liang sun
Multi-Agent Investment Research Team: A-Share Stock Selection and Investment Committee Analysis
It's not an AI assistant, but a virtual investment research team. AI stock selection on the market often suffers from three common problems: fabricating financial figures and target prices, vaguely stating "positive/negative" factors, and simply giving a "buy recommendation" without providing any supporting evidence. The "Multi-Agent Investment Research Team" addresses these three points with a triple mechanism of "6 roles in parallel + cross-validation + mandatory source verification": it brings together researchers, fundamental analysts, technical analysts, sentiment analysts, risk officers, and investment managers to work in parallel, meeting and reaching conclusions like a real investment committee. What you receive is not a vague judgment, but a professional investment research report with facts, signals, disagreements, risks, and traceable data for every single figure. Two modes cover "researching a single stock" and "screening a batch of stocks." Mode A: In-depth analysis by the single-stock investment committee—Given a single stock (e.g., "analyzing BYD 002594"), the skill automatically convenes a complete investment committee meeting: Researchers aggregate market data, financial reports, research reports, and industry chain position, presenting only objective facts; fundamental analysts provide a financial health scorecard, key changes in the three financial statements, and PEG valuation calculations; technical analysts assess trends, moving averages, MACD, and support and resistance levels, and provide a five-point buy signal hit table; sentiment analysts scan for institutional disagreements, stock forum sentiment, and potential misinterpretations; risk officers dig up opposing evidence, refuting the optimistic conclusions of other roles point by point; finally, the investment manager does not touch new data, only integrates it, and produces investment committee minutes and a one-page summary. Mode B - Multi-condition Stock Selection Screening: From your specified range (CSI 300, a specific industry concept sector, or your own stock pool), a three-layer funnel filtering process is used: First, L1 financial hard screening (three consecutive quarters of growth, ample cash flow, PEG < 1 or a significant increase in contract liabilities); then L2 technical timing (platform breakout, golden cross of moving averages, breakout with increased volume, strong pullback with reduced volume, MACD crossing above the zero line); finally, L3 information verification (research report ratings and industry chain logic, eliminating targets with "purely technical analysis without fundamental logic"). After the candidate list is generated, the Top N targets can be automatically connected to Mode A for in-depth analysis. You will receive the following deliverables: Option A: Fixed delivery of the "five-piece set": ① A comprehensive analysis report integrating all six roles; ② Data sources and evidence tables, with each key conclusion corresponding to "Data → Source → Date"; ③ Meeting minutes of the investment committee (topics → viewpoints → disagreements → consensus → variables to be tracked); ④ A risk list sorted by severity (high/medium/low); ⑤ A one-page summary by the investment manager, condensing the core logic, key variables, validation points, and confidence levels. Option B: Delivery of a candidate stock list (code | name | hit criteria | key data | source | trigger date) + explanation of screening criteria and definitions, with the option to optionally include the complete five-piece set for top candidates. All deliverables are saved as files, with filenames including the target and date for easy reuse and archiving.
Professional Course Topology Diagram and Curriculum Program Material Generator
Input a professional course schedule, and the system automatically identifies competency lines and prerequisite relationships, generates a high-definition topology diagram using code, and produces training program description materials; it is fact-anchored with zero fabrication, and the diagram is created only after verification. Input the professional course offerings in any university's talent training program (PDF/image/text table are all acceptable), and the system automatically identifies the topological relationships between courses—grouping them by professional competency lines, identifying strong prerequisite dependencies and competency responses, locating core hub nodes, handling limited elective/specialized courses, and connecting them to the internship in the 5th/6th semester; then, it uses HTML/SVG code to accurately render a high-definition course topology diagram with clear Chinese characters and unblurred lines (not an AI-generated diagram), and finally produces a talent training program curriculum system description material that conforms to the style of teaching documents and includes diagrams. Built-in fact-anchoring and manual verification gate control: course names/nature/class hours are strictly faithful to the original table with zero fabrication; prerequisite relationships are clearly marked as teaching logical inferences and are verified by the user before diagram creation; information such as unfounded curriculum standard adjustments is never fabricated.
Book2Skill — Distill Any Book
Stop taking notes — turn books into callable skills. Turn any book into a set of skills you can actually call inside YouMind. An eight-stage pipeline reads the whole book, extracts its methods, stress-tests them, and registers each as a one-click skill — every one carrying its source, triggers, and boundaries. Distill, never fabricate. Stop collecting reading notes; put the book to work.
AI Personality Distillation: Boss Soul Cloning
Clones any big shot into a conversational AI personality. 4 functional points: (1) Automatic collection of corpus → Pain point: Searching for materials all over the internet and copying and pasting one by one (2) Detailed reading and archiving of each article → Pain point: "AI acts well but only says clichés" (3) Word-for-word preservation of the original words → Pain point: "AI talks nonsense in a serious manner" (4) One-click assembly into a skill → Pain point: Not understanding the technology and not knowing how to implement it. How to use: After opening, just say "Make me an AI version of XXX [e.g., Charlie Munger]", and it will take care of the rest - you only need to nod to confirm at a few key points. Social proof: It has been fully run with [Charlie Munger] - 6 public corpus distilled 25 misjudgment psychology lists, 32 word-for-word original words and an AI Munger that can have a conversation... Five red lines are written in the process to ensure "distillation rather than fabrication" from the mechanism.
