In-Depth Research with Horizontal-Vertical Analysis

Methodology Origin The Horizontal-Vertical Analysis was proposed by digital life Khazix, integrating core ideas from Saussure's diachronic-synchronic analysis in linguistics, longitudinal-cross-sectional research design in social sciences, business school case study methods, and competitive strategy analysis. It forms a universal research framework applicable to products/companies/concepts/people. The core principle remains unchanged: pursue temporal depth in vertical analysis, pursue concurrent breadth in horizontal analysis, and finally converge to form judgments.

Preparations

Environment Preparation

1. Confirm PDF conversion script is available: This Skill comes with

   scripts/md_to_pdf.py

   (based on WeasyPrint), which is used to convert the final Markdown report into a beautifully formatted PDF. Ensure dependencies are installed:

   pip install weasyprint markdown --break-system-packages

   .
2. Writing Style: This Skill has a built-in complete writing style guide (see the "Writing Style" section below), no need to load additional skills.

Clarify Research Object

1. Research Object: Specific product name/company name/concept name/person name
2. Type: Product, company, concept, person, or other?
3. Research Motivation (optional): Why research it? What recent events have occurred?
4. Special Focus Areas (optional): Are there any specific directions to delve into?

Step 1: Online Information Collection

Parallel Search Strategy

• Sub-Agent 1 — Vertical Information: Research the origin of the object, background of the founders, development history, key events, version iterations, financing, strategic shifts, and crises
• Sub-Agent 2 — Horizontal Information: Identify competitors, characteristics and user reputation of each competitor, industry comparative reviews, market share
• Sub-Agent 3 (only for complex objects): Supplementary information, such as in-depth background of founders, changes in industry environment, user community discussions (GitHub issues, Reddit, Twitter/X, Zhihu, etc.)

You need to obtain information online. Use the following tools:

• WebSearch: Used to discover information sources and obtain summary and keyword results
• WebFetch: Used to extract content from specific URLs when the URL is known
• If the web-access skill is installed in the user's environment (check if

  /mnt/.claude/skills/web-access/SKILL.md

  exists), prioritize loading it and following its guidelines, as it provides stronger browser CDP capabilities
• Search strategy: First use WebSearch to discover information sources and clues, then use WebFetch for in-depth extraction after finding specific URLs
• Search multiple times with combinations of keywords, do not give up after just one search
• Primary sources are better than secondary sources: Official blog > Original reports from authoritative media > Reproduction/aggregation
• Must check arxiv for academic research objects: If the research object involves academic concepts, algorithms, AI models, technical paradigms, etc., must obtain relevant papers through the arxiv API. Call method:

  curl -s "https://export.arxiv.org/api/query?search_query=all:keyword1+AND+all:keyword2&max_results=10"

  , or access the same URL using WebFetch. Returns XML format, including title, author, abstract, publication date, PDF link. Adjust keyword combinations and result quantity as needed. After finding key papers, use WebFetch to read the paper page (

  https://arxiv.org/abs/paper-ID

  ) to get more details.

Information Source Priority

export.arxiv.org/api/query

Information Sufficiency Self-Check

• Vertical: Can you tell a complete story? Are there obvious information gaps?
• Horizontal: Is the competitor list complete? Have major players been missed? Is there enough information about each competitor for comparison?
• Sources: Are key facts supported by reliable sources? Are there any judgments based solely on a single source?

Step 2: Vertical Analysis (Diachronic / Longitudinal)

Content Requirements

Length

Step 3: Horizontal Analysis (Synchronic / Cross-sectional)

First Judge Competitor Situation

Comparison Dimensions

Length

Step 4: Horizontal-Vertical Intersection Insights

1. How history shaped the current competitive position: Which decisions and events in the vertical history determined its current position in the horizontal comparison?
2. Vertical comparison of competitors: If major competitors are also placed on the timeline, what are the differences in their origins and evolution paths? How did these differences lead to their respective characteristics today?
3. Historical roots of advantages: Can each core advantage today be traced back to a specific node or decision in history?
4. Historical roots of disadvantages: Can each core disadvantage today be traced back to a historical decision? Did any "good decisions" in the past become burdens today?
5. Future deduction: Based on vertical trends and horizontal competitive landscape, propose three scenarios—the most likely, the most dangerous, and the most optimistic—each with logical support.

Length

Writing Style

Core Elements Borrowed from Khazix's Writing Style

Elements Not Borrowed from Khazix's Writing Style

• Excessive colloquialism: The report can have a conversational tone, but do not fill it with phrases like "this thing" "dude" "so awesome". Occasional embellishments are acceptable, but the density should be much lower than in official account articles.
• Abandonment of subheadings: Official account articles pursue a smooth flow without subheadings. Research reports are different; without a clear structure and navigation, readers will get lost in 10,000-30,000 words. Reports need clear chapter structures.
• Punctuation restrictions can be relaxed: Official account articles prohibit colons and dashes. Research reports can use them normally because reports prioritize information transmission efficiency. However, the usage habit of "" can be retained.
• Fixed tail: Do not add the "like/follow/star" tail of official accounts.

