viceroy-skills

微信公众号短评论发布技能，专为「直隶按察使」公众号的卡兹克风格评论方向定制。 适用：一事一议的短观点、热评reaction、资讯点评（**1000-1500字**，2026-06-04 佳哥确认）。 触发条件：用户说「评论」「热评」「观点」「点评」「说两句」。 ⚠️ **branding 检查铁律（用户零容忍红线，2026-06-04 与 zhiliGitHub 对齐）**： - 作者字段**固定填 `刘生`**（2 字符），**禁止**填 `卡兹克` / 留空 / 其他名字 - 文章底部「作者：xxx」也填 `刘生` - 禁用字符串清单：`['卡兹克', 'khazix', 'zhiliGitHub', 'zhiliComments', '本文由', '一键三连', '扫码', 'wzglyay', '自动发布', 'jacardl']` - 完整检查方式见 zhiliGitHub SKILL.md 顶部「🚫 branding 检查铁律」段

微信公众号长文发布技能，专为「直隶按察使」GitHub 黑马项目方向定制（1500-2000字）。 触发：用户说「写文章」「发长文」「GitHub」「黑马」。 技能边界：本技能只管 GitHub 黑马长文，**不替兄弟技能定规范**。短评/Reaction → `social-media/zhilicomments/`；日常复盘 → `openclaw-imports/zhili-publish/`。完整边界见 `references/skill-boundary.md`。 执行前必读：①写 markdown 草稿（1500-2000字）；② `python3 scripts/render_zhili_article.py /tmp/draft.md /tmp/article.html`；③ 配图调用 `zhili-illustration`；④ `python3 scripts/validate_zhili_article.py --title "<标题>"`；⑤ pre-submit 清单；⑥ `python3 scripts/push.py [--html /tmp/article.html]` 推草稿。配图最多 5 张，封面 16:9 PIL 裁剪 900×383。旧脚本 `publish_zhili.py` 已废弃（2026-06-28）。⚠️ 不要凭记忆手写 HTML CSS，用脚本生成。

直隶按察使公众号长文输出技能。基于素材（视频/音频/文字稿）产出 4000-5500 字的中文公众号长文（HTML + markdown 双格式），**自动对接 zhili-publish 推送到草稿箱**。 触发条件（满足任一即触发）： - 用户说"基于 [素材] 写一篇公众号长文"/"写一篇 4000-5000 字" - 用户粘贴视频链接（B 站/YouTube/播客）+ "写长文"/"写成公众号" - 已有 transcript_clean.txt + 用户说"扩写"/"做长文"/"写成文章" - 用户说"zhili long [主题]"或"长文输出" - 用户说"这篇写完后直接发到草稿箱"/"一键发布到直隶按察使" 不适用：短评（用 zhilicomments-publish）、教程（用 khazix-writer）、GitHub 项目介绍（用 khazix-writer 模板） ...

🛰️ 雷达数据采集（生产版 v5.1）：金价改用新浪hq_GC+USD/CNY换算，新增沪金Au99.99国内现货价，新增gold_note字段

Evidence-backed GEO keyword and Prompt Map research for Chinese AI visibility. Use when asked to generate GEO keywords, AI-search prompts, prompt/question coverage, Nanobot prompt-map JSON, or grouped questions for 品类问题、场景问题、品牌问题、竞品比较问题; also use for this GEO diagnostic page's Step 3 prompt analysis.

招投标 / 标书分析 Skill。默认输出两个 Markdown 文档：①应标清单（符合性、经济标、商务标、技术标、技术规格强制响应、视频/证明材料、排除项、P0-P4执行优先级）；②交付物清单（中标后系统/数据/模型/平台/云资源/部署/文档/验收/付款/售后/知识产权/安全/违约与澄清）。触发条件：标书分析、招投标分析、应标清单、投标材料清单、评分表拆解、交付物清单、RFP/RFQ/RFI、招标文件分析。

Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.

