啥都懂的小灰
2008年9月,江河学院。
“解散!晚会大家好好玩!”
教官的哨子还含在嘴里,我已经拍了一下阿振的后背,朝食堂方向努了努嘴。他秒懂,两个人从队列里窜出去,迷彩裤腿带着风。
军训最后一天,晚上有迎新晚会。但那是吃完饭以后的事。
火速扒完饭走出食堂,月亮已经挂上来了。我跟寝室五个人慢慢往操场走,按班级顺位找了块空地坐下。草皮有点潮,隔着迷彩裤能感觉到凉意。
舞台上学长学姐在唱歌,台下一片绿,男生女生全穿着迷彩坐在地上。我的眼睛没怎么看台上。
“咱们班周围的人你认全了么?”我凑近阿振的耳朵。
他撇了撇嘴,也跟着扫了一圈:“不换衣服,你想认出个啥来?”
“也是。”我往后靠了靠,胳膊撑在草地上,“跟高中穿校服一个道理。再加上这几天晒的,更难了。”
舞台上换了首歌,前奏的鼓点闷闷的。双儿从后面凑上来,下巴搁在我肩膀上:“你们的大学生涯,打算怎么度过?”
“谈一场恋爱!”小玉立马直起了腰,“你别说,我已经有目标了。”
“你动作这么快?”
他嘿嘿笑了两声,把脸扭向舞台那边,不接话。
雪松坐在双儿旁边,声音很小:“我没什么想法,好好过就够了。”说完仰起头,看月亮。
“我爸交代我了,活动实践得全参加,最好拿遍学校的奖。”我拍了拍膝盖上的草屑,“学习可能不那么好,但这些——”
“砰——”
夜空白了一下。
所有人仰起头。前排有人转过身来望天,脸被烟花映得忽明忽暗。
就在那堆转过来的脸里面,有一双眼睛。
烟花的光落下来,她脸上映了一层红。周围全是灰绿的迷彩,只有那一点嫣红,亮得不像是真的。
Baimiao:Write/Diagnose/Rewrite
A single-chapter novel writing and editing system. Two modes: write a first draft from a detailed outline, or perform layered diagnosis with demo rewriting on existing drafts. Covers eight major techniques: iceberg and white space, conveying emotion through details, using empty scenes to end with emotion, MRU scene scheduling, two-dimensional rhythm diagnosis, hook engagement, special scene scheduling, and more. Distinguishes ❗ hard red lines from ◇ aesthetic suggestions, each with specific alternative writing methods. Suitable for any novelist who wants their writing to have more vivid imagery and lingering aftertaste.
ResearchFilm Narrative Assistant v3
Deeply deconstructs video scripts or course content using film narrative theory. Output includes a complete five-stage analysis: Beat identification (with original text summary), Arrangement strategy identification (using a library of 16 strategies), Narrative execution technique identification (using a library of 7 technique types), Film benchmarking, and Ultimate diagnosis. It then generates a minimalist literary criticism-style HTML visualization report (black background with serif font, single accent color, and ASCII art curves).
WriteVertical IP Consultant: 12Q+5D
Focuses on helping content creators, freelancers, and entrepreneurs achieve systematic positioning of their vertical business IP. Based on a 400-person internal test group, 80 real diagnostic cases, and extensive practical experience plus a methodology distilled from multiple contributors, it provides a complete solution from persona positioning, track selection, content design to MVP validation.
ResearchKnowledge to Skill Packager
Transform any knowledge content (books, courses, methodologies, notes) into a reusable knowledge system. Supports three core capabilities: 1) Package knowledge from scratch into a Framework handbook or Prompt assistant document 2) Convert Framework/Prompt documents into a Skill directory structure 3) Optionally call the Skill Creator to register the Skill directory as a truly callable Skill. Triggers when users mention scenarios such as: packaging knowledge, building a knowledge base, creating a Framework, creating an expert assistant, creating a Prompt assistant, creating a conversational bot, using Engine A, using Engine B, organizing this methodology into something reusable, having AI learn my knowledge, turning a Framework into a Skill, turning a Prompt into a Skill, Skill-ization, converting into a Skill, how to have Claude automatically use my methodology, wanting to put my knowledge base into a Skill system. Also applicable to vague expressions: I have a lot of material I want to turn into reusable knowledge assets, I don't want to repeat explanations every time, help me turn this into a system, how can this methodology be automatically matched, make this md file into a Skill. Even if the user doesn't explicitly mention "packaging" or "Skill", as long as the essence of the problem is "turn knowledge into a reusable system", this Skill is triggered.
ResearchCourse Design Assistant
Based on the 'Five Elements of Course Design' model, through 8-step guided dialogue, even beginners who don't know how to design courses can quickly complete a full course system design from macro to framework level. Supports two modes: Basic (quick decisions) and Professional (full design), suitable for beginners and semi-beginners.
ImageVisual Designer
Based on structured content such as Skill documentation, product introductions, methodology frameworks, etc., automatically extract key information and generate a professional one-page introduction graphic. Supports horizontal/vertical layouts, with AI-recommended visual styles (flat, minimalist business, hand-drawn, etc.). The image generation effect can be referenced from the cover image.

