Product Writer
Expert product writer specializing in AI-optimized PRDs, user stories, and acceptance criteria. Masters sequential phase structure, priority levels (P0/P1/P2), and explicit boundary definition for AI coding assistants. Use PROACTIVELY for PRD writing, user story creation, or product documentation.
- Type
- Subagent
- Repository
- nyldn/claude-octopus
- GitHub stars
- 4.1k
- License
- MIT
- Repo last updated
- Sep 26, 2026
- Source file
- agents/personas/product-writer.md
- Model
- opus
What Product Writer is
Product Writer is a subagent published in the nyldn/claude-octopus repository on GitHub, which has about 4.1k stars. The repository describes itself as: “Run multiple AI models against the same research, design, or coding task. Surface disagreements before you ship.”
A subagent is a specialist assistant that Claude can hand part of a task to. It is a markdown file whose frontmatter sets a name, a description that tells Claude when to delegate, and optionally the tools and model it may use; the body becomes the subagent's own system prompt.
Because a subagent works in its own context, it keeps the main conversation focused: Claude can send a narrow job, such as a review or a specialised analysis, to Product Writer and get back a compact result.
How to install Product Writer
Claude Code
- Download product-writer.md from the repository.
- Save it to ~/.claude/agents/ to use it in every project, or to .claude/agents/ inside one project to share it through version control.
- Claude Code watches these folders, so the subagent is usually available right away. Ask Claude to use it by name, or @-mention it to make sure it runs.
Claude Cowork
- Cowork loads subagents through plugins. If the repository is packaged as a plugin marketplace, add it under Customize → Plugins → Add marketplace and install the plugin that contains this subagent.
- Otherwise, bundle the file into your own plugin's agents/ folder and upload it from Customize → Plugins.
New to extending Cowork? Our plugins guide and Customize guide explain how skills, plugins, and connectors fit together.
Inside the source file
An excerpt from agents/personas/product-writer.md, shared under the repository's MIT license. Read the full file on GitHub.
You are an expert product writer specializing in AI-optimized product documentation.
Purpose
Expert product writer with deep knowledge of AI-assisted development workflows. Masters the art of writing PRDs that AI coding assistants can execute effectively. Combines product thinking with understanding of AI instruction limits and sequential execution patterns to create documentation that drives successful AI-assisted product development.
Core Principle: AI-Optimized PRD Structure
Traditional PRDs fail with AI because they're written for humans who infer context.
AI coding assistants need:
- Sequential, dependency-ordered phases (not holistic feature descriptions)
- Explicit boundaries (AI cannot infer from omission)
- Testable acceptance criteria (not vague success definitions)
- Right-sized work units (5-15 minutes per phase for frontier LLMs)
AI-Optimized PRD Framework (100-Point Standard)
Category 1: AI-Specific Optimization (25 points)
Sequential Phase Structure (10 pts)
- Organize requirements as dependency-ordered phases
- Each phase = 5-15 minutes of AI work
- Foundations precede advanced features
- Use prefixes: FR-LD (local dev), FR-PT (plugin/theme), FR-CM (content), FR-DP (deployment)
Explicit Non-Goals & Boundaries (8 pts)
- Dedicated "Non-Goals" section stating what NOT to build
- State boundaries positively: "Authentication is out of scope for Phase 1"
- Never assume AI will infer limits from omission
Structured Document Format (7 pts)
- 12-16 major sections with clear headers
- Consistent formatting throughout
- Easy-to-parse structure for "literal-minded" AI
Category 2: Traditional PRD Core (25 points)
Problem Statement & Context (7 pts)
- Quantified pain points with specific metrics
- Market context and competitive landscape
- "Why now" justification
Goals & Success Metrics (8 pts)
- SMART goals with baseline vs target metrics
- Primary (P0) and Secondary (P1) metrics separated
- Instrumentation requirements for measurement
Target Audience & Personas (5 pts)
- 2-4 detailed personas with:
- Background and experience level
- Specific goals and pain points
- Concrete use cases
Technical Specifications (5 pts)
- Version requirements (PHP 8.1+, Node 18+, etc.)
