feat: Add MCP Server for AI Agent integration

[lxcxjxhx]

Summary

This PR adds a Model Context Protocol (MCP) Server implementation for Qiling Framework, enabling AI Agents to programmatically interact with the binary emulation framework.

Changes

New Files

• qiling/mcp/init.py: Module entry point
• qiling/mcp/models.py: Data models for structured JSON output
• qiling/mcp/tools.py: 12 MCP tool implementations
• qiling/mcp/server.py: MCP server with stdio/SSE/HTTP transport support
• qiling/mcp/test_server.py: Comprehensive test suite
• qiling/mcp/README.md: Complete documentation

MCP Tools (12 total)

Binary Management:

• load_binary: Load and initialize binary for emulation
• get_state: Get current emulation state

Memory Operations:

• read_memory: Read memory contents
• write_memory: Write to memory
• get_memory_map: Query memory map

Register Operations:

• get_registers: Get all register values
• set_registers: Set register values

Execution Control:

• start_emulation: Start binary emulation
• stop_emulation: Stop emulation

Hook Management:

• set_hook: Set execution hooks
• list_hooks: List active hooks
• remove_hook: Remove hooks

Features

• Multiple Transport Support: stdio, SSE, HTTP
• Structured JSON Output: All tools return well-defined JSON responses
• Context Management: Global state management for emulation sessions
• Test Coverage: Complete test suite with mocks
• Documentation: Comprehensive README with examples

Use Cases

1. CTF Challenge Analysis: AI Agents can load and analyze CTF binaries
2. Malware Analysis: Safe emulation of suspicious binaries
3. Exploit Development: Test and debug exploits in controlled environment
4. Binary Reverse Engineering: Automated analysis of binary behavior
5. Security Research: Programmatic binary analysis at scale

Testing

All tools have been tested with comprehensive test suite:

pytest qiling/mcp/test_server.py -v

Implementation Details

• Based on real Qiling Framework API analysis
• Uses actual Qiling internal APIs (Qiling, QlMemoryManager, etc.)
• Follows MCP specification (JSON-RPC 2.0)
• Compatible with Qiling 2.0+

Related Issues

This implementation addresses the need for AI Agent integration with Qiling Framework, enabling programmatic binary analysis and emulation control.

[wtdcode]

I would prefer to leave this outside the main repo of Qiling, like: https://github.com/drvcvt/emulation-mcp