Absolute Forbidden Zones (Still Applicable)

• Clichés: "First... Second... Finally", "To sum up", "It is worth noting that", "It is not difficult to find"
• Empty adjectives: "empower", "starting point", "create a closed loop"
• Textbook openings: "In the era of rapid AI development", "With the continuous progress of technology"
• High-frequency problematic words: "To put it bluntly", "What does it mean?", "This means", "Essentially", "In other words", "Undeniably"
• Vague tool names: Do not say "AI tool" "a certain model"; use specific names
• Fabricated scenarios: If certain information cannot be found, honestly mark "This information is temporarily unavailable", never fabricate it

Write in Plain Language

Step 5: Generate PDF Report

scripts/md_to_pdf.py

Conversion Process

1. Complete the Markdown manuscript first: Write the complete report in standard Markdown format and save it as

   [研究对象]_横纵分析报告.md

2. Install dependencies (if not installed):

   pip install weasyprint markdown --break-system-packages

3. Run the conversion script:
   bash

   python [skill目录]/scripts/md_to_pdf.py input.md output.pdf --title "研究对象名称" --author "数字生命卡兹克"

4. The script will automatically generate an intermediate HTML file (for debugging) and the final PDF

Built-in Typesetting Specifications of the Script

md_to_pdf.py

• Page: A4, margins 25mm top / 20mm left and right / 20mm bottom
• Cover page: Automatically generated, including title (28pt dark blue), subtitle "横纵分析法深度研究报告", author information, decorative divider
• Color scheme: H1 title=#1a5276 dark blue, H2=#1e8449 green, H3=#2e86c1 light blue, H4=#5b2c6f purple, body text=#2c3e50 dark gray
• Font: CSS fallback chain

  "Droid Sans Fallback", Helvetica, Arial, sans-serif

  , automatically handles mixed Chinese and English typesetting
• Body text: 10.5pt, line spacing 1.75, justified, orphan/widow control
• Quote block: 3pt dark blue left vertical line + light gray background
• Table: Full width, dark blue header with white text, zebra stripe rows
• Header: "Report Title | 横纵分析法深度研究报告" (not displayed on the first page)
• Footer: "Page X" (not displayed on the first page)
• The first H1 in Markdown will be automatically extracted as the cover title and will not appear repeatedly in the main text

Markdown Writing Notes

• Use

  # Title

  as the report title on the first line (will be automatically used for the cover)
• Immediately after the title, you can write meta-information lines in the format

  > 研究时间：... | 所属领域：... | 研究对象类型：...

  , which will be extracted to the cover
• Use

  ##

  as the main chapter title (Vertical Analysis, Horizontal Analysis, Horizontal-Vertical Intersection, etc.)
• Use

  ###

  and

  ####

  as sub-chapters
• Use standard Markdown table syntax for tables
• Use

  >

  syntax for quotes
• Use

  **text**

  for bold

End Content

• Information Sources: List of all cited sources, marked with URL and access time
• Methodology Description: Briefly explain the origin of the Horizontal-Vertical Analysis (1-2 sentences are sufficient)

Report Structure Template

封面页

目录

一、一句话定义
[用一句话说清楚这个东西是什么]

二、纵向分析：从诞生到当下
[完整的纵向叙事，6000-15000字]

三、横向分析：竞争图谱
[横向对比分析，3000-10000字]

四、横纵交汇洞察
[交叉分析和未来推演，1500-3000字]

五、信息来源
[所有引用的来源列表]

File Naming and Delivery

[研究对象名称]_横纵分析报告.pdf

Adaptation for Different Types of Research Objects

Overall Length Overview

Quality Inspection Checklist

• Is the vertical axis a narrative story? Does it have causal logic and contextual context? Is it not a chronological list?
• Is the background and motivation of the founder/initiator sufficiently in-depth?
• Each key node is fully expanded, no important details are skipped to compress the length?
• Is the decision logic restored? Not just "what happened", but also "why this choice was made"?
• Is the competitor scenario judgment in the horizontal axis correct (A/B/C)? Is the competitor analysis in-depth enough?
• Does the user reputation section cite real user voices? Not just official propaganda?
• Does the horizontal-vertical intersection produce new judgments, not an abbreviation of previous content?
• Are the three scenarios in future deduction supported by logic?
• Does the writing style have rhythm and readability? Is it not a cold consulting report?
• None of the items in the absolute forbidden zones are violated?
• All key facts are marked with information sources?
• Information that cannot be found is honestly marked as "temporarily unavailable", no fabrication?
• Is the PDF typesetting beautiful, structure clear, and readable?
• Is the total word count within the range of 10,000-30,000 words? ",