将 Skills 同步到 GitHub 仓库（viceroy-skills 公开仓库 + 私有专属仓库），并维护符合 khazix-skills 标准的 README。当用户说"同步skills到GitHub"、"上传skill到github"、"推送skill到远程"、"更新viceroy-skills仓库"、"维护GitHub上的skill仓库"、"更新 README"时使用此技能。包含完整的 GitHub 推送流程和标准 README 生成规范。

维护技能库整洁，系统性整理、归类、查重、清理废弃技能（含项目源码误入 skills/ 目录的情况），**以及在 install 前对 skill 做安全审计（调用 skillspector）**。**新能力（v2.2）**：对指定 skill 进行合规/规范检查（YAML frontmatter / description 规范 / 多余文件 / 目录结构），并与 viceroy-skills 云端同名 skill 比对，**自动三向 merge**，保证本地和云端都是最新文件。**新能力（v2.7）**：分析技能使用情况——直接查询 SQLite 数据库 session 历史，区分「mtime 活跃」和「实际被调用过」。触发词：「检查 skill 规范」「合规检查」「跟云端比对」「merge skill」「同步 skill 到云端」「检查并更新 skill」「哪些 skill 没用过」「清理未使用技能」「检查技能使用情况」。

从 PostgreSQL 读取今日雷达数据，分 4 条飞书消息发送（金价+AI+政治+GitHub），周六周报，月末月报

为「直隶按察使」公众号长文统一配图。读取 HTML → 提取 shot list → mmx image generate 生图 → 上传微信素材 → 注入 HTML。 触发条件：用户说「配图」「生成插图」「zhili-illustration」「给文章配图」。

Image generation via 9Router local gateway (cx/gpt-5.5-image etc.). Use when generating images for articles, reports, or social media, and 9Router is the available image backend.

写直隶按察使公众号 GEO 垂直系列文章。当用户说「写一篇 GEO 文章」「帮我写 GEO""「GEO 系列」「GEO 选题」「GEO 文章」「写个 GEO 稿」或类似需求时触发。为直隶按察使公众号撰写 GEO 垂直类文章——对比 SEO 与 GEO、分析 GEO 案例、解读 AI 搜索趋势、讲解品牌在 AI 时代的信息基础设施。故事驱动 + 数据锚点 + 结尾洞察，格式为微信草稿箱 HTML。

通用中国专利挖掘发现与交底书生成全流程：扫描项目文档挖掘专利点、讨论融合、基于脱敏模版生成技术交底书、联网查新、生成后自检含逻辑闭环与公式参数一致性。| Patent mining, disclosure drafting, prior-art search, and consistency self-check.

开源中文配图技能 + IP 角色库。用"挑认知锚点 → 现编隐喻 → 反 PPT 自检"的方法, 为中文深度文/方法拆解生成由固定角色出演的正文配图(不是通用插画,不是样式库选风格)。 自带 31 个原创 IP 角色(手绘线稿 15 + 谐音梗 meme 16)。 触发:用户说"配图""正文配图""IP 配图""给这篇配图""隐喻配图""选个角色配图""配图 shot list"。

人味儿写作 — 打磨、润色、改写别人的文字时，保住文字背后那个人的存在感。 Use when: 用户拿来一段自己写的文字，要求打磨、润色、修改、精简、改写、 "帮我改改"、"这段怎么样"。也适用于代笔后用户说"AI 味儿太重"、"没有人味儿"、 "感受不到我"的返工场景。触发词：打磨文案、润色、帮我改、改写、修改这段、 人味儿、AI 味儿。内置两层：底层原理（动不动、怎么动）+ 事后检查清单 （动过的地方有没有 AI 味儿，源自 Wikipedia Signs of AI writing / blader humanizer）。

生成横向翻页网页 PPT（单 HTML 文件），含 WebGL 背景、章节幕封、数据大字报、图片网格等模板。提供两种风格：① "电子杂志 × 电子墨水"（衬线 + 流体背景 + 暖色） ② "瑞士国际主义"（无衬线 + 网格点阵 + IKB/柠檬黄/柠檬绿/安全橙高亮）。当用户需要制作分享 / 演讲 / 发布会风格的网页 PPT，或提到"杂志风 PPT"、"瑞士风 PPT"、"Swiss Style"、"horizontal swipe deck"时使用。

Scan, explain, select, install, and call local AI models across Ollama, Hugging Face cache, MLX Whisper, whisper.cpp, LM Studio, and common Mac model stores. Use when the user asks to list local models, understand what installed models do, check whether models are healthy, manage local model caches, choose a local model for a task, run local inference/transcription/embedding, or prefer Apple Silicon / Mac-optimized models.

