Add MCP Toolz server

[taylorleese]

Summary

This PR adds MCP Toolz to the Docker MCP registry - a server for context management and todo persistence for Claude Code with AI feedback from ChatGPT, Claude, Gemini, and DeepSeek.

Server Details

• Name: MCP Toolz
• Category: Productivity
• GitHub: https://github.com/taylorleese/mcp-toolz
• PyPI: https://pypi.org/project/mcp-toolz/ (v0.3.7)
• Docker Image: Will be published to mcp/mcp-toolz

Features

Context Management

• Save and retrieve conversation contexts, code snippets, suggestions, and errors
• Search contexts by query, tags, or type
• Link contexts across sessions for continuity
• Project organization by directory

Todo Persistence

• Create snapshots of todo lists that persist across sessions
• Restore previous todo states
• Track todos per project with timestamps
• Never lose your place when restarting Claude Code

AI Feedback

• Get feedback from ChatGPT (via OpenAI API) on saved contexts
• Get feedback from Claude (via Anthropic API) on saved contexts
• Get feedback from Gemini (via Google API) on saved contexts
• Get feedback from DeepSeek API on saved contexts
• Compare insights from multiple AI models

MCP Resources

• mcp-toolz://contexts/project/recent - Recent contexts for current project
• mcp-toolz://contexts/project/sessions - List of recent sessions
• mcp-toolz://todos/recent - Recent todo snapshots
• mcp-toolz://todos/active - Active todo snapshot for current project

Tools Provided

Context Tools (5):

• context_save - Save a new context entry
• context_list - List recent contexts
• context_search - Search contexts by query/tags
• context_get - Get full context details by ID
• context_delete - Delete a context

Todo Tools (6):

• todo_save - Save a new todo snapshot
• todo_list - List recent todo snapshots
• todo_search - Search todo snapshots
• todo_get - Get full todo snapshot details
• todo_restore - Restore a todo snapshot
• todo_delete - Delete a todo snapshot

AI Feedback Tools (4):

• ask_chatgpt - Get ChatGPT's opinion on a context
• ask_claude - Get Claude's opinion on a context
• ask_gemini - Get Gemini's opinion on a context
• ask_deepseek - Get DeepSeek's opinion on a context

Total: 15 tools

Configuration

Optional environment variables:

• MCP_TOOLS_DB_PATH - Database path (default: /home/app/.mcp-toolz/contexts.db)
• OPENAI_API_KEY - For ChatGPT feedback (optional)
• ANTHROPIC_API_KEY - For Claude feedback (optional)
• GOOGLE_API_KEY - For Gemini feedback (optional)
• DEEPSEEK_API_KEY - For DeepSeek feedback (optional)
• MCP_TOOLS_MODEL - OpenAI model to use (default: gpt-5)
• MCP_TOOLS_CLAUDE_MODEL - Anthropic model to use (default: claude-sonnet-4-5-20250929)
• MCP_TOOLS_GEMINI_MODEL - Google model to use (default: gemini-2.0-flash-thinking-exp-01-21)
• MCP_TOOLS_DEEPSEEK_MODEL - DeepSeek model to use (default: deepseek-chat)

Testing

• ✅ Docker image builds successfully locally
• ✅ All 15 MCP tools defined and tested
• ✅ Package published to PyPI (v0.3.7)
• ✅ CI/CD with GitHub Actions (tests, linting, coverage)
• ✅ Integration tests passing
• ✅ OpenSSF Best Practices badge
• ✅ OpenSSF Scorecard 6.9/10
• ✅ Property-based fuzzing with Hypothesis

License

MIT License

Checklist

• Created server.yaml with all required fields
• Created tools.json with all 15 tools
• Dockerfile exists in source repository
• Server is published to PyPI
• License permits consumption (MIT)
• Documentation includes usage instructions
• Integration tests passing

---

Part of taylorleese/mcp-toolz#11 - Publishing MCP Toolz to official registries and community directories.

Related PRs:

• Add MCP Toolz to third-party servers modelcontextprotocol/servers#2900 - Official MCP third-party servers list
• Add MCP Toolz server #353 - Docker MCP registry
• Add MCP Toolz to Knowledge & Memory section punkpeye/awesome-mcp-servers#1444 - Awesome MCP Servers community list