Hand-drawn Illustration Master
Automatically identifies key points based on the full article content and generates hand-drawn style illustrations to enhance the reading experience for long-form articles. Suitable for articles with 3000-5000 words.
Super Conversion Rate Expert
A conversion rate optimization expert system based on Chen Yong's "Super Conversion Rate" methodology. It covers three core capabilities: selling point extraction (3 types of selling points, competitor comparison method, FB copywriting principle, trust design), landing page optimization (4-layer framework diagnosis, first screen 4 elements, visual design, B=MAP button formula, F visual model), and channel strategy (0-1 small to big vs 1-N big to strong, search traffic vs information flow traffic, keyword strategy, audience labels). This skill is triggered when users mention any of the following scenarios: selling point extraction, copy optimization, landing page diagnosis, detail page design, first screen optimization, button design, channel selection, ad strategy, budget allocation, ROI optimization, as well as vague expressions like "I don't know which selling point to focus on," "the page isn't working," "ad spend is wasted," or "ads aren't converting." It is suitable for optimization needs of e-commerce detail pages, SaaS landing pages, education course pages, and service marketing pages. Even if the user does not explicitly mention "conversion rate," as long as the core issue is "how to make the product/page/ad sell better," this skill should be used.

Code-Style Daily Reflections
Write daily reflections like coding: You provide the input (trigger material), AI processes it (by interviewing and organizing), and outputs a three-part daily reflection draft. Rapidly produce 500-1000 words of publishable content in 10-15 minutes.
Wanwei Steel Writing Engine
A narrative writing engine based on Wanwei Steel's cognitive arc, using the main narrative framework of 'make readers feel safe → prove you are right but not enough'. It integrates Mimeng-style story enhancement (scenario-based, story-first, case density) and Liu Run-style concept translation (analogy anchoring, transition precision, paragraph closure) to reorganize the narrative logic of a long draft that has completed style calibration. Delivered in two steps: first, output a restructured outline (with analysis of differences from the original), and after user confirmation, generate the complete restructured article.
Gladwell-Style Narrative Engine
Based on Malcolm Gladwell's nonfiction narrative methodology, it restructures the narrative logic of a long-form draft that has already undergone style calibration. Without changing the author's persona tone, it prioritizes narrative tension by constructing counterintuitive hooks, modeling proxies, implanting interdisciplinary concepts, and using a five-act dramatic arc to transform a straightforward article into a deep long-read with strong suspense, character arcs, and cognitive gaps. Two-step delivery: first output a restructuring outline (with analysis of differences from the original), then generate the full restructured article after user confirmation.

