qwen-code

中文版 | English

English

Sandbox runtime for the Qwen coding agent — supports qwen-code-driven autonomous deployment and provides a Python 3.12 + Node.js 22 AI coding foundation.

Environment

This template comes preinstalled with the following tools and dependencies:

Category	Details
Base image	python:3.12-slim
Python	Python 3.12 + venv (at `/opt/venv`)
Node.js	Node.js 22.x + npm
Python packages	`dashscope`, `openai>=1.0`, `httpx`, `rich`
Node.js packages	`@qwen-code/qwen-code@0.23.0`
System tools	git, curl, wget, ripgrep, jq, build-essential
Resources	2 CPU / 4096 MB memory

Installation

Install from a local path:

abx template install ./qwen-code --registry-type local

Install from GitHub:

abx template install alibaba-agent-sandbox/awesome-templates//qwen-code

Usage

Create a sandbox instance and pass in the API key:

abx create --template qwen-code --env DASHSCOPE_API_KEY=sk-xxx

Call a Qwen model inside the sandbox via the DashScope SDK:

abx exec <sandbox-id> -- python3 -c "
import dashscope
from dashscope import Generation
response = Generation.call(model='qwen-turbo', prompt='Write a quicksort algorithm')
print(response.output.text)
"

The OpenAI-compatible interface is also supported:

abx exec <sandbox-id> -- python3 -c "
from openai import OpenAI
client = OpenAI(
    api_key='your-dashscope-key',
    base_url='https://dashscope.aliyuncs.com/compatible-mode/v1'
)
response = client.chat.completions.create(
    model='qwen-turbo',
    messages=[{'role': 'user', 'content': 'Write a quicksort'}]
)
print(response.choices[0].message.content)
"

Using the Python SDK:

from agent_sandbox import Sandbox

sandbox = Sandbox.create_sync(
    template="qwen-code",
    env={"DASHSCOPE_API_KEY": "sk-xxx"}
)

result = sandbox.commands.run("python3 --version && node --version")
print(result.stdout)

Configuration

Environment Variables

Variable	Description	Required
`DASHSCOPE_API_KEY`	Alibaba Cloud DashScope API key	Yes
`WORKSPACE`	Working directory path, defaults to `/workspace`	No

Manual AI Coding Tool Installation

This template already ships with the qwen-code CLI, ready to use:

# Use qwen-code for NL-based deployment
abx run <sandbox-id> qwen --prompt "Deploy this FastAPI project"

# Or use the qwen command directly inside the sandbox
abx exec <sandbox-id> -- qwen -p "Analyze the project and deploy it" --yolo

To install additional tools:

# Install extra Python packages
abx exec <sandbox-id> -- pip install <agent-package>

# Install extra Node.js tools
abx exec <sandbox-id> -- npm install -g <agent-tool>

Notes

API key security: Never hardcode DASHSCOPE_API_KEY into the template; always inject it via the --env option or environment variables.
Network access: The DashScope API (dashscope.aliyuncs.com) must be reachable — make sure the sandbox has outbound internet connectivity.
Compatible interface: DashScope offers an OpenAI-compatible mode, so Qwen models can be called directly through the openai library.
Virtual environment: Python packages are installed in /opt/venv; no manual activation is needed since PATH is preconfigured.

中文

通义千问（Qwen）编码 Agent 的沙箱运行环境，支持 qwen-code 驱动的 AI 自主部署，提供 Python 3.12 + Node.js 22 的 AI 编程基础环境。

环境说明

本模板预装以下工具和依赖：

类别	内容
基础镜像	python:3.12-slim
Python	Python 3.12 + venv 虚拟环境（位于 `/opt/venv`）
Node.js	Node.js 22.x + npm
Python 包	`dashscope`、`openai>=1.0`、`httpx`、`rich`
Node.js 包	`@qwen-code/qwen-code@0.23.0`
系统工具	git、curl、wget、ripgrep、jq、build-essential
资源配置	2 CPU / 4096 MB 内存

安装方式

从本地安装：

abx template install ./qwen-code --registry-type local

从 GitHub 安装：

abx template install alibaba-agent-sandbox/awesome-templates//qwen-code

使用示例

创建沙箱实例并传入 API Key：

abx create --template qwen-code --env DASHSCOPE_API_KEY=sk-xxx

在沙箱中通过 DashScope SDK 调用 Qwen 模型：

abx exec <sandbox-id> -- python3 -c "
import dashscope
from dashscope import Generation
response = Generation.call(model='qwen-turbo', prompt='编写一个快速排序算法')
print(response.output.text)
"

也支持使用 OpenAI 兼容接口：

abx exec <sandbox-id> -- python3 -c "
from openai import OpenAI
client = OpenAI(
    api_key='your-dashscope-key',
    base_url='https://dashscope.aliyuncs.com/compatible-mode/v1'
)
response = client.chat.completions.create(
    model='qwen-turbo',
    messages=[{'role': 'user', 'content': '编写一个快速排序'}]
)
print(response.choices[0].message.content)
"

使用 Python SDK：

from agent_sandbox import Sandbox

sandbox = Sandbox.create_sync(
    template="qwen-code",
    env={"DASHSCOPE_API_KEY": "sk-xxx"}
)

result = sandbox.commands.run("python3 --version && node --version")
print(result.stdout)

配置说明

环境变量

变量名	说明	必填
`DASHSCOPE_API_KEY`	阿里云 DashScope API 密钥	是
`WORKSPACE`	工作目录路径，默认 `/workspace`	否

手动安装 AI 编码工具

本模板已预装 qwen-code CLI 工具，可直接使用：

# 使用 qwen-code 进行 NL 部署
abx run <sandbox-id> qwen --prompt "部署这个 FastAPI 项目"

# 或在沙箱内直接使用 qwen 命令
abx exec <sandbox-id> -- qwen -p "分析项目并部署" --yolo

如需安装额外工具：

# 安装额外 Python 包
abx exec <sandbox-id> -- pip install <agent-package>

# 安装额外 Node.js 工具
abx exec <sandbox-id> -- npm install -g <agent-tool>

注意事项

API Key 安全：请勿将 DASHSCOPE_API_KEY 硬编码到模板中，始终通过 --env 参数或环境变量注入。
网络访问：需要访问 DashScope API（dashscope.aliyuncs.com），确保沙箱具备外网连接能力。
兼容接口：DashScope 提供 OpenAI 兼容模式，可通过 openai 库直接调用 Qwen 模型。
虚拟环境：Python 包安装在 /opt/venv 中，使用时无需手动激活，PATH 已自动配置。

Deploy and local e2e

Install and register this template:

abx template install alibaba-agent-sandbox/awesome-templates//qwen-code

From this directory (build, push, and register):

abx deploy

From the repository root, with agent-sandbox-server and httpx installed:

python3 scripts/run_local_e2e.py

That starts python3 commands.py in this directory and runs e2e_test.py.

Sandbox runtime for the Qwen coding agent…