微信公众号长文发布技能，专为「直隶按察使」公众号的 GitHub 黑马项目方向定制。 适用：GitHub Trending 黑马开源项目介绍文章（1500-2000字，流式叙事）。 触发条件：用户说「写文章」「发长文」「GitHub」「黑马」。 **执行前必读**：每次写 HTML 前必须按顺序执行以下三步，再开始写作： 1. 读取规范参考文件 `references/streambert-reference.html`，提取 CSS 检查清单（字体、层级、高亮、分隔线、blockquote、标签行、作者信息位置等） 2. 按规范生成 HTML，内联所有 CSS 3. 生成完毕后，对照第 1 步的检查清单逐项验证，合格后再推送草稿

Synthesize user interviews, survey open-ended responses, transcripts, and synthetic user Q&A into evidence-backed product research insights, opportunity backlogs, next research plans, and knowledge graphs. Use when the user asks to analyze qualitative research for product strategy, MVP design, concept validation, customer insight, or cross-industry research synthesis.

GEO 策略分析报告生成技能 — 基于多维度 CSV 数据（可见度×准确率×人群×关键词×竞品×AI模型）生成完整的 GEO 优化策略报告，包含四象限优先级矩阵、竞品 Buzzword 分析、预算分配和分阶段执行路线图。

The world's most comprehensive AI-powered webpage audit tool. Analyzes 15+ key dimensions including AI readability, structured data, SEO, accessibility, mobile-friendliness, performance, security, image optimization, and social media integration. Delivers actionable recommendations with scored assessments.

从 AI 回答文本中高精度识别目标品牌的竞品，支持知识库持久化和增量积累。当用户粘贴 AI 平台回答并要求识别竞品、分析某品牌竞争格局、从回答中提取竞品信息时触发。也适用于：持续追踪某一品牌的竞品变化、多平台 AI 回答批量竞品分析。

GEO 关键词研究 Skill。为指定品牌生成行业推荐型 + 品牌直指型关键词报告， 支持 DAG 并发搜索模式，产出两份 Markdown 文件。

雷达数据采集 Skill — 采集金价、国际政治、AI热讯，写入 PostgreSQL 向量数据库，验证新闻真实性，按日期归档。

Analyze brand or product relevance against CID audience data. Use when Codex needs to evaluate a brand, product, product family, or URL against CID age-city audiences; generate conversion and brand-mind audience rankings; compute Affinity Index and Chi-Square evidence; inspect CID Excel or PostgreSQL data; create auditable reports with sources, product attributes, numeric evidence, and plain-language explanations.

Scrapling — adaptive web scraping framework for Python. Handles everything from single requests to full-scale crawls. Supports anti-bot bypass (Cloudflare Turnstile), browser automation (Playwright), and adaptive element parsing. Use as fallback when 9Router web-fetch fails.

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 pre-Defuddle HTML-to-Markdown pipeline when needed. If local browser capture fails entirely, it can fall back to the hosted defuddle.md API. Supports two modes - auto-capture on page load, or wait for user signal (for pages requiring login). Use when user wants to save a webpage as markdown.

