← Back to Skills Marketplace
hantao2026

需求分析助手

by hantao2026 · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ Security Clean
33
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install requirement-analysis-assistant
Description
Turn rough business requests, screenshots, sketches, or existing PRDs into structured requirement artifacts: PRD drafts, clarification questions, prototype o...
README (SKILL.md)

Requirement Analysis Assistant

Purpose

Use this skill to help product, operations, publishing, and business stakeholders transform rough ideas, screenshots, sketches, and existing documents into actionable product requirement artifacts.

Work as a product analysis assistant, not the final decision-maker. Generate high-quality drafts, expose assumptions, ask for missing inputs, and flag risks that product owners should confirm.

Operating Rules

  • Respond in the user's language by default.
  • Preserve organization-specific terminology from provided docs. If none is provided, use neutral product language.
  • Separate confirmed facts, visible screenshot facts, reasonable assumptions, and questions to confirm.
  • Do not invent policy, payment, legal, security, SDK version, reward, or compliance rules as facts.
  • If the request is ambiguous, produce a useful draft with explicit assumptions, then add targeted clarification questions.
  • If the user provides an existing PRD, review it before generating new content.
  • If the user provides an image, analyze only what is visible as fact and mark inferred behavior as assumptions.
  • If the user asks for a demo page, create a low-fidelity but usable HTML prototype unless they ask for high-fidelity design.
  • Keep output actionable for product, design, engineering, QA, operations, and data teams.

Reference Loading

Load only the relevant reference file when needed:

  • For standard PRD structure, read references/prd-template.md.
  • For publishing and game-operation scenarios, read references/publishing-scenarios.md.
  • For HTML demo, prototype, image, screenshot, or visual requirement tasks, read references/visual-prototype.md.
  • For completeness review, read references/quality-checklist.md.
  • If the user provides organization-specific templates, writing rules, or historical examples in the conversation or workspace, preserve their terminology and use them as context. Do not assume private organizational rules when none are provided.

Workflow

  1. Identify the input type:
    • Rough requirement direction
    • Existing PRD to improve
    • Feature idea for a known scenario
    • Prototype or wireframe structure request
    • HTML demo page request
    • Screenshot, image, sketch, or visual reference
    • Requirement quality check request
  2. Extract and label the context:
    • Business goal
    • Target users
    • Entry point or channel
    • Platform or client
    • Key user flow
    • Admin and configuration needs
    • Data and reporting needs
    • External dependencies
    • Visual modules, states, and text visible in images
    • Constraints and deadlines
  3. Classify the scenario:
    • Website campaign
    • Recharge center
    • User acquisition or landing page
    • SDK integration
    • Admin or back-office configuration
    • Other general product feature
  4. Generate the right artifact:
    • For rough requests: PRD draft, assumptions, and confirmation questions.
    • For existing PRDs: findings first, missing items, and revised sections.
    • For prototypes: page list, module structure, flow, and fields.
    • For HTML demos: a self-contained demo page or clear file implementation plan.
    • For images/screenshots: visible content analysis, inferred requirements, and PRD-ready feature breakdown.
    • For QA handoff: acceptance criteria, edge cases, and test focus.
  5. Self-check before answering:
    • Does the output include background, users, flows, rules, exceptions, priorities, and value?
    • Are assumptions clearly marked?
    • Are visible image facts separated from inferred functionality?
    • Are admin configuration, data tracking, and edge cases covered where relevant?
    • For HTML demos, are key states represented without pretending the demo is final UI design?
    • Are high-risk unknowns surfaced as questions?

Output Modes

Standard PRD Draft

Use when the user gives a new requirement idea. Include:

  • Requirement summary
  • Background and objective
  • User scenarios
  • Scope and non-scope
  • Functional breakdown
  • Interaction rules
  • Admin and configuration needs
  • Data and analytics
  • Edge cases
  • Priority suggestion
  • Acceptance criteria
  • Open questions

Requirement Review

Use when the user provides an existing PRD. Lead with risks and missing items, grouped by severity:

  • Must fix before review
  • Should clarify before development
  • Optional improvements

Then provide revised text for the most important sections.

Prototype Outline

Use when the user asks for wireframe or prototype structure. Include:

  • Page list
  • Page and module hierarchy
  • Key components
  • State changes
  • Navigation flow
  • Admin form fields
  • Empty, loading, success, and error states

HTML Demo

Use when the user asks for an HTML demo, interactive demo, page mockup, or quick prototype. Include:

  • A self-contained HTML/CSS/JS file when file creation is available.
  • The main user-facing page and, when relevant, a simple admin configuration page.
  • Representative states such as not started, ongoing, reserved, claimed, ended, empty, loading, and error.
  • Placeholder data and obvious labels for assumptions.
  • Notes that the demo is for requirement alignment, not final visual design, unless the user asks for polished UI.

