yang0

Turn a 001–367 hand-drawn and realistic style number and image theme into bilingual prompts, using model capability data to decide when core traits and a numbered reference image are required.

极简高效的文章与视频封面设计技能。根据输入文章或视频字幕文件自动提炼200字核心摘要、推算受众群体、默认输出微信公众号标准 2.35:1 超宽画幅封面（兼顾X 5:2与安全留白），或按需匹配小红书(3:4)、B站/YouTube/短视频(16:9/9:16)场景；深度协同全库 365 种视觉风格与36种经典配色；视频封面模式自动注入“视频封面设计，网感强，吸引人眼球”，并结合“先设计隐喻再出图，主标题明显，小字少或者没有小字”核心指令，一键输出高质量结构化封面生图提示词。

深度分析文章脉络，精准规划插图点位与视觉意象。对知识类/教程类文章默认启动“认知降负与高保真图替繁文机制”，严禁删除公式与量化指标，强制协同 170 种排版图型（IG-*信息图/SC-*社媒卡）与 365 种视觉风格/36 种经典配色，杜绝微缩杂物摆拍；对非知识类文章保持叙事氛围与意境。严格执行双向间距（400~1200字）与全篇章节均衡覆盖，支持自主生图或全自动插图回填排版。

通用手绘风社区发帖与静态页面/资产发布 Skill。打通 handraw-style.com (Discourse) 官方论坛账号系统与静态画廊发布系统。涵盖双重能力：(1) 全板块论坛发帖与管理（创作交流、好用的工具、综合大厅、运营与变现、实战作品与爆款、手绘skill）：支持发布通用长帖、出版级手绘提示词卡片、自动配图直传短链、回帖、帖子查询与删除；(2) 风格画廊与静态资产一键部署上云（index.html、tutorials.html、assets.html、367…

内置300多个画风，图型，主题色。支持纯图模式，图文模式，可以给文章做封面，批量制作插图。可以批量制作海报，漫画，小红书组图。信息图，ip设计稿，电商图甚至真人摄影都随意拿捏。

专门生成极简结构化 AI 海报提示词的专业设计技能。彻底摒弃冗长描述与套路化格式，深度协同项目 365 种视觉风格与 36 种经典主题色画廊，按核心业务与设计字段输出提示词（主题、业务场景、受众、内容密度、情感基调、主题色、editorial风格、画风），让生图模型（如 GPT Image 2/2.5）自主完成原生高质量构图与排版。

专为 HandRaw 新【画风 (Style)】或新【图型 (Layout)】发布专栏文章量身定制的实测配图规划与批量生图 Skill。因为发布的本质是实测手绘 Skill 的画风与图型在商业场景中的表现边界，所以必须将商业主题与手绘 Skill 中的图型/画风深度咬合：当适合的商业主题推导出来后，强制为每个主题严密搭配契合的 HandRaw 排版图型（或跨流派画风）；随后严格执行【五维实体异构与零雷同铁律】，5…

跨媒介“角色视觉语言 × 场景视觉语言”双风格共存融合提示词生成技能。支持融合两套不同视觉风格编号（#001–#367），或融合风格编号与写实风格，并可按需注入图型（170 种）与主题色（36种），严格输出造型解耦、物理光影与叙事共存的高质量生图提示词。

x-expert 短推爆款创作总控 Skill（Master Short Tweet Orchestrator）。 严格按序串联 3 大专业节点 Skill（evidence-researcher ➔ hook-and-headline-writing ➔ topic-reviewer）， 从一手事实核验 (evidence_pack)、并发撰写 5 篇完整不同风格候选推文 (S1≤25, S2≤35)、5 维量化打分大榜评选 TOP 1 冠军， 最终以标准 Thread 双帖模板完整落盘至…

