neurodivergent-workplace-toolkit
mogoldb/neurodivergent-workplace-toolkit
MCP server providing communication assistance for neurodivergent professionals in technical workplaces
Instalação
npx convez-ai mogoldb/neurodivergent-workplace-toolkit1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
README
A tradução para o português deste README ainda está na fila.
Neurodivergent Workplace Toolkit
Workplace communication assistance for neurodivergent professionals in technical roles.
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
Install from PyPI:
pip install neurodivergent-workplace-toolkit
Then add the server to your MCP client. For Claude Desktop, edit claude_desktop_config.json (see INSTALL.md for Q CLI, Cursor, and other clients):
{
"mcpServers": {
"neurodivergent-workplace-toolkit": {
"command": "nwt"
}
}
}
Restart your client and test it:
Can you check this message: Hey team, thoughts on approach B?
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.
Usage Examples
Full examples: EXAMPLES.md
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.
For Developers/Contributors
If you want to contribute to the development of this tool, you can install it from source.
Installation from Source
Requirements: Python 3.10+
git clone https://github.com/mogoldb/neurodivergent-workplace-toolkit.git
cd neurodivergent-workplace-toolkit
pip install -e ".[dev]"
pytest # verify all tests pass
Install from Git (pin a tag)
To install a specific release directly from Git:
pipx install "git+https://github.com/mogoldb/neurodivergent-workplace-toolkit.git@v0.1.0"
# or
pip install "git+https://github.com/mogoldb/neurodivergent-workplace-toolkit.git@v0.1.0"
Either way, the nwt command runs the server:
nwt
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
License
MIT License - Use however helps you best.
Acknowledgments
Designed by and for neurodivergent professionals in technical workplaces.
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 INSTALL.md
- Usage examples: See EXAMPLES.md
- 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.