Contributing to Claude Code Writing System

September 5, 2025 · View on GitHub

We welcome contributions that improve the writing quality and research capabilities of this template! This project helps content creators build personalized AI writing systems, and your expertise can make it work better for everyone.

🎯 Current Priority Areas

We're currently focused on improvements in these key areas:

1. Agent Prompt Improvements

LinkedIn agent optimization (hooks, engagement, professional tone)
Newsletter agent refinement (subject lines, personal voice, structure)
Conversational agent enhancement (social media posts, podcast Q&A scripts)

2. Research & Writing Workflow

Improving the /research command effectiveness
Enhancing the /write command output quality
Suggesting high-quality default research sources
Streamlining the research-to-publication pipeline

📋 Contribution Requirements

Before You Submit

All contributions must include:

Before/After Comparison: Test your changes against real content
Output Examples: Show actual generated content with your improvements
Brief Explanation: 2-3 sentences on why the "after" version is better
Testing Evidence: Demonstrate the improvement works consistently

Example Contribution Format

## Improvement: LinkedIn Agent Hook Generation

**Before**: Generic opening lines that don't engage
**After**: Compelling hooks that drive comments and shares

**Test Results**:
- Original prompt generated: "Here's an interesting trend in remote work..."
- Improved prompt generated: "Most productivity advice for remote workers is still stuck in 2019."

**Why Better**: The improved version uses a contrarian hook that immediately challenges assumptions, leading to higher engagement rates.

🚀 How to Contribute

Small Improvements (Preferred)

Focus on specific, targeted enhancements:

Better LinkedIn post hooks
More engaging newsletter subject lines
Improved research source prioritization
Enhanced voice consistency in outputs

Larger Changes

If you have ideas for significant workflow redesigns or major new features:

Discuss first: Email info@womendefiningai.com before implementing
Get approval: Wait for green light before spending time on large changes
Break it down: Consider splitting large ideas into smaller contributions

🔄 Contribution Process

1. Set Up Your Environment

# Fork this repository on GitHub
# Clone your fork
git clone https://github.com/YOUR-USERNAME/claudecode-writertemplate.git
cd claudecode-writertemplate

# Install Claude Code if you haven't already
# Follow the README instructions for setup

2. Make Your Improvements

For Agent Prompts: Edit files in .claude/agents/
For Commands: Modify files in .claude/commands/
For Research Sources: Update context/research-sources.md

3. Test Thoroughly

Run your improved agents/commands on real content
Document the before/after results
Verify improvements are consistent across different topics

4. Submit Your Contribution

# Create a new branch
git checkout -b improve-linkedin-agent

# Make your changes and commit
git add .
git commit -m "Improve LinkedIn agent hook generation"

# Push and create pull request
git push origin improve-linkedin-agent

5. Pull Request Requirements

Your PR description must include:

What you changed (specific agent/command/workflow)
Why you changed it (what problem it solves)
Before/after examples (real output comparisons)
Testing details (how you verified the improvement)

🎨 Quality Standards

Good Contributions

✅ Solve specific, real problems users face
✅ Include concrete before/after examples
✅ Improve output quality measurably
✅ Maintain consistency with the overall system
✅ Are thoroughly tested on multiple topics

What We Don't Need

❌ Generic writing examples (users should customize with their own)
❌ Untested prompt changes
❌ Major architectural changes without prior discussion
❌ Platform agents for niche platforms (focus on major ones first)
❌ Changes that make the system more complex without clear benefit