Sequential Thinking MCP Server

A Model Context Protocol (MCP) server focused on structured sequential thinking capabilities, designed to integrate with Cline's Memory Bank. This server helps break down complex problems into structured sequential steps, track reasoning chains, and store thinking patterns.

Features

• Create and manage sequential thinking chains for problem-solving
• Track chains of thought with validation at each step
• Store and retrieve reasoning patterns
• Analyze the quality of reasoning processes
• Visualize thinking pathways
• Seamlessly integrate with the Memory Bank system

Architecture

The server consists of the following core components:

• Sequential Thinking Engine: Manages thinking chains, steps, and reasoning validation
• Memory Bank Connector: Integrates with Cline's Memory Bank
• Tag Manager: Implements a comprehensive tagging system
• Visualization Generator: Creates visual representations of thinking chains
• Utilities: File storage, thinking validation, and other helpers

Available Tools

The server provides the following MCP tools:

create_thinking_chain

Create a new sequential thinking process with specified parameters.

• Input: problem description, thinking type, context
• Output: chain_id and initial structure

add_thinking_step

Add a step to an existing thinking chain.

• Input: chain_id, step description, reasoning, evidence
• Output: updated step information

validate_step

Validate logical connections between steps.

• Input: chain_id, step_id
• Output: validation results, potential issues

get_chain

Retrieve a complete thinking chain.

• Input: chain_id
• Output: full chain with all steps

generate_visualization

Create visual representation of a thinking chain.

• Input: chain_id, format (mermaid, json, text)
• Output: visualization code/data

save_to_memory

Save a thinking chain to Memory Bank.

• Input: chain_id, memory_name, tags
• Output: confirmation and memory_id

load_from_memory

Load a thinking chain from Memory Bank.

• Input: memory_id or search parameters
• Output: complete chain

search_related_thinking

Find related thinking chains based on parameters.

• Input: keywords, tags, thinking_type
• Output: list of relevant chains

apply_template

Apply a reasoning template to current thinking.

• Input: template_name, problem_context
• Output: pre-structured thinking chain

Thinking Types

The server supports various thinking types, each with specific patterns and structures:

• Analytical - Break down, analyze, synthesize
• Creative - Diverge, explore, converge
• Critical - Question, evaluate, conclude
• Systems - Map, analyze, model
• First-Principles - Identify, break down, reassemble
• Divergent - Generate alternatives, explore
• Convergent - Analyze, evaluate, select
• Inductive - Observe, pattern, hypothesize
• Deductive - Premise, logic, conclusion

Templates

The server includes ready-to-use reasoning templates to jumpstart the thinking process:

• First Principles Analysis - Break down a complex problem into its fundamental principles
• Systems Thinking Analysis - Analyze complex systems holistically

Installation

1. Ensure Node.js v14+ is installed
2. Clone the repository
3. Install dependencies:

   npm install
   

Usage

1. Start the server:

   node index.js
   

2. The server will be available as an MCP server that you can connect to via Claude/Cline

Memory Bank Integration

This server is designed to integrate with Cline's Memory Bank, allowing:

1. Reading from Memory Bank files (projectbrief.md, activeContext.md, etc.)
2. Storing complete thinking chains as structured memories
3. Updating activeContext.md with reasoning outcomes
4. Creating links between reasoning and Memory Bank sections

Example Tool Usage

// Example: Create a new thinking chain
{
  "problem": "How to improve user engagement on our platform",
  "thinking_type": "systems",
  "context": "Our user engagement metrics have decreased by 15% over the past quarter"
}

// Example: Add a thinking step
{
  "chain_id": "3a7e4fc0-5c1d-4b9f-9d1a-8b5e7c5a9d3e",
  "description": "Identify key components of the engagement system",
  "reasoning": "User engagement consists of several interconnected components including onboarding, core user actions, notification systems, and retention mechanisms.",
  "evidence": "Analysis of our user journey maps and analytics data",
  "confidence": 0.8
}

// Example: Generate a visualization
{
  "chain_id": "3a7e4fc0-5c1d-4b9f-9d1a-8b5e7c5a9d3e",
  "format": "mermaid",
  "options": {
    "showValidation": true,
    "showConfidence": true
  }
}

Tag System

The server implements a comprehensive tagging system with multiple dimensions:

• Thinking Type - analytical, creative, critical, systems, etc.
• Domain - business, science, technology, art, etc.
• Complexity - simple, moderate, complex
• Status - draft, validated, complete
• Custom - user-defined tags

License

MIT