nanzhipro

针对指定 PDF 或 Office 文档（.pdf/.docx/.xlsx/.pptx）进行全文解析、检索与深度理解，识别并汇总隐私/敏感信息（PII、金融、健康、账号凭据等），输出审计报告与风险提示。用于合规审查、数据脱敏准备、外发前检查或隐私扫描。

将平庸、模板化、AI 味重的草稿重构为更有人味、更有张力、更有说服力的叙事。Use when Codex needs to improve or design stories, speeches, essays, founder narratives, sales copy, leadership messaging, positioning narratives, or AI-generated drafts by applying RENT (Relatability, Ease, Novelty, Tension), Pyramid Principle, Big Idea, Story Spine, Hero's Journey, audience-muse alignment, vulnerability, taste-based editing, or explicit requests about Super Skill, Storytelling Economy, Vortex of Mid, 叙事升维, 共鸣, 钩子, 张力, 脆弱性, 品味, or 先独立思考后 AI 辅助.

极客级 YouTube 播客提取与可视化方案。遵循电影感字幕视觉标准 (Cinema Style)，通过词窗重叠算法去重字幕，并利用 Playwright 渲染高质量金句卡片。适用于需要将长视频转化为高价值社交分享内容的场景。

专业书籍核心内容萃取专家。基于语义深度分析，提炼书籍 50% 的核心高价值内容，构建费曼讲解主线。

EPUB 高清转换专家。技术脚本生成出版级 PDF/MD，支持内置字体提取与高清排版重构。

抓取微信公众号文章，支持全文提取、图文混排、图片本地化下载及概要生成。在需要访问微信公众号链接并转化为 Markdown 时调用。

Downloads Xiaoyuzhou podcasts, transcribes using FunASR, and generates raw, structured, and optimized text files. Invoke when user provides a Xiaoyuzhou episode link.

Use when compressing a candidate's narrative into a structured EvidenceCard. Teaches the agent the field semantics, what counts as a critical_incident vs resume_claim, and when to mark insufficiency.

Use when interviewing a candidate. Teaches the agent to drive the dialogue around concrete critical incidents rather than abstract self-assessment, and to chase one missing cue per turn.

Use when naming a cognitive mechanism from evidence cards. Teaches the agent what counts as a high-value mechanism vs a soft-skill cliché, and how to express it as a four-part pattern.

Use when coordinating the seven-stage extraction pipeline. Tells the Orchestrator when to delegate, when to gate, when to escalate.

Use when composing the final extraction_report. Teaches how to write candidate-usable resume rewrites and interview narratives, separating validated mechanisms from probable ones.

Use when reviewing a proposed mechanism for flaws. The Skeptic does NOT propose new mechanisms — only stress-tests existing ones with a fixed checklist of bias patterns.

Bootstrap and operate an LLM Wiki: a persistent Markdown knowledge base with immutable raw sources, an LLM-maintained wiki, a SCHEMA.md operating contract, thin runtime pointer files, EXTEND.md preferences, and optional local BM25 search. Use when the user asks to create or bootstrap an LLM wiki, ingest sources, query or lint a wiki, configure EXTEND.md preferences, or set up BM25/full-text search for an LLM Wiki.

Set up a Phase-Contract long-running workflow that drives an AI agent to work continuously for 5+ hours on a single project. Use when the user wants to plan a large migration, new product build, big refactor, long-form writing project, compliance remediation, or any multi-phase task that exceeds a single context window. Triggers: '让 AI 连续工作', '长任务规划', '分阶段执行', 'phase 规划', 'plan 规划', '多阶段迁移', '大型重构规划', 'long-running AI workflow', 'phase-contract', 'planctl', 'manifest + execution 双文档', '拆 phase', '规划压缩恢复', '给 AI 写 plan', '项目分阶段', '续跑规划', '合同链规划', '长任务续跑工作流'. Scaffolds plan/manifest.yaml, plan/common.md, plan/phases/, plan/execution/, a planctl scheduler script, plus synchronized agent instruction files (.github/copilot-instructions.md for Copilot, CLAUDE.md for Claude Code, AGENTS.md for Codex) so phase state is externalized and AI work can survive context compression across all three agents.

把一个功能 / 变更 / 重构 / 集成 / 迁移 / 演进需求,转写为符合 arc42 12 章框架的中文技术方案文档,写入本地文件。当用户要求"写技术方案 / 写设计文档 / 出架构方案 / 写 RFC / 写 spec / 生成 arc42",或在实现非琐碎变更前需要先建立结构化方案时使用。输入:需求描述(必填)、目标文件路径(可选)、规模等级 S/M/L(可选)。输出:单个 markdown 文件,含 arc42 全部 12 章节,逐章经过契约校验与跨章一致性门禁。下游 AI 可以基于该文档直接进行代码实现,人类可以基于该文档进行架构评审。

Remove LLM citation artifacts such as `citeturn31view0turn16view0` from Markdown or text reports while preserving all report content. Use when the user asks to delete citation markers, cite artifacts, `turn...view/search` markers, or private-use glyph sequences from generated files without editing the surrounding prose, tables, punctuation, or whitespace.

