neurodivergent-workplace-toolkit by mogoldb - MCP Server

Neurodivergent Communications MCP Server

Communication assistance tools for neurodivergent professionals in technical workplaces.

Works with Claude Desktop, Amazon Q CLI, and any MCP-compatible AI system.

What This Does

Helps with communication challenges common in neurodivergent professionals:

✅ Check messages before sending (tone, clarity, completeness)
✅ Decode vague messages from others (implicit asks, real deadlines)
✅ Prep for meetings (talking points, questions, what to expect)
✅ Scaffold complex documents (structure before reading)
✅ Organize scattered thoughts into clear messages
✅ Catch up on long email threads (extract what you need to know)
✅ Decide call vs text (based on complexity and urgency)
✅ Get unstuck on reading (strategy to start overwhelming docs)
...and more!

You shouldn't spend 45 minutes editing a 3-line email.

Quick Start

1. Install

See for detailed instructions.

TL;DR:

# Install FastMCP
python3 -m pip install --user fastmcp

# Clone this repo
git clone https://github.com/mogoldb/neurodivergent-comms-mcp.git

# Configure Claude Desktop or Q CLI (see INSTALL.md)

2. Try It

See for real-world usage examples.

Quick test:

Can you check this message: Hey team, thoughts on approach B?

Your AI assistant should analyze it for clarity, tone, and effectiveness!

How to Use

Once installed, just talk naturally to your AI assistant:

Examples:

"Check this email before I send it: [paste]"
→ Calls check_message tool

"I got this confusing message from my manager: [paste]
What do they actually want?"
→ Calls decode_message tool

"Help me prep for tomorrow's architecture review"
→ Calls prep_meeting tool

"I need to read this 50-page design doc but I'm overwhelmed"
→ Calls scaffold_document or unstuck_reading tool

The AI automatically picks the right tool based on what you need.

What's Included

11 Communication Tools

Your AI assistant can call these functions to help you:

check_message - Analyze message drafts for clarity, tone, structure
decode_message - Extract explicit and implicit meaning from confusing messages
prep_meeting - Generate talking points and preparation for meetings
scaffold_document - Preview document structure before deep reading
check_tone - Validate tone and flag potential misinterpretations
call_or_text - Recommend communication method (call/text/video)
synthesize_thoughts - Organize scattered thoughts into clear message
catch_up_thread - Summarize long email/Slack threads
summarize_meeting - Extract decisions and action items from notes
ask_clarity - Draft polite messages asking for clarity
unstuck_reading - Get unstuck when unable to start reading a document

5 Communication Resources

Background knowledge automatically loaded for the AI:

Message clarity patterns - How to structure clear messages
Context interpretation - Decoding vague phrases and implicit meaning
Tone calibration - Professional levels and tone markers
Meeting structure - How to prep, participate, and follow up
Document scaffolding - Strategies for approaching complex documents

Design Philosophy

This toolkit treats neurodivergent communication patterns as different, not deficient.

Core principles:

Bottom-up processing - Start with details, build to summary
Explicit over implicit - Make assumptions visible
Structure first - Scaffolding before content
External working memory - Toolkit holds context for you
Decision reduction - Clear paths, fewer choices

Designed for: ADHD, autism, dyslexia, and executive function differences.

Why MCP?

Model Context Protocol (MCP) is an open standard that makes these tools available across multiple AI systems:

✅ Claude Desktop (personal use)
✅ Amazon Q CLI (work environments)
✅ Cursor (coding with AI)
✅ Any MCP-compatible system

Write once, use everywhere. No platform lock-in.

Installation

Full guide:

Requirements:

Python 3.10+
Claude Desktop or Q CLI
5 minutes

Quick install for Claude Desktop (macOS):

# Install FastMCP
python3 -m pip install --user fastmcp

# Clone repo
git clone https://github.com/mogoldb/neurodivergent-comms-mcp.git
cd neurodivergent-comms-mcp

# Configure Claude Desktop
mkdir -p ~/Library/Application\ Support/Claude
cat > ~/Library/Application\ Support/Claude/claude_desktop_config.json << 'EOF'
{
  "mcpServers": {
    "neurodivergent-comms": {
      "command": "python3",
      "args": [
        "/FULL/PATH/TO/neurodivergent-comms-mcp/src/server.py"
      ]
    }
  }
}
EOF

# Update the path above, then restart Claude Desktop

Usage Examples

Full examples:

Quick tests:

Check message:

Check this email: Hey team, just wanted to circle back...

Decode message:

My manager said "when you get a chance" - what do they actually mean?

Prep meeting:

Help me prep for Architecture Review. I'm the tech lead.

Troubleshooting

Server won't start

# Test manually
python3 /path/to/neurodivergent-comms-mcp/src/server.py

# Should start without errors (Ctrl+C to stop)

Tools not showing up

Completely quit and restart Claude Desktop (Cmd+Q)

Check config file exists:

cat ~/Library/Application\ Support/Claude/claude_desktop_config.json

Verify Python 3.10+: python3 --version

Wrong Python version

Need 3.10+. Install from python.org

More help: See troubleshooting section.

File Structure

neurodivergent-comms-mcp/
├── src/
│   ├── server.py              # Main MCP server (11 tools, 5 resources)
│   └── resources/             # Communication rule files
│       ├── message-clarity.md
│       ├── context-interpretation.md
│       ├── tone-calibration.md
│       ├── meeting-structure.md
│       └── document-scaffolding.md
├── INSTALL.md                 # Installation guide
├── EXAMPLES.md                # Real-world usage examples
├── README.md                  # This file
├── pyproject.toml            # Python package config
└── example-qcli-config.json  # Example Q CLI config

Contributing

Found a communication pattern that would help? PRs welcome!

Areas for contribution:

Additional communication scenarios
More example prompts
Platform-specific installation guides
Bug fixes and improvements

Roadmap

v1.0 (Current):

✅ 11 core communication tools
✅ 5 communication resource files
✅ Claude Desktop support
✅ Q CLI support

Future:

Publish to PyPI (installable via pip install neurodivergent-comms-mcp)
Add workplace-specific integrations (Slack, email, etc.)
Conversation memory (remember your communication preferences)
Additional tools based on user feedback
Support for more MCP clients (Cursor, etc.)

License

MIT License - Use however helps you best.

Acknowledgments

Special thanks to:

The MCP community for the protocol
FastMCP for the Python framework
Everyone who provided feedback on communication patterns

Support

Installation help: See
Usage examples: See
Issues: Open an issue on GitHub
Questions: Discussions welcome!

Remember: You're not deficient. Your communication style is just different. This toolkit helps bridge that gap.