BinaryAnalysis-MCP
Enables analysis of binary files (PE, ELF, Mach-O, COFF) by providing tools to retrieve info, headers, sections, imports, exports, libraries, security hardening, signatures, and COFF object file details.
README
BinaryAnalysis-MCP
An MCP server for analysing PE, ELF, Mach-O, and COFF binary files using LIEF. Pass an absolute file path to any tool and the format is auto-detected.
Tools
| Tool | Description |
|---|---|
get_binary_info |
Quick triage — format, architecture, entry point, section/import/export counts, NX & PIE flags |
get_binary_headers |
Full header dump (PE DOS/COFF/Optional, ELF header, Mach-O header) |
get_binary_sections |
All sections with name, size, virtual address, entropy, permissions, image base, and entry point |
get_binary_imports |
Imported functions grouped by library (PE by DLL, ELF by shared library, Mach-O by dylib) |
get_binary_exports |
Exported functions/symbols with ordinals, addresses, and forwarding info |
get_binary_libraries |
Dynamic library dependencies (DLLs / shared objects / dylibs) |
get_binary_security |
Security hardening — ASLR, DEP/NX, SEH, CFG, RELRO, stack canaries, code signing |
get_binary_signatures |
Code-signing details — PE Authenticode/x509 certs, Mach-O LC_CODE_SIGNATURE/CodeDirectory |
get_coff_info |
COFF object file analysis — header, sections, symbols, and relocations |
Requirements
- Python 3.10+
- Dependencies listed in
requirements.txt:mcp[cli]— Model Context Protocol SDKlief>=0.17.0— binary parsing library
Installation
git clone https://github.com/Ap3x/BinaryAnalysis-MCP.git
cd BinaryAnalysis-MCP
python -m venv .venv
# Windows
.venv\Scripts\activate
# macOS / Linux
source .venv/bin/activate
pip install -r requirements.txt
Running the server
python server.py
The server communicates over stdio using the MCP protocol.
MCP client configuration
Claude Desktop
Add the following to your Claude Desktop config file:
- Windows:
%APPDATA%\Claude\claude_desktop_config.json - macOS:
~/Library/Application Support/Claude/claude_desktop_config.json
{
"mcpServers": {
"binary-analysis": {
"command": "python",
"args": ["C:/path/to/BinaryAnalysis-MCP/server.py"],
"env": {}
}
}
}
If you're using a virtual environment, point directly to the venv Python:
{
"mcpServers": {
"binary-analysis": {
"command": "C:/path/to/BinaryAnalysis-MCP/.venv/Scripts/python.exe",
"args": ["C:/path/to/BinaryAnalysis-MCP/server.py"],
"env": {}
}
}
}
Claude Code (CLI)
In your project's .mcp.json:
{
"mcpServers": {
"binary-analysis": {
"command": "python",
"args": ["C:/path/to/BinaryAnalysis-MCP/server.py"],
"env": {}
}
}
}
Generic MCP client (stdio)
Any MCP-compatible client can launch the server as a subprocess:
{
"command": "python",
"args": ["/absolute/path/to/server.py"],
"transport": "stdio"
}
Example usage
Once connected, ask your MCP client to call the tools with an absolute file path:

Analyse the security hardening of C:\Windows\System32\notepad.exe
List all imported DLLs for /usr/bin/ls
Show me the PE headers of C:\Windows\explorer.exe
Example output
get_binary_info — C:\Windows\System32\notepad.exe
{
"file": "C:/Windows/System32/notepad.exe",
"format": "PE",
"entrypoint": "0x1400019b0",
"imagebase": "0x140000000",
"is_pie": true,
"has_nx": true,
"sections": 8,
"imported_functions": 339,
"exported_functions": 0,
"libraries": 56,
"machine": "AMD64",
"subsystem": "WINDOWS_GUI",
"has_signatures": false,
"has_tls": false,
"has_resources": true,
"has_rich_header": true,
"has_relocations": true
}
get_binary_security — C:\Windows\System32\notepad.exe
{
"aslr_dynamic_base": true,
"aslr_high_entropy_va": true,
"dep_nx_compat": true,
"seh": true,
"guard_cf": true,
"force_integrity": false,
"appcontainer": false,
"is_pie": true,
"has_nx": true,
"signed": false,
"format": "PE"
}
get_binary_sections — C:\Windows\System32\notepad.exe
{
"format": "PE",
"image_base": "0x140000000",
"entrypoint": "0x19b0",
"count": 8,
"sections": [
{
"name": ".text",
"virtual_address": "0x1000",
"size": 159744,
"entropy": 6.2826,
"virtual_size": 157410,
"sizeof_raw_data": 159744,
"characteristics": ["CNT_CODE", "MEM_EXECUTE", "MEM_READ"]
},
{
"name": ".rdata",
"virtual_address": "0x29000",
"size": 45056,
"entropy": 5.8039,
"virtual_size": 42456,
"sizeof_raw_data": 45056,
"characteristics": ["CNT_INITIALIZED_DATA", "MEM_READ"]
},
{
"name": ".data",
"virtual_address": "0x34000",
"size": 4096,
"entropy": 1.624,
"virtual_size": 10048,
"sizeof_raw_data": 4096,
"characteristics": ["CNT_INITIALIZED_DATA", "MEM_READ", "MEM_WRITE"]
},
{
"name": ".rsrc",
"virtual_address": "0x3a000",
"size": 126976,
"entropy": 7.0998,
"virtual_size": 123344,
"sizeof_raw_data": 126976,
"characteristics": ["CNT_INITIALIZED_DATA", "MEM_READ"]
}
]
}
Truncated to 4 of 8 sections for brevity.
Project structure
server.py — entrypoint: imports tools, runs mcp
app.py — FastMCP instance
helpers.py — parse_binary, hex_addr, safe_str, safe_enum, format_name, _error
tools/
__init__.py — imports all tool modules (triggers @mcp.tool registration)
info.py — get_binary_info
headers.py — get_binary_headers
sections.py — get_binary_sections
imports.py — get_binary_imports
exports.py — get_binary_exports
libraries.py — get_binary_libraries
security.py — get_binary_security + _pe_security, _elf_security, _macho_security
certificates.py — get_binary_signatures (PE Authenticode/x509, Mach-O LC_CODE_SIGNATURE)
coff.py — get_coff_info
tests/
conftest.py — shared fixtures and sample file paths
test_helpers.py — tests for helpers.py utilities
test_info.py — tests for get_binary_info
test_headers.py — tests for get_binary_headers
test_sections.py — tests for get_binary_sections
test_imports.py — tests for get_binary_imports
test_exports.py — tests for get_binary_exports
test_libraries.py — tests for get_binary_libraries
test_security.py — tests for get_binary_security
test_coff.py — tests for get_coff_info
binary-samples/ — test binaries (git submodule)
.github/workflows/
tests.yml — CI: runs pytest on push/PR to main
Pairs well with
This MCP pairs well with GhidraMCP — an MCP server that exposes Ghidra's reverse engineering capabilities. Use BinaryAnalysis-MCP for quick static triage (headers, imports, security flags) and GhidraMCP for deeper decompilation and control-flow analysis.
License
This project is licensed under the GNU General Public License v3.0.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。