Skip to content
/doc-coauthoringOfficial

Guide users through a structured workflow for co-authoring documentation. Use when writing documentation, proposals, technical specs, decision docs, or similar structured content.

DocumentationWritingWorkflowCollaborationΒ· 2 min read

Quick import: Download the .md file and save it to .claude/commands/ (Claude Code), .cursorrules (Cursor), or paste as a system prompt in ChatGPT, Gemini, or any LLM API.

#What it does

The Doc Co-Authoring skill provides a structured three-stage workflow for collaborative document creation. It acts as an active guide, walking users through context gathering, iterative refinement, and reader testing to produce documentation that actually works for its audience.

#How to use

Trigger this skill when writing docs, creating proposals, drafting specs, PRDs, design docs, decision docs, RFCs, or any substantial writing task.

Help me write a technical design doc for our new auth system
I need to draft a product requirements document

#Skill instructions

#Stage 1: Context Gathering

Close the gap between what the user knows and what Claude knows:

  1. Meta-context questions: Document type, audience, desired impact, format constraints
  2. Info dumping: User provides all relevant context in any format -- stream of consciousness, links, channel references
  3. Clarifying questions: 5-10 targeted questions based on gaps identified during the dump
  4. Exit condition: Questions show understanding of edge cases and trade-offs, not just basics

#Stage 2: Refinement and Structure

Build the document section by section through brainstorming, curation, and iterative refinement:

  1. Structure agreement: Propose 3-5 sections appropriate for the doc type
  2. Per-section workflow:
    • Clarifying questions about what to include
    • Brainstorm 5-20 options for content
    • User curates (keep/remove/combine with justification)
    • Gap check for missing elements
    • Draft the section
    • Iterative refinement through surgical edits
  3. Near completion: Full document review for flow, consistency, redundancy, and filler

#Stage 3: Reader Testing

Test the document with a fresh Claude instance (no context bleed) to catch blind spots:

  1. Predict reader questions: Generate 5-10 realistic questions readers would ask
  2. Test with sub-agent: Fresh Claude reads only the document and answers questions
  3. Additional checks: Ambiguity, false assumptions, contradictions
  4. Fix and iterate: Loop back to refinement for any problematic sections

#Final Review

When reader testing passes, the user does a final read-through to verify facts, links, and technical details before the document is complete.

This skill is from the Anthropic Skills Repository.

AnthropicΒ·
View all skills