sequential-thinking-claude-code

sequential-thinking-claude-code

An enhanced version of the original mcp-sequentialthinking-tools by Scott Spence, which was adapted from the MCP Sequential Thinking Server.

This enhanced edition transforms the original framework into an intelligent tool recommendation system with:

• 26 pre-configured tools including all Claude Code tools
• Automatic tool recommendations based on thought content
• Deep integration with Basic Memory MCP for knowledge management
• Pattern-based intelligence that suggests the right tools at the right time
• Context-aware caching with intelligent tool result management
• Performance metrics tracking for optimization insights
• 100% test coverage with comprehensive validation

Credits: Original repository by Scott Spence. Basic Memory tools from Basic Machines Co. Enhanced with permission.

Thoughts

I think the biggest benefit is the improved utilization of branching, revisiting, and thought revision. Thoughts are much more dynamic and intentional, spread out throughout the conversation at the correct time.

Of course, Claude Code already does a good job of tool function use, but with the integration of sequential thinking with explicit function calls for those tools makes everything more seamless and responses are significantly better.

There is also caching for tools and thoughts, session continuation and context management, and a bunch of other improvements.

I want to create separate config files for the most popular MCP tools with instructions for users on how to add the specific tool calls of the MCP in the config for ST. It's just an array so it would be simple, just have to identify all the MCPs and their tools, or just the MCPs themselves because agents will know what to do with them once recommended. I found the specific function/tool tools in the recommendation improves the responses and creates more of a seamless flow.

I would do this for other agents as well like Codex, Cline, Augment, etc.. I actually have a copy modified for Augment Code, going to make a branch for that soon.

What's New in This Enhanced Edition

🚀 Key Improvements

• Auto-recommends tools based on your thought content (no manual specification needed!)
• 26 pre-integrated tools ready to use out of the box
• Intelligent pattern matching that understands context
• Basic Memory integration for persistent knowledge management
• Context-aware caching with TTL and size-based eviction
• Performance metrics collection for usage optimization
• Enhanced tool recommendations with contextual awareness

📦 Included Tool Sets

Claude Code Tools (17 tools)

• File operations: Read, Write, Edit, MultiEdit, LS, Glob, Grep
• Execution: Bash, Agent
• Notebooks: NotebookRead, NotebookEdit
• Web: WebFetch, WebSearch
• Tasks: TodoRead, TodoWrite
• Special: StickerRequest

Basic Memory Tools (9 tools)

• mcp__basic-memory__write_note - Record knowledge
• mcp__basic-memory__read_note - Access saved information
• mcp__basic-memory__search_notes - Search your knowledge base
• mcp__basic-memory__build_context - Follow knowledge connections
• And more...

Original Features (Significantly improved utilization of the original functions)

• 🤔 Dynamic and reflective problem-solving through sequential thoughts
• 🔄 Flexible thinking process that adapts and evolves
• 🌳 Support for branching and revision of thoughts
• 🛠️ Automatic tool suggestions based on thought content
• 📊 Pre-configured confidence scores for each tool
• 🔍 Context-aware rationale generation
• 📝 Step tracking with expected outcomes
• 🔄 Progress monitoring with previous and remaining steps
• 🎯 Intelligent alternative tool suggestions**

How It Works

Original Behavior

The original server provided a framework for tool recommendations but required manual tool specification.

Enhanced Behavior

This version automatically analyzes your thoughts and recommends appropriate tools:

Each recommendation includes:

• Confidence score (0-1) based on pattern matching
• Clear rationale explaining why the tool fits
• Priority level for execution order
• Alternative tools that could also work

The enhanced pattern matching understands context like:

• "record", "save", "document" → Basic Memory write_note
• "read", "examine", "check" → Read tool
• "multiple edits" → MultiEdit tool
• "run", "execute" → Bash tool

Example Usage

Here's an example of how the server guides tool usage:

The server tracks your progress and supports:

• Creating branches to explore different approaches
• Revising previous thoughts with new information
• Maintaining context across multiple steps
• Suggesting next steps based on current findings

Configuration

This server requires configuration through your MCP client. Here are examples for different environments:

Cline Configuration

Add this to your Cline MCP settings:

Claude Desktop with WSL Configuration

For WSL environments, add this to your Claude Desktop configuration:

API

The server implements a single MCP tool with configurable parameters:

sequentialthinking_tools

A tool for dynamic and reflective problem-solving through thoughts, with intelligent tool recommendations.

Parameters:

• thought (string, required): Your current thinking step
• next_thought_needed (boolean, required): Whether another thought step is needed
• thought_number (integer, required): Current thought number
• total_thoughts (integer, required): Estimated total thoughts needed
• is_revision (boolean, optional): Whether this revises previous thinking
• revises_thought (integer, optional): Which thought is being reconsidered
• branch_from_thought (integer, optional): Branching point thought number
• branch_id (string, optional): Branch identifier
• needs_more_thoughts (boolean, optional): If more thoughts are needed
• current_step (object, optional): Current step recommendation with:
  • step_description: What needs to be done
  • recommended_tools: Array of tool recommendations with confidence scores
  • expected_outcome: What to expect from this step
  • next_step_conditions: Conditions for next step
• previous_steps (array, optional): Steps already recommended
• remaining_steps (array, optional): High-level descriptions of upcoming steps

Development

Setup

1. Clone the repository
2. Install dependencies:

3. Build the project:

4. Run in development mode:

Publishing

The project uses changesets for version management. To publish:

1. Create a changeset:

2. Version the package:

3. Publish to npm:

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

MIT License - see the LICENSE file for details.

Acknowledgments

• Built on the Model Context Protocol
• Adapted from the MCP Sequential Thinking Server