、结构、重心、边缘、材质、光影、背景和整体气质写清楚，避免只写“生成几个字”“好看”“高级”“有设计感”等空指令。

Idioma do texto original: chinês

sweety-image-naturalizer

为生图任务去除 AI 味，将用户的图像需求改写成更自然、更可信、更少 AI 模板感的视觉创作 brief 和生成提示词。适用于“生图去除AI味”“图片去 AI 味”“写真实感图片提示词”“Pinterest/VHTC 风格图像 brief”“让 AI 图像更像人工策划/拍摄/图解”等请求。

Idioma do texto original: chinês

sweety-negative-short-tweet

生产负能量短推候选。用于用户要写中文 X/Twitter 短推、反鸡汤、冷幽默、自嘲、职场/生活/关系/成长焦虑类短句，或要求“负能量短推”“丧一点”“毒舌但别低级”“没灵感给几个方向”时；根据人设、用户输入、日期/天气/热点/记忆选择 3 个合适方向产出候选，并做平台风险与可信度过滤。

Idioma do texto original: chinês

sweety-post-to-juejin

Use when the user asks to publish, draft, or prepare a Markdown article for Juejin/掘金. Creates a safe browser-based Juejin draft from the original Markdown, supports cover image, category, tags, and brief, and defaults to stopping before final publish unless…

Idioma do texto original: chinês

sweety-post-to-x

Posts content and articles to X (Twitter). Supports regular posts with images/videos and X Articles (long-form Markdown). Uses real Chrome with CDP to bypass anti-automation. Use when user asks to "post to X", "tweet", "publish to Twitter", or "share on X".

Idioma do texto original: inglês

sweety-trend-source-intake

Collects trend, hotspot, and source evidence for content projects with Firecrawl while preserving project judgment boundaries. Use when the user asks to research trends, pick topics or products automatically, find X/Twitter posts, gather platform signals,…

Idioma do texto original: inglês

sweety-code-naming-comments

Cross-stack code naming and comment discipline for software development. Use when Codex writes, reviews, refactors, or explains code and must choose clear identifiers, rename variables/functions/classes/constants/files, add or remove comments, review naming…

Idioma do texto original: chinês

sweety-image-privacy

Use when sanitizing image privacy metadata, removing GPS/source traces, auditing macOS xattrs, or preparing images with iPhone-like camera metadata before publishing or AI-detection checks.

Idioma do texto original: inglês

sweety-karpathy-writing-style

借鉴 Andrej Karpathy 公开技术写作中的结构、语气和解释方式，帮助撰写英文 AI/技术长文，并可从 original/final 两版文本 diff 中自动提取写作规则来更新 SKILL.md。适用于英文技术随笔、年度回顾、AI 行业分析、观点文章、Karpathy 风格写作，以及用户要求从 AI 初稿和最终稿中学习偏好时。

Idioma do texto original: Vários idiomas

sweety-markdown-to-html

纯公众号排版 skill。把 Markdown、纯文本笔记或 Obsidian 风格文章转成微信公众号兼容的内联样式 HTML，支持 30 套主题、主题画廊、AI 排版增强、CJK 修复、脚注、图片处理和容器语法。用户提到公众号排版、微信 Markdown 转 HTML、主题预览、dialogue/gallery/callout 容器或需要兼容 sweety-post-to-wechat 的 Markdown 渲染时使用。

Idioma do texto original: chinês

release-skills

Universal release workflow. Auto-detects version files and changelogs. Supports Node.js, Python, Rust, Codex Plugin, and generic projects. Use when user says "release", "发布", "new version", "bump version", "push", "推送".

Idioma do texto original: inglês

ai-cat-wechat

全自动课程驱动的微信公众号生产技能。按固定栏目时间自动推进儿童线、成人线、晚间实战案例/作品集课和新闻补充，默认贴图优先，贴图发布失败自动降级为文章。

Idioma do texto original: chinês

sweety-article-illustrator

分析文章结构，识别需要视觉辅助的位置，通过类型×风格两维度方法生成配图。当用户说"为文章配图"、"添加插图"、"illustrate article"、"add images"、"generate images for article" 时使用。

Idioma do texto original: chinês

sweety-comic

Knowledge comic creator supporting multiple art styles and tones. Creates original educational comics with detailed panel layouts and sequential image generation. Use when user asks to create "知识漫画", "教育漫画", "biography comic", "tutorial comic", or…

Idioma do texto original: inglês

sweety-compress-image

Compresses images to WebP (default) or PNG with automatic tool selection. Use when user asks to "compress image", "optimize image", "convert to webp", or reduce image file size.