X 平台爆款推文与开篇钩子写作 Skill。以传播性、人性洞察与情绪截停为绝对第一基座，将 2026 新算法（Phoenix 两塔向量检索、实体密度、防降权护栏）作为隐形助推无缝融合。 严格遵循：首句纯钩子抓人（≤25字）➔ 二句亮反差与关键数据（≤35字）➔ 正文 1️⃣ 2️⃣ 3️⃣ 列表自然装载具体实体 ➔ 评论区沉淀外链。 支持 5 大经典中文爆款风格、15 种爆款公式与 4 U's 量化择优。 触发关键词：写推文、出钩子、写开头、hook、headline、出候选、批量写推、生成推文。

对推文/选题候选进行严苛的多维量化打分质检，担任"以挑剔著称的内容质检总监"角色。 坚持传播性与爆款网感第一，同时核验 2026 算法隐形健康指标（实体密度自然达标、0 外链防跳出降权、AI 腔一票否决）。 输入：一组推文候选列表（通常由 hook-and-headline-writing 生成）。 输出：每条推文的 5 维评分、PASS/REVISE/REJECT 判定、可执行修改指令与综合打分大榜。 触发关键词：打分、质检、评分、哪条最好、审核推文、topic review。

全场景通用事实证据研究员。覆盖 5 大类来源（GitHub开源、SaaS商业产品、X推文动态、官方技术文档/论文、真实商业账本），在动笔写推之前强制执行物理事实检索、核验真实一手链接（拒绝臆造假URL）、提取客观参数、强制挖掘局限性与反方声音，输出标准化 evidence_pack 证据包，彻底杜绝推文误解、功能夸大与凭空脑补。 触发关键词：搜证据、核实事实、查数据、找反方、证据包、evidence pack、排坑调研、泛用核验。

微尘印记 (@weichen_ink) 的写作风格 Skill。Build-in-Public 叙事者，AI工具×涨粉变现×个人成长。 用于写推文、项目汇报、经验分享。触发词：「用微尘风格」「微尘写法」「像微尘那样写」「Build-in-Public风格」。 NOT for: 正式商业文案、学术写作、品牌官方公告。

x-expert 深度长文创作总控 Skill（Master Long-form Article Orchestrator）。 支持 3 大素材输入模式（音视频转录 / 素材库线索聚合 / 网络深度调研与高维认知升维）， 串联 7 大专业写作节点 Skill（video-transcript-article + evidence-researcher + dbs-spread + dbs-deconstruct ➔ growth-writing-standards ➔ dbs-resonate ➔…

诊断如何把已经确定的选题做成好内容。用户要求设计内容、检查文案或改善内容表达时使用。

用维特根斯坦与奥派经济学方法拆解模糊的商业概念。用户要求解释一个词、澄清概念边界或识别概念混用时使用。

Orchestrates professional portrait, couple, wedding, family, cosplay, and themed photography planning from direction selection through Look, scene micro-zones, Shot List, and on-demand image generation.

Early professional hairstylist node that designs and locks formal shoot hair for each candidate direction before downstream Look expansion.

Design numbered six-look wardrobe boards for one or any number of participants inside a confirmed portrait theme, preserving full per-person outfit detail.

Confirm portrait, couple, family, group, wedding-party, team, cast, travel, period, or cosplay photo requirements and produce complete per-participant Looks, scenes, props, and a confirmed photography handoff for any number of subjects.

Mandatory formal-shoot makeup and skin optimization node for portrait photography.

Turns a real-person reference into a reusable polished identity reference through a visual three-panel beauty exploration.

Turn confirmed one-person or multi-person portrait styling handoffs into detailed eight-shot plans, numbered grids, and selected detail-render briefs without reducing per-participant specificity.

Orchestrates professional portrait, couple, wedding, family, cosplay, and themed photography planning from direction selection through Look, scene micro-zones, Shot List, and on-demand image generation.

Coordinate original mascot requests for projects, accounts, or folders, directly producing three distinct complete IP candidates; exclude personal-photo asset packs.

Automatically create and score three distinct mascot directions from readable project context, or redesign one direction holistically after feedback; this node never renders or delivers images.

