deusyu

De-AI-flavor portrait prompt engineering — build photorealistic commercial portrait prompts through an 8-dimension framework (persona, fabric, moment, lens, light & skin, background, finishing, negatives), run a mandatory anti-AI-flavor self-check, then adapt the output to the target model (Midjourney / Stable Diffusion / Flux / GPT-Image / Seedream). Trigger phrases: 人像提示词, 去AI味, 塑料感, 商业人像, 质感人像, 老钱风人像, portrait prompt, photoreal portrait, de-AI portrait, realistic portrait prompt.

把 Mermaid 图表变成可点击的预览链接（mmd.dyu.sh），省去从对话里复制粘贴代码。 Turn Mermaid diagrams into clickable preview links — no copy-paste. 触发: 当你在回复中输出了 Mermaid 图表 / ```mermaid 代码块时，主动在其后附上预览链接； 或用户说 "打开这个图", "预览这个图", "mermaid 链接", "在浏览器里看这个图", "open this diagram", "mermaid link", "preview mermaid"。

Check whether a proposed company, product, app, or App Store seller name is already used or likely to conflict. Use when the user provides a name and asks to check duplicates, availability, registrability, domain availability, App Store seller suitability, social/project/package-name collisions, trademark risk, state LLC name availability, D-U-N-S/Apple identity risk, or wants a concise clearance report before forming a US LLC or choosing a brand. Can also generate names, but duplicate checking is the primary behavior.

判断 Mac 或 MacBook 是否值得换新。按真实瓶颈而不是芯片代数做决策，比较 Apple Silicon M/Pro/Max/Ultra，输出不用换、平替、轻升级、真升级、必要升级。

实时汽车问答与对比分析系统。当用户询问任何买车、选车、汽车参数对比、车型评测、价格分析相关问题时触发此 Skill。 触发场景（只要涉及以下任一情形就必须使用此 Skill）： - 车型参数对比："小米SU7和Model 3哪个好"、"国产车和特斯拉对比" - 配置/价格查询："Model Y 焕新版座椅加热有吗"、"问界M9多少钱" - 真实车主评价："XX车口碑怎么样"、"懂车帝评分" - 购车决策辅助："20-30万预算推荐什么车"、"新能源SUV怎么选" - 车辆功能查询："这款车支持V2L吗"、"有没有露营模式" - 销量/市场数据："2024年最畅销新能源车" 即使用户没有明确说"帮我对比"或"查一下"，只要话题涉及具体车型的任何属性，都应该主动触发此 Skill 进行实时数据检索，而不是依赖训练数据回答。

天气查询 — 实时天气、天气预报、生活指数、城市搜索。数据源为和风天气 (QWeather)。 触发词: 天气, 气温, 下雨, 预报, weather, forecast, 穿什么, 洗车, 紫外线, UV, 空气, 今天天气, 明天天气, 北京天气, 或任何 "[城市] 天气" 格式的输入。

中国节假日/调休查询 — 查某天是工作日还是休息日、是否调休补班、全年假期安排。 触发词: 节假日, 调休, 补班, 放假, 上班, holiday, workday, 今天上班吗, 明天放假吗, 春节放几天, 国庆放假, 下个工作日, 或任何 "[日期] 是否放假/上班" 格式的输入。

汇率换算 — 查询实时汇率、历史汇率、汇率趋势。数据源为欧洲央行 (ECB)。 触发词: 汇率, 换算, exchange rate, convert currency, 多少钱, 美元, 欧元, 日元, 英镑, USD, EUR, JPY, GBP, 或任何 "[金额] [货币A] 转/换/to [货币B]" 格式的输入。

把一批 Harness Engineering 调研候选（文章/论文/工具的 URL）走完「抓取→翻译→评审→收录→清理」流水线，整合进本仓库 works/ 与 references/articles.md，并保持 C1–C12 一致性检查全绿。当用户说"处理这批调研候选 / 收录这些链接 / 整理 translate / 把这几篇翻译收进来"时使用。这是仓库给自己用的策展 harness。

Translate books (PDF/DOCX/EPUB) into any language using parallel sub-agents. Converts input -> Markdown chunks -> translated chunks -> HTML/DOCX/EPUB/PDF.

AI-powered resume system. Manages structured experience data and generates tailored LaTeX resumes for different job targets. Use when user says 'cv', 'generate resume', 'make resume', 'init resume', 'add experience', '简历', '生成简历', '初始化简历', or '添加经历'.

Turn vague concerns, decisions, and forecasts into resolvable, updatable, scoreable probabilistic predictions. Use when the user asks "will X happen", "should I do Y", "how likely is Z", expresses anxiety about a future outcome, weighs a bet or risk, or wants calibrated probability tracking. Triggers on Chinese (概率、预测、应不应该、会不会、焦虑、值不值得、风险) and English (probability, forecast, predict, what's the chance, should I, will X happen). Also handles update/settle/review/coach modes for an existing forecast ledger.

通过脚本直连 Google Maps Platform API 完成地理编码、逆地理编码、路线规划、地点搜索、地点详情、海拔查询和时区查询。用户要求"Google Maps 查询""国际路线规划""地点搜索"或需要用命令行脚本调用 Google Maps API 时使用。

Deep travel research workflow. Input a city name, auto-profiles the city's character, dynamically selects research dimensions, generates structured knowledge document (org-mode) + portable reference cards (PNG). Adapts to any city type: ancient capitals, coastal cities, ethnic regions, modern metropolises. Use when user says '旅行研究', '出发前功课', 'travel research', or provides a city name with intent to do deep travel preparation.

锦书 (jinshu) —— 把 Markdown 用 mdnice 风格主题排版成微信公众号样式，实时预览/一键复制，并可直接发到公众号草稿箱。Use when user says "jinshu", "锦书", "公众号排版", "Markdown 转公众号", "发公众号草稿", or wants WeChat Official Account typesetting with mdnice themes.

Analyze facial/body photos to detect signs of cosmetic surgery or aesthetic procedures. Use when the user uploads a photo and asks to identify cosmetic work, detect plastic surgery, assess facial naturalness, check if someone has had work done, analyze before/after photos, or evaluate aesthetic procedure signs. Also trigger when users ask about specific procedures visible in photos (fillers, Botox, rhinoplasty, jaw contouring, etc.), compare photos for surgical changes, or want a "naturalness score" for a face. Works with single images, multiple comparison images, and video screenshots. 触发词：整容检测、看看有没有整、自然度评分、鉴定一下、有没有动过。

Deep research and decision analysis for car purchases. Takes candidate car models, runs a 6-phase pipeline (basic dossiers, negative investigation, safety rate analysis, industry baseline, cross-comparison, weighted decision matrix), and outputs a complete research package with actionable recommendations.