Idioma do texto original: inglês

sweety-cover-image

生成文章封面图，支持五维定制（类型、配色、渲染风格、文字、氛围），10 种配色方案和 7 种渲染风格。支持超宽 (2.35:1)、宽屏 (16:9)、方形 (1:1) 等比例。当用户说"生成封面"、"制作封面图"、"generate cover image"、"create article cover" 时使用。

Idioma do texto original: chinês

sweety-danger-gemini-web

Generates images and text via reverse-engineered Gemini Web API. Supports text generation, image generation from prompts, reference images for vision input, and multi-turn conversations. Use when other skills need image generation backend, or when user…

Idioma do texto original: inglês

sweety-danger-x-to-markdown

Converts X (Twitter) tweets and articles to markdown with YAML front matter. Uses reverse-engineered API requiring user consent. Use when user mentions "X to markdown", "tweet to markdown", "save tweet", or provides x.com/twitter.com URLs for conversion.

Idioma do texto original: inglês

sweety-format-markdown

Formats plain text or markdown files with frontmatter, titles, summaries, headings, bold, lists, and code blocks. Use when user asks to "format markdown", "beautify article", "add formatting", or improve article layout. Outputs to {filename}-formatted.md.

Idioma do texto original: inglês

sweety-image-gen

基于 OpenAI、Google、Relay（OpenAI 兼容中转）、OpenRouter、DashScope、即梦、Seedream 和 Replicate API 的 AI 图像生成。支持文本生成图像、参考图、宽高比设置及从已保存提示词文件批量生成。默认顺序执行；当用户已有多个提示词或需要稳定的多图吞吐量时，使用批量并行生成。当用户要求生成、创建或绘制图像时使用。

Idioma do texto original: chinês

sweety-infographic

生成专业信息图，支持 21 种布局和 20 种视觉风格。分析内容、推荐布局×风格组合，生成可发布的信息图。当用户说"生成信息图"、"信息图"、"可视化"、"高密度信息大图"、"infographic"、"visual summary" 时使用。

Idioma do texto original: chinês

sweety-post-to-weibo

Posts content to Weibo (微博). Supports regular posts with text, images, and videos, and headline articles (头条文章) with Markdown input via Chrome CDP. Use when user asks to "post to Weibo", "发微博", "发布微博", "publish to Weibo", "share on Weibo", "写微博", or "微博头条文章".

Idioma do texto original: inglês

sweety-slide-deck

Generates professional slide deck images from content. Creates outlines with style instructions, then generates individual slide images. Use when user asks to "create slides", "make a presentation", "generate deck", "slide deck", or "PPT".

Idioma do texto original: inglês

sweety-translate

Translates articles and documents between languages with three modes - quick (direct), normal (analyze then translate), and refined (analyze, translate, review, polish). Supports custom glossaries and terminology consistency via EXTEND.md. Use when user asks…

Idioma do texto original: inglês

sweety-url-to-markdown

Fetch any URL and convert to markdown using Chrome CDP. Saves the rendered HTML snapshot alongside the markdown, uses an upgraded Defuddle pipeline with better web-component handling and YouTube transcript extraction, and automatically falls back to the…

Idioma do texto original: inglês

sweety-xhs-images

生成小红书信息图系列，支持 11 种视觉风格和 8 种布局。将内容拆分为 1-10 张卡通风格图片，针对小红书互动优化。当用户说"小红书图片"、"XHS images"、"RedNote infographics"、"小红书种草"或想要中文平台社交媒体信息图时使用。

Idioma do texto original: Vários idiomas

sweety-youtube-transcript

Downloads YouTube video transcripts/subtitles and cover images by URL or video ID. Supports multiple languages, translation, chapters, and speaker identification. Caches raw data for fast re-formatting. Use when user asks to "get YouTube transcript",…

Idioma do texto original: inglês

tujie-wanwu-wechat

为公众号图解万物生成并发布微信草稿的中文专用技能。根据用户提供的话题、URL、关键词或自动发现的热点，完成资料检索、价值筛选、成稿生成、图片路由、去重记录与公众号草稿发布。适用于图解科普、结构拆解、历史事件、国漫剧情、AI 与技术解释类内容。

Idioma do texto original: chinês