deepseek-harness

中文版 | English

English

Sandbox runtime for a DeepSeek model-driven AI coding agent, supporting code generation, completion, and intelligent programming assistance.

Environment

This template comes preinstalled with the following tools and dependencies:

Installation

Install from a local path:

abx template install ./deepseek-harness --registry-type local

Install from GitHub:

abx template install alibaba-agent-sandbox/awesome-templates//deepseek-harness

Usage

Create a sandbox instance and pass in the API key:

abx create --template deepseek-harness --env DEEPSEEK_API_KEY=sk-xxx

Call a DeepSeek model inside the sandbox via the OpenAI-compatible interface:

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

Using the Python SDK:

from agent_sandbox import Sandbox

sandbox = Sandbox.create_sync(
    template="deepseek-harness",
    env={"DEEPSEEK_API_KEY": "sk-xxx"}
)

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

Configuration

Environment Variables

Manual Agent Plugin Installation

This template provides a base development runtime. To install a specific agent plugin or framework:

# Install a Python agent framework
abx exec <sandbox-id> -- pip install langchain langgraph

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

Notes

• API key security: Never hardcode DEEPSEEK_API_KEY into the template; always inject it via the --env option or environment variables.
• API compatibility: DeepSeek offers an OpenAI-compatible interface with the base URL https://api.deepseek.com.
• Network access: The DeepSeek API must be reachable — make sure the sandbox has outbound internet connectivity.
• Dual runtimes: The environment supports both Python 3.12 and Node.js 22 — pick whichever fits your plugin needs.
• Virtual environment: Python packages are installed in /opt/venv; no manual activation is needed since PATH is preconfigured.

中文

DeepSeek 模型驱动的 AI 编程 Agent 沙箱运行环境，支持代码生成、补全和智能编程辅助。

环境说明

本模板预装以下工具和依赖：

安装方式

从本地安装：

abx template install ./deepseek-harness --registry-type local

从 GitHub 安装：

abx template install alibaba-agent-sandbox/awesome-templates//deepseek-harness

使用示例

创建沙箱实例并传入 API Key：

abx create --template deepseek-harness --env DEEPSEEK_API_KEY=sk-xxx

在沙箱中通过 OpenAI 兼容接口调用 DeepSeek 模型：

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

使用 Python SDK：

from agent_sandbox import Sandbox

sandbox = Sandbox.create_sync(
    template="deepseek-harness",
    env={"DEEPSEEK_API_KEY": "sk-xxx"}
)

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

配置说明

环境变量

手动安装 Agent 插件

本模板提供基础开发运行时，如需安装特定 Agent 插件或框架：

# 安装 Python Agent 框架
abx exec <sandbox-id> -- pip install langchain langgraph

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

注意事项

• API Key 安全：请勿将 DEEPSEEK_API_KEY 硬编码到模板中，始终通过 --env 参数或环境变量注入。
• API 兼容性：DeepSeek 提供 OpenAI 兼容接口，base URL 为 https://api.deepseek.com。
• 网络访问：需要访问 DeepSeek API，确保沙箱具备外网连接能力。
• 双运行时：环境同时支持 Python 3.12 和 Node.js 22，可根据插件需求选择。
• 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//deepseek-harness

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.