Implement MCP (Model Context Protocol) support for neural network memory

[konard]

🎯 Solution Overview

This PR implements MCP (Model Context Protocol) support for neural network memory, allowing GPTs and AI assistants to use the links database as persistent associative memory storage.

🔥 Key Features

• Full MCP Server Implementation: JSON-RPC 2.0 over stdio communication
• Neural Network Memory Operations:
  • store_memory - Store information as associative links
  • search_memory - Find memories by content or name
  • update_memory - Modify existing memory links
  • delete_memory - Remove memories from storage
• MCP Resources: Access to all links and search capabilities
• MCP Prompts: Templates for common memory operations
• CLI Integration: --mcp-server (alias --mcp) option

🧠 Why This Solves the Issue

Issue #56 requested "Support MCP for neural networks" to "allow GPTs to remember anything they need." This implementation:

1. Provides Persistent Memory: Neural networks can store information that survives across conversations
2. Uses Associative Storage: Links naturally represent relationships between concepts
3. Enables Efficient Retrieval: Fast search and filtering of stored memories
4. Offers Structured Data: All memories stored as triplets (index, source, target)
5. Supports Named References: Memories can be labeled for easy access

📋 Technical Implementation

• MCP Server (McpServer.cs): Full JSON-RPC 2.0 protocol implementation
• Data Models (McpModels.cs): Complete MCP message format definitions
• CLI Integration: Extended Program.cs with MCP server mode
• Dependencies: Added StreamJsonRpc for robust JSON-RPC support

🚀 Usage Example

# Start MCP server for neural network memory
clink --mcp-server

# Neural networks can now use memory operations:
# - Store: {"content": "User prefers coffee", "name": "preference"}  
# - Search: {"query": "coffee"}
# - Update: {"id": "1", "name": "new_preference"}
# - Delete: {"name": "old_memory"}

✅ Benefits for Neural Networks

• Long-term Memory: Information persists beyond single conversations
• Associative Relationships: Links connect related concepts naturally
• High Performance: Efficient storage and retrieval operations
• Standard Protocol: Compatible with any MCP client
• Rich Capabilities: Resources, tools, and prompts for complete memory management

This makes link-cli the ideal backend for neural network memory systems, enabling AI assistants to build and maintain sophisticated knowledge bases.

---

🤖 Generated with Claude Code

Fixes #56

[konard]

🤖 Solution Log

This log file contains the complete execution trace of the AI solution process.

📎 Log file uploaded as GitHub Gist (361KB)
🔗 View complete solution log

---

Log automatically attached by solve.mjs with --attach-solution-logs option