kweaver-ai/skills · Archived

generate-prd

Generate complete Chinese PRD documents from rough product requests.

First seen Mar 23, 2026

Installation

$ npx skills add kweaver-ai/skills --skill generate-prd

Summary

  • Generate complete Chinese PRD documents from rough product requests.
  • Use when the user asks to create, refine, or standardize a PRD, turn ambiguous requirements into an Agile-friendly product requirements document, or fill a PRD template by first asking focused clarification questions and then saving the final markdown file.

Stronger alternatives

This repository is archived — consider an actively maintained alternative.

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from kweaver-ai/skills.

npx skills add kweaver-ai/skills

Browse all from kweaver-ai/skills

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Repository health

Default branch main
Open issues 1
Status Archived

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 5,073 B
  • docs SUMMARY.md 346 B

History

  1. First seen on skills.sh
  2. First recorded snapshot · 2 installs

SKILL.md

Generate PRD

Overview

Turn incomplete product ideas into a complete Chinese PRD that a delivery team can discuss and implement. Use the bundled template at ./assets/prd-template.md as the required output structure and do not change its section order unless the user explicitly asks for a different format.

Workflow

Follow these phases in order. Do not skip a phase unless the user explicitly asks for a partial result.

Phase 1: Clarify Missing Information

  1. Read the user's request and propose a concise requirement name.
  2. Read ./assets/prd-template.md before asking questions.
  3. Compare the request against every template section and identify missing or weak information.
  4. Ask focused clarification questions one by one or in very small batches.
  5. Prioritize gaps that block later sections:

- business background and problem evidence - target users and scenarios - scope boundaries - functional details and business rules - success metrics, release constraints, and dependencies

  1. Stop asking questions when the remaining unknowns can be filled with reasonable, explicit assumptions.
  2. If assumptions are necessary, label them clearly in the final PRD as 待确认 rather than pretending they were provided.

Phase 2: Generate The PRD

  1. Use the exact structure from ./assets/prd-template.md.
  2. Replace every placeholder with real content derived from the user's request, follow-up answers, or explicit assumptions.
  3. Write the entire document in Chinese.
  4. Use professional product language and active voice.
  5. Save the document as <需求名称>-prd.md.
  6. Keep the requirement name short, specific, and file-system safe. Replace illegal path characters with - when needed.

Phase 3: Validate And Improve

  1. Review the PRD section by section.
  2. Remove placeholder text such as 【功能名称】, [待补充], [例如:], [功能1], or [不包含1].
  3. Verify that goals are SMART and measurable.
  4. Verify that functional requirements are actionable enough for design and engineering discussion.
  5. Verify that the tone supports collaboration rather than issuing commands.
  6. Tighten ambiguous wording. Replace weak terms such as 大约, 可能, 尽量, or similarly vague phrasing.
  7. Save the improved version to the same file path.

Questioning Rules

  • Ask only for information that materially improves the PRD.
  • Ask concrete questions tied to missing template fields.
  • Prefer questions that elicit measurable answers, constraints, examples, or rules.
  • When the user gives a broad goal, drill into baseline, target value, and deadline.
  • When the user gives a feature idea without boundaries, ask what is explicitly out of scope.
  • When the user cannot provide data, ask for the best available proxy such as current process time, error rate, ticket volume, or conversion rate.

Writing Rules

  • Keep the template headings and numbering format intact.
  • Preserve markdown readability and concise phrasing.
  • Use bullet lists for scope, rules, risks, and acceptance criteria when that improves clarity.
  • Express objectives with metrics, thresholds, owners, or time bounds whenever possible.
  • Write user stories in the form 作为……,我希望……,从而…….
  • For 需求范围, clearly separate In Scope and Out of Scope.
  • For each functional requirement, include description, user value, interaction flow, business rules, boundary conditions, and exception handling.
  • For 验收标准, write testable Given-When-Then statements.

Quality Bar

Enforce these principles in every output:

  • Data-driven: ground the problem statement in observed behavior, user research, or operational data.
  • SMART goals: make every core objective specific, measurable, achievable, relevant, and time-bound.
  • Concise and clear: prefer precise statements over long explanations.
  • Executable: provide enough detail for implementation and cross-functional review.
  • Collaborative: phrase open questions, dependencies, and assumptions in a way that invites discussion.

Template Handling

  • Always load ./assets/prd-template.md before drafting.
  • Treat the template as mandatory output scaffolding.
  • Update dates to the current working date when the user does not provide one.
  • Fill unknown owner names or status with explicit values such as 待确认 only when the user did not specify them.
  • Do not leave example text or placeholder markers in the final file.

Output Pattern

When responding after writing the file:

  1. State the saved file path.
  2. Summarize the requirement name and any major assumptions.
  3. Call out unresolved items that still need stakeholder confirmation.