sliver-mcp 1.0.0

MCP server for Sliver C2 — AI agent automation for post-exploitation

sliver-mcp · English / 中文

---

English

MCP server for Sliver C2 — let AI agents automate post-exploitation.

V1.0 — 17 tools for session interaction, command execution, file operations, and reconnaissance.

Prerequisites

• Python >= 3.10
• uv (Python package manager)
• sliver-py (gRPC bindings)
• A running Sliver C2 server
• A Sliver operator config file (generated by sliver-client import or sliver-server operator)

Quick Start

# Clone both repos
git clone [e-mail]:hi-unc1e/sliver-mcp.git
cd sliver-mcp
git clone [e-mail]:hi-unc1e/sliver-py.git

# Install & run
uv sync
uv run sliver-mcp --config /path/to/operator.cfg

Configuration

Option	CLI	Default
Operator config	--config PATH	(required)
Downloads dir	--downloads-dir DIR	~/Downloads/sliver/
gRPC timeout	--timeout SECONDS	60
Debug logging	--debug	off

Claude Desktop Setup

{
  "mcpServers": {
    "sliver": {
      "command": "uv",
      "args": [
        "run",
        "--directory", "/path/to/sliver-mcp",
        "sliver-mcp",
        "--config", "/path/to/operator.cfg"
      ]
    }
  }
}

Tools

Category	Tool	Description
Server	server_version	Get Sliver server version
Sessions	list_sessions	List all active sessions
	session_info	Get detailed session info
	kill_session	Kill a session
Execute	execute	Run shell command (auto/cmd/powershell/sh)
Files	list_files	List directory contents
	change_directory	Change working directory
	working_directory	Get current directory
	make_directory	Create directory
	remove_path	Remove file/directory
	download_file	Download file to local
	upload_file	Upload file to remote
Recon	list_processes	List running processes
	network_interfaces	Get network interfaces
	network_connections	Get network connections
	get_environment	Get environment variable
	screenshot	Take screenshot (PNG + base64)

Downloads

Files are saved to: {downloads_dir}/{hostname}_{session_id}/{filename}.

Architecture

sliver-mcp (MCP server, Python)
  └── sliver-py (gRPC client library)
        └── grpcio → Sliver Server (gRPC + mTLS)

No Sliver client binary needed — sliver-py connects directly via gRPC.

Development

# Run tests (mock, no server needed)
uv sync --extra dev
uv run pytest tests/ -v --ignore=tests/test_smoke.py

# Run integration tests (needs real server, CI only)
SLIVER_INTEGRATION_TEST=1 uv run pytest tests/test_smoke.py -v

License

MIT

---

中文

针对 Sliver C2 的 MCP 服务器 — 让 AI agent 自动化后渗透操作。

V1.0 — 17 个工具，覆盖会话交互、命令执行、文件操作和信息收集。

前置条件

• Python >= 3.10
• uv（Python 包管理器）
• sliver-py（gRPC 绑定库）
• 运行中的 Sliver C2 服务器
• Sliver 操作员配置文件（通过 sliver-client import 或 sliver-server operator 生成）

快速开始

# Clone 两个仓库
git clone [e-mail]:hi-unc1e/sliver-mcp.git
cd sliver-mcp
git clone [e-mail]:hi-unc1e/sliver-py.git

# 安装并运行
uv sync
uv run sliver-mcp --config /path/to/operator.cfg

配置

配置项	CLI	默认值
操作员配置文件	--config PATH	（必填）
下载目录	--downloads-dir DIR	~/Downloads/sliver/
gRPC 超时	--timeout SECONDS	60
Debug 日志	--debug	off

Claude Desktop 配置

{
  "mcpServers": {
    "sliver": {
      "command": "uv",
      "args": [
        "run",
        "--directory", "/path/to/sliver-mcp",
        "sliver-mcp",
        "--config", "/path/to/operator.cfg"
      ]
    }
  }
}

工具列表

类别	工具	说明
服务端	server_version	获取 Sliver 服务器版本
会话	list_sessions	列出所有活跃会话
	session_info	获取会话详细信息
	kill_session	终止会话
执行	execute	执行命令（auto/cmd/powershell/sh）
文件	list_files	列出目录内容
	change_directory	切换工作目录
	working_directory	获取当前目录
	make_directory	创建目录
	remove_path	删除文件/目录
	download_file	下载文件到本地
	upload_file	上传文件到远程
侦察	list_processes	进程列表
	network_interfaces	网络接口信息
	network_connections	网络连接
	get_environment	获取环境变量
	screenshot	截屏（PNG + base64）

下载文件

文件保存路径：{下载目录}/{主机名}_{会话ID}/{文件名}。

架构

sliver-mcp (MCP 服务器, Python)
  └── sliver-py (gRPC 客户端库)
        └── grpcio → Sliver 服务器 (gRPC + mTLS)

无需 Sliver 客户端二进制文件 — sliver-py 直接通过 gRPC 连接。

开发

# 运行测试（Mock，无需服务器）
uv sync --extra dev
uv run pytest tests/ -v --ignore=tests/test_smoke.py

# 运行集成测试（需真实服务器，CI 环境）
SLIVER_INTEGRATION_TEST=1 uv run pytest tests/test_smoke.py -v

License

MIT