About mcp_reverse_engineering
mcp_reverse_engineering is a sandboxed interface for running reverse-engineering utilities through a CLI, Python API, or MCP server. It is intended for binary analysis while restricting tool execution with filesystem isolation, timeouts, and argument validation.
Key features
- 01Unified access to analysis utilities
- 02Filesystem-isolated execution
- 03Execution timeouts and argument validation
- 04CLI, Python API, and MCP server
Use cases
- 01Extract printable strings from a binary
- 02Disassemble a binary with objdump
- 03Inspect ELF headers with readelf
MCP support
Exposes reverse-engineering tools such as strings, objdump, readelf, and binwalk to MCP clients. Execution uses a sandbox workspace and configured safety restrictions.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEmcp-reverse-engineering
A sandboxed MCP (Model Context Protocol) tool for reverse engineering that provides a unified interface to various reverse engineering tools with security restrictions.
Purpose
This project provides a secure, sandboxed environment for executing reverse engineering tools via CLI or MCP protocol. It wraps common reverse engineering utilities (strings, objdump, readelf, binwalk, etc.) with safety features like filesystem isolation, timeouts, and argument validation.
Install
pip install mcp-reverse-engineering
Or for development:
pip install -e ".[dev]"
MCP Server Installation
To use as an MCP server with Claude Desktop:
mcp install src/mcp_reverse_engineering/server.py
Usage
CLI
# Extract strings from a binary
mcp-re --tool strings --file /path/to/binary
# Disassemble a binary
mcp-re --tool objdump --args "['-d']" --file /path/to/binary
# Analyze ELF headers
mcp-re --tool readelf --args "['-h', '-s']" --file /path/to/elf
# Run binwalk for firmware analysis
mcp-re --tool binwalk --file /path/to/firmware.bin
Python API
from mcp_reverse_engineering import ReverseEngineeringEngine
# Create engine with default config
engine = ReverseEngineeringEngine(
workspace="./workspace",
timeout=30,
)
# List available tools
print(engine.list_available_tools())
# Execute a tool
result = engine.execute_tool("strings", ["-n", "8"], "/path/to/binary")
print(result)
MCP Server
from mcp_reverse_engineering.server import mcp, strings, objdump, readelf, binwalk
# Run the server (stdio transport for Claude Desktop)
if __name__ == "__main__":
mcp.run()
API
ReverseEngineeringEngine
Main class for executing reverse engineering tools.
engine = ReverseEngineeringEngine(
workspace: str = "./workspace", # Sandbox directory
timeout: int = 30, # Tool execution timeout
config_path: str | Path | None = None, # YAML config path
)
Methods:
execute_tool(tool_name: str, args: List[str], file_path: Optional[str] = None) -> str- Execute a toollist_available_tools() -> List[str]- List enabled toolsget_tool_documentation(tool_name: str) -> Dict[str, Any]- Get tool docsget_mcp_tools() -> List[Dict[str, Any]]- Get MCP tool schemas
Available Tools
| Tool | Category | Description |
|---|---|---|
| file | file_tools | Determine file type |
| strings | file_tools | Extract printable strings |
| hexdump | file_tools | Hexadecimal dump |
| xxd | file_tools | Hexadecimal dump |
| objdump | binary_tools | Disassemble binary |
| readelf | binary_tools | Read ELF headers |
| binwalk | firmware_tools | Firmware analysis |
Development
# Clone the repository
git clone https://github.com/daedalus/mcp_reverse_engineering.git
cd mcp_reverse_engineering
# Install dependencies
pip install -e ".[test]"
# Run tests
pytest
# Format code
ruff format src/ tests/
# Lint
ruff check src/ tests/
# Type check
mypy src/
# Install pre-commit hooks
pip install pre-commit
pre-commit install
MCP Server Configuration
mcp-name: io.github.daedalus/mcp-reverse-engineering
Requirements
- Python 3.11+
- External tools: binwalk, radare2, ghidra, etc. (must be installed separately)
License
MIT
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.