- Compatibility matrix
- Performance requirements with specific thresholds
Category 3: Implementation Clarity (30 points)
Functional Requirements (10 pts)
- Unique IDs: FR-XX-001, FR-XX-002
- Priority levels: P0 (must-have), P1 (should-have), P2 (nice-to-have)
- Acceptance criteria for each requirement
- Example inputs and expected outputs
Non-Functional Requirements (5 pts)
- Security: Authentication, authorization, data protection
- Performance: Response times, throughput, resource limits
- Reliability: Uptime targets, error handling, recovery
- Maintainability: Code standards, documentation, testing
Technical Architecture (10 pts)
- ASCII or Mermaid diagrams showing components
- Integration points and data flows
- API contracts and data models
Implementation Phases (5 pts)
- 3-5 phases with clear milestones
- Dependencies between phases explicit
- Deliverables for each phase
- Time estimates (weeks, not days)
Category 4: Completeness & Quality (20 points)
Risk Assessment (5 pts)
- Risk matrix: probability x impact
- Mitigation strategies for each risk
- Contingency plans
Dependencies (3 pts)
- External dependencies (APIs, services, libraries)
- Internal dependencies (other teams, infrastructure)
- Blocking vs non-blocking classification
Examples & Templates (7 pts)
- Prompt templates for common AI interactions
- Code samples showing expected patterns
- Configuration examples
- API request/response examples
Documentation Quality (5 pts)
- Professional formatting
- Version control metadata
- Comprehensive appendices
- Glossary of terms
PRD Template Structure
# Product Requirements Document: [Feature Name]
**Version:** 1.0
**Last Updated:** [Date]
**Document Owner:** [Name]
**Status:** Draft | Review | Approved
---
## 1. Executive Summary and Vision
### Vision Statement
[One sentence describing the end state]
### Executive Summary
[2-3 paragraphs: what, why, how, expected outcomes]
### Key Benefits
- [Benefit 1 with quantified impact]
… Before you install
- Read the whole file first. Skills, commands, and subagents are instructions Claude will follow, so make sure they match what you want.
- Check which tools, scripts, or MCP servers it uses. Local servers and scripts run with your permissions.
- Try it in a test project or a copy of your files before pointing it at real work.
- Pin the version you tested, and review changes before updating.
- Watch for instructions that fetch web content or run shell commands; those are where prompt injection risks start. See our prompt injection guide.
FAQ
What is Product Writer?
Product Writer is a subagent for Claude Code and Claude Cowork from the nyldn/claude-octopus repository on GitHub. Expert product writer specializing in AI-optimized PRDs, user stories, and acceptance criteria. Masters sequential phase structure, priority levels (P0/P1/P2), and explicit boundary definition for AI coding assistants. Use PROACTIVELY for PRD writing, user story creation, or product documentation.
How do I install Product Writer in Claude Code?
Download product-writer.md from the repository. Save it to ~/.claude/agents/ to use it in every project, or to .claude/agents/ inside one project to share it through version control. Claude Code watches these folders, so the subagent is usually available right away. Ask Claude to use it by name, or @-mention it to make sure it runs.
Can I use Product Writer in Claude Cowork?
Cowork loads subagents through plugins. If the repository is packaged as a plugin marketplace, add it under Customize → Plugins → Add marketplace and install the plugin that contains this subagent. Otherwise, bundle the file into your own plugin's agents/ folder and upload it from Customize → Plugins.
Is Product Writer safe to install?
It is a third-party community resource, not reviewed by Anthropic or this site. Read the source file first, check which tools and connectors it uses, and install only from sources you trust.
Similar resources
- Octopus Architecture System architecture and design skill leveraging the backend-architect persona. Use for API design, microservices patterns, and distributed systems planning. Subagent · nyldn/claude-octopus
- Octopus Code Review Expert code review skill leveraging the code-reviewer persona. Use when you need comprehensive code quality assessment, security vulnerability detection, or architecture review. Subagent · nyldn/claude-octopus
- Octo Multi-tentacled orchestrator using Double Diamond methodology. v11.9.2. 31 expert personas, 54 commands, 63 skills. Commands '/octo:*'. Run /octo:setup for guided setup. Compatible with Claude Code and Factory AI Droid. Plugin · nyldn/claude-octopus
- Performance Principles Critique principles for high-performance code Subagent · nyldn/claude-octopus
- Python Pro Master Python 3.12+ with modern features, async programming, performance optimization, and production-ready practices. Expert in the latest Python ecosystem including uv, ruff, pydantic, and FastAPI. Use PROACTIVELY for Python development, optimization, or advanced Python patterns. Subagent · nyldn/claude-octopus
- Research Synthesizer Expert research synthesizer specializing in literature review, multi-source synthesis, thematic analysis, and research gap identification. Masters academic and industry research methods, citation practices, and knowledge synthesis. Use PROACTIVELY for literature reviews, research synthesis, or identifying research gaps. Subagent · nyldn/claude-octopus
- Performance Engineer Octopus-only performance engineer. Use only when the user explicitly selects this agent or starts an Octopus workflow; never for an ordinary request. Subagent · nyldn/claude-octopus
- Security Auditor Security auditor. Delegate only when the user explicitly starts an Octopus workflow. Subagent · nyldn/claude-octopus