AngelosZou/dsh-python-env

Workspace-scoped Python virtual environment management for a DeepSeek Harness project — discover, create, install into, and remove virtual environments without sandbox, network, or subprocess pitfalls.

Project Overview项目介绍

This is a native DeepSeek Harness (DSH) plugin that enables agent-facing Python virtual environment management for DSH workspaces. It ships five functional tools: pyenv_discover, pyenv_create, pyenv_install, pyenv_uninstall, and pyenv_remove, along with a dedicated python-env skill and system prompt guidance for the agent. It runs python -m venv and pip through DSH’s host subprocess channel instead of a sandboxed shell, which avoids common pitfalls with network access and sandbox restrictions that often break pip commands. It also includes automatic mirror and proxy fallback for network issues, retrying across multiple public PyPI mirrors if the default index fails.

The plugin is designed for AI coding agents that use DSH for Python development, and it works cross-platform across Windows, macOS, and Linux. It automatically handles differences in path layout and interpreter chain naming between operating systems, such as Scripts folders on Windows versus bin folders on macOS and Linux. A typical workflow starts with the agent calling pyenv_discover to scan for existing virtual environments in the current workspace. If no suitable environment exists, the agent creates a new one with pyenv_create, installs required dependencies with pyenv_install, and removes packages or the entire environment when no longer needed.

To install the plugin from npm, run the command dsh plugin --profile web add dsh-python-env, then restart the DSH backend to load the new plugin properly. It requires Node.js version 20 or newer, a DSH profile built from @deepseek-ai/dsh-base, and Python 3.8 or newer available on the system PATH or specified explicitly by the user. It has no third-party dependencies, so you do not need to install uv, virtualenv, or any other extra tools to use it. It is released under the permissive open-source MIT license, with no cost for personal or commercial use.

这是一个原生 DeepSeek Harness (DSH) 插件,专为 DSH 项目提供工作区范围的 Python 虚拟环境管理能力。它提供五个核心工具:pyenv_discover、pyenv_create、pyenv_install、pyenv_uninstall、pyenv_remove,还附带专用的 python-env 技能和系统提示引导。它通过 DSH 的宿主进程子通道调用标准 python -m venv 和 pip,而非沙箱 shell,避开了常见的网络和执行陷阱。

该插件面向使用 DSH 进行 Python 开发的 AI 编程代理,适配 Windows、macOS、Linux 多平台,能够自动处理不同系统下的路径布局和解释器路径差异。典型工作流为:代理先调用 pyenv_discover 扫描当前工作区已有的虚拟环境,若没有则调用 pyenv_create 创建新环境,之后通过 pyenv_install 安装依赖包,完成开发后可卸载包或删除整个环境。

该插件依赖 Node.js 20 以上版本、基于 @deepseek-ai/dsh-base 构建的 DSH 配置,以及系统 PATH 中或可指定路径的 Python 3.8 以上版本,无第三方依赖,不需要 uv 或 virtualenv 等额外工具。它采用 MIT 开源许可证,安装方式为执行 dsh plugin --profile web add dsh-python-env,安装后需要重启 DSH 后端才能生效。

Pre-install check安装前体检Compatibility · Security兼容性 · 安全性 1 warning1 项注意
  • Only 4 stars - very few users, little community feedback星标只有 4,几乎没人在用,遇到问题缺少社区反馈
DSH walks through these 9 checksDSH 会逐条核对这 9 项

Compatibility兼容性

  • DSH, Node, OS and profile requirementsDSH 版本 / Node 版本 / 操作系统 / profile 是否满足要求
  • External dependencies and runtimes (Electron / Python / Docker, ...)外部依赖与运行时(Electron / Python / Docker 等)是否齐备
  • Conflicts with installed plugins: command names, skill / tool names, ports, duplicate MCP registration与已装插件是否冲突:命令名、skill / tool 重名、端口占用、重复 MCP 注册

Security安全性

  • Repo matches the facts registered here; archived or abandoned?仓库是否与页面登记一致,是否归档或长期停更
  • Safety of preinstall / install / postinstall and install.sh / setup.ps1preinstall / install / postinstall 与 install.sh、setup.ps1 是否安全
  • curl|bash, download-then-execute, obfuscation, unrelated domains → stop immediatelycurl|bash、下载即执行、混淆代码、无关域名 → 立刻停止
  • Typosquatting or unmaintained packages among the new dependencies新增依赖里有没有 typosquatting 或无人维护的包
  • Requested permissions vs. what the feature actually needs申请了哪些权限、是否超出功能所需(filesystem / network / shell / clipboard)
  • Any sudo / admin requirement, plus uninstall and rollback是否要求 sudo / 管理员权限,以及卸载与回滚方式

Anything uncertain must be marked unknown with a note on how to confirm it. This site's signal screen is a static snapshot, not a security audit.拿不准的必须标「未知」并说明要我怎么确认。本站的信号筛查是静态快照,不能替代安全审计。

Or use CLI install (for developers)或使用命令行安装(适合开发者)

CLI Install命令行安装

dsh plugin --profile web add dsh-python-env

把 AngelosZou/dsh-python-env 加入你的 DSH 配置(web profile)即可启用。

READMEREADME

dsh-python-env

English | 中文

Workspace-scoped Python virtual environment management for a DeepSeek Harness project — discover, create, install into, and remove virtual environments without sandbox, network, or subprocess pitfalls.

License: MIT Node.js >= 20 npm version GitHub issues Awesome DSH Plugin

A DeepSeek Harness plugin that gives one project (workspace) agent-facing Python virtual environment management:

  • Five model tools — pyenv_discover, pyenv_create, pyenv_install, pyenv_uninstall, pyenv_remove — plus the python-env skill and a system-prompt guidance section.
  • Runs the standard library python -m venv / pip through the platform subprocess channel (host process) instead of the sandboxed shell, so venv creation, ensurepip bootstrapping, and package-index network access work where shell-side pip fails.
  • Mirror and proxy fallback — on a network-classified failure, installs retry across PyPI mirrors (TUNA, Aliyun, USTC) and probe common local proxy ports; index / proxy arguments pin either.
  • Workspace confinement — every path resolves inside the session workspace (case-insensitive on Windows); caches and temp state live under <workspace>/.dsh-pyenv/; commands are argv arrays (no shell); the global Python environment, host pip cache, and system temp are never touched.
  • Cross-platform — Windows / macOS / Linux layouts and interpreter chains (Scripts vs bin, py -3 vs python3).
  • No third-party dependency — no uv, no virtualenv, no other plugin. A pip-less environment is repaired offline via ensurepip.
  • Session policy parity — the mutating tools consult the session's sandbox policy and refuse to run in read-only sessions; discovery stays available everywhere.

Showing the opening section of the README — the full document lives in the repository以上为 README 开头摘要,完整文档在仓库内 · View the full README on GitHub →在 GitHub 查看完整 README →

← 上一个 Prev dsh-tts-bridge 下一个 Next deepseek-web-import →