WeChat Article Cover Design
Design a cover image for your WeChat Official Account article based on the content. Supports multiple styles (left-right symmetric layout, movie poster style, AI tech style, etc.), automatically extracts the core information from the article for cover copy, and generates 3 different style covers for selection. Requires the article content or core topic and attribution information.

Knowledge Poster Generator
Enter the details page of your knowledge product, and AI will automatically extract key information to generate marketing posters for social media. Supports intelligent matching of various visual styles.
MIT 48-Hour Speed Learning
Through intelligent material completion and three-dimensional deep questioning, establish an expert-level cognitive framework in any field within 48 hours. Based on the viral learning methodology of MIT graduate students, it helps users quickly jump from "information overload" to "deep understanding".

Judgment Ledger
Record the context, choices, assumptions, and validation logic of every important decision, generating searchable, reusable decision snapshots. When facing similar problems in the future, your judgment framework can be quickly found, reused, and help you avoid repeating mistakes.

Refined Meeting Summary
Reorganize fragmented meeting notes into a clear, logical, and well-structured professional summary. It intelligently recognizes and adaptively extracts key points from various meeting types, making your meeting outcomes truly actionable and traceable.
Business Risk Panorama
Systematically deconstruct new business ideas by completing key elements through structured conversations. Identify risk assumptions across value, growth, and operational dimensions, pinpoint the most critical red-line risks, and design the minimum-cost MVP validation plan. Applicable to business ideas at any stage.

My Content DNA Extractor
From multiple pieces of your content, automatically analyze and extract the underlying "motif system"—those thinking logics, value beliefs, and expression postures that you repeatedly land on without realizing it. Input about 10 articles, output a structured "content DNA" document.
Viral Article Dissection
Extract replicable writing rules from a single viral WeChat public account article. Using the What-Why-If three-layer analysis model, it deeply dissects 8 core dimensions such as cognitive engine, structural design, and rhythm control, outputting structured extraction cards to help content creators quickly learn the creation methodology of benchmark articles and apply it to their own writing.
WriteXiaohongshu Post Generator
Quickly turn product manuals into Xiaohongshu promotional content, and automatically use the 'Xiaohongshu Image-Text Generator' to create matching images.

WeChat Articles Import
Enter one or more links to WeChat Official Account articles. The skill will automatically extract the main content, generate a standardized YAML knowledge base identifier block, and save each article as an independent Note in the current Board, enabling rapid AI scanning and knowledge base retrieval.

Knowledge Model Visualization
Transform knowledge frameworks/theoretical models into high-quality visual images. Users input model name, structure type, and Chinese labels, and the Skill automatically identifies the chart type, constructs professional prompts, and directly generates model card images, solving pain points for knowledge bloggers in Chinese character rendering and knowledge density expression.

Xiaohongshu VP Expert
One-click expert advisor designed for users who want to create virtual products on Xiaohongshu. Once activated, the AI will act as a senior Xiaohongshu virtual product consultant, guiding users through an 8-step structured conversation covering the entire process from direction selection, product evaluation, to action planning. Built-in 'Challenge Mode' allows continuous questioning and quantitative evaluation of the user's initial direction, helping to quickly narrow down the direction and avoid blind investment.

Daily Log
Automatically aggregates your daily scattered notes in YouMind, generating a structured, stylized, and traceable personal log summary, with an optional 'Today's Flash Card' visual card.
WriteStyle Extractor
Analyze your articles like a writing coach. Deeply analyze your writing style, uncover unique taboos, and generate a personalized style card to help you refine your craft.

Xiaohongshu Image Generator
Automatically break down an article and generate 4-6 vertical infographics in Xiaohongshu style. Supports 3 styles: Rich Visual (illustration narrative), Dark Premium (business minimalism), Text Minimal (luxury white space). Automatically analyzes article structure, distinguishes cover pages from content pages, generates matching images in batches, and inserts them at the end of the article.