将多个 Markdown 文件智能合并为结构化报告 HTML，突出重点洞察与数据。触发条件：(1) 用户提到"合并报告"、"生成报告"、"报告 HTML"、"访谈合并" (2) 用户指定目录下的多个 MD 文件需要整合 (3) 用户要求将 markdown 转为可汇报的 HTML 格式 (4) 用户提到 taste-skill 或 taste-skill 风格报告

End-of-session knowledge cleanup with OCD-level rigor — reconciles project docs (CLAUDE.md, README.md, docs/) and agent memory against the code so nothing rots. 会话结束后对项目文档和记忆进行洁癖级审查与同步。MUST trigger when the user says: "sync up", "tidy up docs", "update memory", "clean up docs", "/sync", "/neat", "同步一下", "整理文档", "整理一下", "更新记忆", "梳理一下", "收尾", "这个阶段做完了", "新人能直接上手", or any phrase suggesting a dev milestone where knowledge needs reconciliation. Also trigger when the user reports stale docs, conflicting memories, or wants a clean handoff to teammates or other agents. Bare "整理" / "tidy" with prior dev context counts — do not under-trigger. Cross-platform: works on Claude Code, OpenAI Codex, OpenCode, and OpenClaw.

Comprehensive SEO analysis for any website or business type. Full site audits, single-page analysis, technical SEO (crawlability, indexability, Core Web Vitals with INP), schema markup, content quality (E-E-A-T), image optimization, sitemap analysis, and GEO for AI Overviews/ChatGPT/Perplexity. Industry detection for SaaS, e-commerce, local, publishers, agencies. Triggers on: SEO, audit, schema, Core Web Vitals, sitemap, E-E-A-T, AI Overviews, GEO, technical SEO, content quality, page speed, structured data.

Technical SEO audit across 9 categories: crawlability, indexability, security, URL structure, mobile, Core Web Vitals, structured data, JavaScript rendering, and IndexNow protocol. Use when user says "technical SEO", "crawl issues", "robots.txt", "Core Web Vitals", "site speed", or "security headers".

Search Experience Optimization: reads Google SERPs backwards to detect page-type mismatches, derives user stories from search intent signals, and scores pages from multiple persona perspectives. Identifies why well-optimized pages fail to rank by analyzing what Google rewards for each keyword. Use when user says "SXO", "search experience", "page type mismatch", "SERP analysis", "user story", "persona scoring", "why isn't my page ranking", "intent mismatch", or "wireframe".

Analyze existing XML sitemaps or generate new ones with industry templates. Validates format, URLs, and structure. Use when user says "sitemap", "generate sitemap", "sitemap issues", or "XML sitemap".

Detect, validate, and generate Schema.org structured data. JSON-LD format preferred. Use when user says "schema", "structured data", "rich results", "JSON-LD", or "markup".

Programmatic SEO planning and analysis for pages generated at scale from data sources. Covers template engines, URL patterns, internal linking automation, thin content safeguards, and index bloat prevention. Use when user says "programmatic SEO", "pages at scale", "dynamic pages", "template pages", "generated pages", or "data-driven SEO".

Strategic SEO planning for new or existing websites. Industry-specific templates, competitive analysis, content strategy, and implementation roadmap. Use when user says "SEO plan", "SEO strategy", "SEO planning", "content strategy", "keyword strategy", "content calendar", "site architecture", or "SEO roadmap".

Deep single-page SEO analysis covering on-page elements, content quality, technical meta tags, schema, images, and performance. Use when user says "analyze this page", "check page SEO", "single URL", "check this page", "page analysis", or provides a single URL for review.

Maps intelligence for local SEO — geo-grid rank tracking, GBP profile auditing via API, review intelligence across Google/Tripadvisor/Trustpilot, cross-platform NAP verification (Google/Bing/Apple/OSM), competitor radius mapping, and LocalBusiness schema generation from API data. Three-tier capability: free (Overpass + Geoapify), DataForSEO (full intelligence), DataForSEO + Google (maximum coverage). Use when user says "maps", "geo-grid", "rank tracking", "GBP audit", "review velocity", "competitor radius", "maps analysis", "local rank tracking", "Share of Local Voice", or "SoLV".

Local SEO analysis covering Google Business Profile optimization, NAP consistency, citation health, review signals, local schema markup, location page quality, multi-location SEO, and industry-specific recommendations. Detects business type (brick-and-mortar, SAB, hybrid) and industry vertical (restaurant, healthcare, legal, home services, real estate, automotive). Use when user says "local SEO", "Google Business Profile", "GBP", "map pack", "local pack", "citations", "NAP consistency", "local rankings", "service area", "multi-location", or "local search".