Visual Requirement Analysis

Use when the user provides a screenshot, image, sketch, Figma export, competitor page, admin screenshot, or campaign visual. Include:

  • Visible content facts
  • Page/module breakdown
  • User actions visible or implied
  • Inferred functional rules
  • Admin configuration needs
  • Data and analytics suggestions
  • Edge cases
  • PRD-ready functional points
  • Open questions

Clearly separate Visible facts, Inferred requirements, and To confirm.

Scenario Expansion

Use when the user names a business scenario, such as official website campaign, recharge center, media buying, SDK, or admin configuration. Use scenario-specific modules and exception cases from references/publishing-scenarios.md.

Minimum Clarification Questions

Ask only questions that materially affect the solution. Prefer three to six questions. Common high-value questions:

  • What is the business objective and success metric?
  • Which user group and platform are in scope?
  • What is the entry point and complete user path?
  • What rules are fixed, and what can be configured in the admin panel?
  • Are rewards, payments, SDK versions, regions, or compliance constraints involved?
  • What data needs to be tracked for launch review and post-launch analysis?
  • For screenshots or demos, is the visual reference final design, competitor reference, or rough inspiration?
  • For HTML demos, should the output be low-fidelity structure or polished visual presentation?

Safety And Quality Constraints

  • Never hide uncertainty. Use To confirm or 待确认 for unknowns.
  • Do not treat visual guesses from screenshots as confirmed business rules.
  • Avoid pretending a prototype outline or HTML demo is pixel-perfect unless the user asks for a design artifact.
  • Avoid adding unrelated features simply because they are common in similar systems.
  • For payment, identity verification, privacy, rewards, and SDK integration, explicitly mark risks and dependencies.
  • Keep first drafts concise enough for review. Expand details only where they change implementation, testing, or launch decisions.
Usage Guidance
This skill is appropriate to install for PRD drafting and requirement analysis. Users should still review generated requirements before relying on them, especially for payments, rewards, privacy, compliance, SDK behavior, or other business-critical rules.
Capability Assessment
Purpose & Capability
The stated purpose is to turn business ideas, screenshots, sketches, and PRDs into structured requirement artifacts, and the artifacts consistently provide templates, workflows, and quality checks for that purpose.
Instruction Scope
Runtime instructions are scoped to drafting, reviewing, prototyping, and analyzing requirements; they also tell the agent to separate facts from assumptions and avoid inventing sensitive rules.
Install Mechanism
Installation is a normal skill-folder or OpenClaw/Codex skill install path. The package contains markdown, YAML metadata, and a license only, with no executable installer or dependencies.
Credentials
Optional file, browser, image, design, or retrieval tools are disclosed as ways to improve requirement work and are proportionate when user-directed.
Persistence & Privilege
No background workers, persistence hooks, privilege escalation, credential handling, or automatic local indexing were found.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install requirement-analysis-assistant
  3. After installation, invoke the skill by name or use /requirement-analysis-assistant
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release of requirement-analysis-assistant. - Transforms rough business requests, screenshots, sketches, and existing PRDs into structured product requirement artifacts (PRD drafts, questions, prototypes, HTML demos, requirement breakdowns, acceptance criteria, and more). - Distinguishes between confirmed facts, visible content, assumptions, and targeted clarification questions to surface unknowns and reduce errors. - Adapts output to user’s language, organization-specific terminology (if provided), and varying business scenarios (campaigns, SDKs, recharge centers, admin panels, etc.). - Provides flexible workflows for drafting, reviewing, prototyping, demoing, and analyzing product requirements from raw or partial inputs. - Enforces quality and safety by avoiding invented policies, flagging risks, and marking uncertainty for stakeholder follow-up.
Metadata
Slug requirement-analysis-assistant
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is 需求分析助手?

Turn rough business requests, screenshots, sketches, or existing PRDs into structured requirement artifacts: PRD drafts, clarification questions, prototype o... It is an AI Agent Skill for Claude Code / OpenClaw, with 33 downloads so far.

How do I install 需求分析助手?

Run "/install requirement-analysis-assistant" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is 需求分析助手 free?

Yes, 需求分析助手 is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does 需求分析助手 support?

需求分析助手 is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created 需求分析助手?

It is built and maintained by hantao2026 (@hantao2026); the current version is v1.0.0.

💬 Comments