How to Create a PRD Using the /write-prd Command in PM-Skills

The /write-prd command executes a four-step workflow that transforms a simple feature idea or problem statement into a complete Product Requirements Document by gathering context, invoking the create-prd skill, and outputting an 8-section markdown file.

The phuryn/pm-skills repository provides a structured command system for product management workflows. When you need to create a PRD using the /write-prd command, you interact with an automated workflow defined in pm-execution/commands/write-prd.md that orchestrates the create-prd skill to generate professional documentation.

Understanding the /write-prd Workflow

The command implementation follows a deterministic four-phase process. According to the source code in pm-execution/commands/write-prd.md (lines 20‑108), the workflow progresses from initial input to final document delivery.

Step 1 – Understand the Feature

The command accepts diverse input formats at invocation. Located at lines 20‑27 of the command definition, this phase parses your submission whether it is a feature name, a problem statement, a user request, or an uploaded file (PDF, research doc, or strategy deck).

Step 2 – Gather Context

The system issues conversational prompts to capture critical gaps in your initial input. As defined in lines 31‑39, the command queries for:

• The core user problem
• Target user segments
• Success metrics
• Technical or business constraints
• Prior art and existing solutions
• Scope preferences (MVP vs. full feature)

Step 3 – Generate the PRD

The workflow delegates document generation to the create-prd skill defined in pm-execution/skills/create-prd/SKILL.md. This skill populates an 8-section markdown template:

1. Executive Summary
2. Background
3. Objectives
4. Target Users
5. User Stories
6. Solution Overview
7. Open Questions
8. Timeline

Step 4 – Review & Iterate

After generating the draft, the command presents options for refinement (lines 101‑108). You can tighten scope boundaries, run a pre-mortem analysis, split the PRD into discrete user stories, or generate stakeholder updates. The final document saves as PRD-<feature>.md in your workspace.

Invocation Syntax and Examples

You can trigger the workflow through the PM-Skills interface (CLI or web UI) using several input patterns.

Simple feature description:

/write-prd SSO support for enterprise customers

Problem-oriented phrasing:

/write-prd Users are dropping off during onboarding — we need to fix step 3

Uploaded document workflow:

/write-prd [attach PDF, research doc, or strategy deck]

The 8-Section PRD Structure

When you create a PRD using the /write-prd command, the output follows a standardized template. The create-prd skill ensures each document contains:

## Product Requirements Document: [Feature Name]

**Author**: [Name]  
**Date**: [Date]  
**Status**: Draft  

### 1. Executive Summary

[High-level description of the solution and expected impact]

### 2. Background & Context

[Problem statement, market research, and current state analysis]

### 3. Objectives

[Measurable goals and success criteria]

### 4. Target Users

[User personas and segments affected]

### 5. User Stories

[Format: As a [user], I want [goal], so that [benefit]]

### 6. Solution Overview

[Functional requirements and technical approach]

### 7. Open Questions

[Items requiring stakeholder input or further research]

### 8. Timeline

[High-level milestones and delivery estimates]

Key Source Files

Understanding the underlying architecture helps when customizing or debugging the workflow:

• pm-execution/commands/write-prd.md – Defines the command interface, invocation examples, and the four-step workflow orchestration (lines 20‑108).
• pm-execution/skills/create-prd/SKILL.md – Contains the template engine and section definitions used to render the final markdown document.
• pm-execution/README.md – Provides the execution suite overview and integration context for the /write-prd command.

Summary

• The /write-prd command automates PRD creation through a structured four-step workflow defined in pm-execution/commands/write-prd.md.
• Input flexibility allows plain text descriptions, problem statements, or uploaded documents.
• Context gathering occurs through targeted questions about users, metrics, constraints, and scope.
• The create-prd skill generates an 8-section markdown document (Executive Summary through Timeline).
• Iteration support enables pre-mortems, scope tightening, and user story decomposition before finalizing.

Frequently Asked Questions

What input formats does the /write-prd command accept?

The command accepts any text input including feature names, problem statements, user requests, or uploaded files such as PDFs, Word documents, and strategy decks. The parser handles these flexibly during the "Understand the Feature" phase (lines 20‑27).

How long does the PRD generation process take?

The initial document generation happens immediately after you answer the context-gathering questions. The create-prd skill renders the 8-section markdown template in seconds, though the total time depends on how much iteration you perform during the review phase.

Can I customize the PRD template sections?

The standard template includes eight fixed sections defined in pm-execution/skills/create-prd/SKILL.md. While the command workflow itself is standardized, you can modify the output after generation or fork the skill definition to create custom templates for your organization.

Where is the final PRD saved?

The command saves the completed document as PRD-<feature>.md in your workspace directory. The filename derives from your initial feature description, and the markdown format ensures compatibility with version control systems and documentation platforms.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}