Internalize an open-source repo by extracting all four layers — Knowledge / Skill / Patterns / Mindset — not just "what it does" but the author's mental models, value trade-offs, and decision frameworks. Produces publishable public research docs, a transferable methodology, and a seed project where you reuse the mindset on a different domain. Portable by design: pure Markdown instructions that work as an installed skill or as a prompt file in any LLM runtime. If your environment lacks task lists or sub-agents, run the same workflow sequentially. 深度内化一个开源仓库:从表层(知识/技能)穿透到冰山之下(思维模式/心智 mindset), 不止学"它是什么",更学作者的决策框架和世界观。产出可公开发布的研究文档 + 可迁移方法论 + 把 mindset 应用到不同业务场景的种子项目。 设计上可移植:纯 Markdown 方法论,既可作为 skill 安装,也可直接交给任意 LLM 执行。 若环境没有任务清单或子代理能力,按相同 phase 串行执行即可。 **Perfect for:** - 想"内化"某个开源仓库的 mindset(不止"会用",而是"会创造类似的东西") - 评估技术选型时,理解作者的设计哲学和取舍框架 - 学习行业标杆背后的思维模型与元认知(meta-cognition) - 想把 X 的方法论搬到自家不同业务场景,而非照搬代码 - 团队培训:让团队学到的是"思考方式",不只"功能列表" **Not ideal for:** - 单纯 code review / bug 修复(用专门 review 工具) - 仅需阅读 README 就够的简

帮助候选人或职场人通过多轮文字互动,从真实工作事件中萃取并命名其反复发挥作用的高价值认知机制,产出可用于简历改写、面试表达与职业发展的结构化报告。不做招聘判定、不做心理诊断。

Set up a disk-backed Phase-AI-Anime workflow for AI anime drama and manga-motion video projects. Use when the user wants to create, plan, continue, reset, or revise AI 漫剧, AI 动漫短剧, webtoon-to-video, anime short drama, storyboard-to-video, image/video/TTS generation pipelines, shot lists, dialogue scripts, asset prompts, audio-video timelines, or final assembly manifests across many sessions. Triggers: 'AI 漫剧', 'AI 动漫视频', '动漫短剧', '漫剧工作流', '分镜脚本', '音画同步', '生图到视频', 'AI drama pipeline', 'anime video workflow', 'storyboard to video', 'image to video', 'TTS timeline', 'phase-ai-anime', 'reset phase', '重置 phase', '从 0 开始'. Scaffolds plan/manifest.yaml, plan/common.md, phase and execution contracts, planctl, profiles, generation job specs, and synchronized agent instruction files so production state survives context compression across Copilot, Claude Code, and Codex.

Develops self-media long-form articles, opinion essays, newsletters, blogs, and columns through Socratic inquiry, six independent writing-cat subagents, public fact checking, article starter packs, and draft skeletons. Use when users ask for AI-assisted writing, topic expansion, writing angles, article structure, public fact support, multi-perspective drafts, self-media posts, deep commentary, or Chinese/English long-form writing. Triggers include AI协作写作、写作猫、六只写作猫、六只思考猫写作、选题、文章选题、写作思路、写作角度、文章结构、自媒体长文、观点文章、深度写作、帮我写文章、帮我拓展思路、帮我找资料、帮我生成初稿、thinking cats for writing、AI writing partner。

Analyzes personal decisions with six independent thinking cats and outputs a Mermaid decision tree. Use when users are making a choice, weighing tradeoffs, or feeling stuck in work, life, business, or investing. Triggers include 决策困难、纠结选择、不知道该怎么选、要不要做某件事、权衡利弊、分析风险、换工作、创业、投资、买房、换车、搬家、六顶思考帽、六只思考猫、thinking hats、thinking cats、帮我分析一下、帮我想想。

Set up a disk-backed Phase-Fiction workflow for long-form fiction projects. Use when the user wants to plan, continue, or revise a novel, novella, serial fiction project, multi-book arc, story bible, scene matrix, or full-manuscript rewrite across many sessions. Triggers: '写小说', '长篇小说规划', '续写长篇', '分阶段写小说', '故事 bible', '人物弧光规划', '章节矩阵', '修订轮次', 'serial fiction workflow', 'novel planning', 'story phase', 'scene matrix', 'revision pass', 'manuscript recovery', 'phase-fiction'. Scaffolds plan/manifest.yaml, plan/common.md, phase and execution contracts, a planctl scheduler script, synchronized agent instruction files, and a final story/README.md index after creation completes.