Portrait Freedom: Two-Step Reverse Workshop
One reference image generates a unique, high-quality portrait. Many people create AI portraits by uploading a selfie and asking, "Generate a high-quality professional portrait for me." The result is often: the face doesn't resemble the person, the style is generic and generic, the skin looks plastic, and the makeup and clothing are completely mismatched with the individual's personality. The root of the problem is that the AI tries to guess both "what you look like" and "what style you should have," failing at both. This skill breaks this down into two stable steps: First, you provide only one reference image (professional portrait, magazine portrait, black and white film, modern Chinese style…any direction). The AI doesn't rush to generate an image; instead, it breaks down the "photographic language" of the image like a photographer—lighting, composition, lens focal length, background, color tone, atmosphere, texture—and writes two complete sets of prompts for "male" and "female" versions. Clothing, makeup, hairstyle, and pose are all written separately for each gender, fundamentally preventing the disastrous result of "putting a man in a silk suspender top." The second step involves choosing your preferred prompt and pairing it with a clear photo of yourself: the reference image defines the style, while your photo defines your identity—preserve your face shape, features, age, and expression, then apply the professional photographic quality of the reference image. The result is a portrait that is instantly recognizable and has a high-quality, professional feel. The reference image controls the style, and your photo defines your identity—it's that simple. Changing the reference image allows you to change your style: ID photos, professional images, magazine covers, black and white film—all done in one process, completely eliminating the homogenization and artificiality of AI portraits. Suitable for: Professionals, social media influencers, job seekers, and anyone who wants a high-end portrait that is uniquely theirs, requiring a professional headshot/ID photo/personal brand portrait.
An eight-person editorial team—an AI-driven expert writing system
An AI-powered collaborative writing system for self-media bloggers, writers, educators, and researchers. It features a streamlined team of 8 virtual experts (style appraiser, topic planner, research analyst, knowledge manager, chief writer, fact checker, editor-in-chief, and layout designer) who collaborate throughout the entire process, from style definition to final draft layout, producing high-quality content that is in-depth, logically sound, and stylistically appropriate.
Automated generation system for Chinese patent technology disclosure documents
The entire process from project documents to deliverable technical disclosure documents includes: patent point mining, online novelty search (prioritizing the State Intellectual Property Office + Google Patents/Scholar dual channels), anonymization and document drafting, and a self-inspection closed loop. It is compatible with the YouMind platform toolchain, automatically scanning Kanban materials, outputting disclosure documents and independent novelty search reports, and supporting generateDiagram generation, iterative revisions, and version management. It incorporates P1-P7 core protocols (factual integrity/automatic anonymization/version protection/confirmation access control/self-inspection isolation/novelty search location/social science context).
National Social Science Topic Incubation
This program helps young humanities and social science teachers who are applying for the National Social Science Fund for the first time to start from a vague research direction and, through a four-stage dialogic incubation process (direction analysis → theoretical transformation → initial draft of argumentation → self-checking and optimization), ultimately produce high-quality topics and initial drafts of argumentation that are suitable for application.
Workflow for creating tutorials and knowledge manuals
A one-stop book generation engine from research to publication. It supports in-depth research, intelligent outlines, AI illustration generation, and professional typesetting, ultimately outputting complete book documents or web pages. Suitable for creating content products such as knowledge manuals, tutorials, and brochures.
Distill a book into a set of executable skills
Distill a book into a set of executable skills that can be directly called within YouMind. An eight-stage sequential pipeline: full book reading → five-category extraction → triple verification → RIA++ specification → relationship networking → red team load testing → one-click registration. Five red lines ensure "distillation, not fabrication," with each skill including its original source, trigger conditions, and usage boundaries. Say goodbye to fragmented reading notes and truly integrate a book into your workflow.
Collector's Edition City Promotional Poster - LIANG DESIGN
Based on the "NAICHA DESIGN · Collector's Edition Poster" creation guidelines, the themes of city/object/abstract are translated into exhibition-grade posters with the texture of thick oil painting, master-level serif typography, diagonal composite color band main framework, and hot-stamped large lettering. v1.1 Iteration: Strengthen the main visual anchor point + three-level font hierarchy + Chinese main character pinyin secondary characters.
Vocational College Thesis Review Assistant: Internet Marketing and Live Streaming E-commerce Major
This thesis review tool is specifically designed for vocational college students majoring in online marketing and live-streaming e-commerce. From the perspective of an associate professor, it provides in-depth review in five steps: "Formal Compliance → Topic Abstract → Main Body Structure → Argumentation Data → Academic Norms and Overall Evaluation." Each step generates a structured scoring sheet and a list of professor's annotations, culminating in a complete scoring sheet and a warm, personalized feedback letter from the professor. It incorporates compliance standards such as the 2026 Ministry of Education's new national standard (GB/T 7713.1—2026), plagiarism rate ≤30%, and AIGC rate ≤25%. It also provides precise error correction for key practical aspects of live-streaming, including product selection, scriptwriting, traffic generation, GMV, and private domain management. Suitable for phased reviews of drafts (first draft, second draft, and final draft).