Render three distinct complete mascot IP candidates from internal art-direction records, or regenerate only one affected candidate after holistic revision.

Inspect mascot candidates for anatomy and scope compliance, then deliver QA-approved choices with plain-language revision guidance and DG/MS gallery links.

Reframe mascot feedback as a whole-design constraint and produce one coherent revised direction for the affected candidate, including vague non-professional feedback.

Read project, account or folder evidence and normalize an original mascot brief before three-candidate art direction; exclude personal-photo asset packs.

Coordinate original character and person IP requests from photos, existing cartoon images, folders, URLs or text descriptions, directly producing three distinct complete IP candidates; specialized for human and character IP design.

Render three distinct complete character IP candidates from internal art-direction records, or regenerate only one affected candidate after holistic revision.

Inspect character IP candidates for human anatomy, fingers, eye symmetry, and scope compliance, then deliver QA-approved choices with plain-language revision guidance and PG/MS gallery links.

Automatically create and score three distinct character IP directions (A 气场觉醒 / B 平行宇宙 / C 潮玩先锋) from Identity Locks and context, mapping PG person forms and MS render styles; no conservative options; no image calls.

Read photo/image evidence or text/link/folder project brief, extract human facial cues, hair, demeanor, outfit, and build normalized Identity Lock before three-candidate art direction.

Reframe character IP feedback or PG/MS codes as a whole-character constraint and produce one coherent revised direction for the affected candidate, including vague feedback.

Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies

Use when you have a written implementation plan to execute in a separate session with review checkpoints

Use when executing implementation plans with independent tasks in the current session

Use at the start of every Codex CLI conversation or new user request to enforce skill-first execution: check and read relevant skills before planning, clarifying, coding, or running tools

Use when you have a spec or requirements for a multi-step task, before touching code

Use when creating new skills, editing existing skills, or verifying skills work before deployment

根据 ArtShift 画廊编号或用户照片，生成风格迁移提示词和图片；未选编号时推荐临时原创方向。

根据 ArtShift 画廊编号或用户指定画风与参考图，生成风格迁移提示词和图片；无编号时推荐已有风格与原创方向。

Capture a specified HTTP(S) webpage as a verified PNG with viewport, full-page, or unique-element modes.

Capture a focused screenshot card from Weibo, Zhihu, X, Xiaohongshu or Douyin through a logged-in local Chrome CDP session.

Collect top-N YouTube hot videos for any domain from the last 24 hours with Patchright, then output title, intro snippet, metrics, and link. Use when asked for a ranked trend digest in a specific field such as AI, finance, robotics, gaming, cybersecurity, or…

Extract subtitle files from YouTube video or playlist links using yt-dlp and authenticated cookies. Use when asked to download manual subtitles or auto captions from YouTube URLs and save them locally as subtitle files (for example .vtt/.srt).

Submit prompts to https://grok.com using an authenticated cookie file and capture the generated answer verbatim. Use when the task requires raw prompt passthrough, raw Grok output capture, and reproducible artifacts without JSON parsing.

Scrape full X tweet text from a seed tweet URL, including same-author continuation posts in the thread and related images, using patchright with Netscape cookies.

生成可按名称选择已注册 IP（默认牙仔；含绒宝、阿龅和确认注册的个人角色）的中文正文配图；当用户同时提到角色、这个IP、显式调用 $rongbao-illustrations 或带IP，以及封面、竖版海报、方图、萌粒/角色锚点/转面/3:4 信息图/贴纸，Baoyu/宝玉 的内容能力，或真人照片个人 IP、本人/博主卡通形象、照片转卡通、人物表情包/动作包时，按注册依赖组合调用目标设计 Skill。用于正文配图、文章插图、配图建议、shot list、去标题/改图和 create|prompt…

根据拍摄主题原创单人姿势，或在用户明确提供 Pose ID 时读取该姿势的完整结构；不做库内姿势推荐或图页引用。