hv-analysis

Overview of hv-analysis skill

hv-analysis is a Chinese-language research skill for turning a product, company, concept, technology, or person into a structured horizontal-vertical analysis report. It is best for users who need more than a definition: they want a credible way to study what something is, how it evolved, how it compares with peers, and what that means now. The hv-analysis skill is especially useful when the goal is to produce a polished PDF research report rather than a quick summary.

What hv-analysis is for

The real job-to-be-done is “help me understand this thing deeply and systematically.” hv-analysis guides a dual-axis workflow: the vertical axis traces the full life cycle over time, and the horizontal axis compares the target against competitors or similar cases at the present moment. That makes it a strong fit for product analysis, company research, market studies, and strategic review.

What makes it different

Unlike a generic deep-research prompt, hv-analysis is built around a specific analytical structure and a final deliverable. It pushes the agent to gather history, map milestones, compare alternatives, and synthesize findings into a report-ready format. That structure helps reduce shallow “what it is” answers and encourages decision-grade analysis.

Best-fit users and use cases

Use hv-analysis when you need to understand positioning, evolution, competitive context, or strategic tradeoffs. It is a good fit for people researching a new tool before adoption, evaluating a company’s trajectory, or preparing an internal briefing. It is less suitable for simple glossary questions or short-form content tasks.

How to Use hv-analysis skill

Install and trigger hv-analysis

Use the skill in your skill directory workflow, then invoke it with a clear research object and intent. A practical install command is: npx skills add KKKKhazix/khazix-skills --skill hv-analysis. In your request, name the target directly and state that you want a horizontal-vertical analysis, for example: “Use hv-analysis to research Notion as a productivity product and compare it with Obsidian and Evernote.”

Give the skill the right input shape

Strong inputs are specific, bounded, and decision-oriented. Include the subject, type, audience, and any angle that matters:

• “Analyze Cursor as an AI coding tool for a founder choosing a dev workflow.”
• “Research OpenAI as a company, with emphasis on product shifts and competitive pressure.”
• “Use hv-analysis for Data Analysis on Figma’s evolution and market position.”

If you only say “帮我研究一下XX,” the skill can still work, but clearer scope produces better comparisons and a cleaner narrative.

Read these files first

Start with SKILL.md to understand the workflow and report structure. Then inspect references/schema.json for the analysis fields the skill expects, and scripts/md_to_pdf.py if you want to understand the final Markdown-to-PDF output path. Those files show what the skill prioritizes: structured research, milestone logic, and presentation quality.

Workflow tips that improve output

For hv-analysis usage, ask for both axes explicitly: history first, then current comparison. If your target has obvious competitors, name them. If you care about a specific lens—pricing, product maturity, ecosystem, or strategic risk—say so up front. When the first draft is too broad, narrow the object or the market slice rather than asking for “more detail” everywhere.

hv-analysis skill FAQ

Is hv-analysis only for Chinese prompts?

The repository content is Chinese-heavy, but the hv-analysis skill can still be used with English requests if the research object and comparison set are clear. The output workflow matters more than the language of the prompt.

What does hv-analysis do better than a normal prompt?

A normal prompt often produces a summary. hv-analysis is designed to produce a research method: vertical history, horizontal comparison, and a final synthesis. That structure is the main reason to install the hv-analysis skill instead of relying on ad hoc prompting.

When should I not use hv-analysis?

Do not use hv-analysis for a simple definition, a headline summary, or a short social post. If you only need “what is X,” this is too much process. It is also a poor fit when you have no real research object yet or when you do not need comparative analysis.

Is hv-analysis beginner-friendly?

Yes, if you can name the thing you want to study and explain why you care. You do not need research methodology knowledge to use it well. The main beginner mistake is asking for a broad topic with no decision context, which weakens the comparison and the final judgment.

How to Improve hv-analysis skill

Start with a sharper research question

The hv-analysis skill works best when the first prompt contains a question, not just a topic. Compare “Research Anthropic” with “Research Anthropic as an AI company, focusing on product strategy, model positioning, and competitor pressure.” The second version gives the analysis a clearer center and better output quality.

Provide better comparison targets

Horizontal analysis gets stronger when you specify peers that matter to your decision. If you are evaluating a product, list the alternatives you would actually choose between. If you leave the comparison set vague, hv-analysis may choose sensible but not always decision-relevant benchmarks.

Control scope to avoid shallow output

The most common failure mode in hv-analysis is trying to cover too many angles at once. If the target is large, constrain the period, market, or use case. For example, ask for “the last three years,” “consumer adoption,” or “developer tooling” so the report stays focused and evidence-rich.

Iterate from structure, not just prose

After the first output, improve hv-analysis by tightening gaps in the vertical timeline, adding a missing competitor, or asking for a stronger synthesis of tradeoffs. If a section feels generic, tell the skill what to emphasize: business model, technical evolution, user adoption, or strategic risk. That is usually more effective than asking it to “make it longer.”