AI Agent Hub
Back to plugins
🤖

dsh-tensorlake-sandbox

Model Inference Updated 2026.08.21

Run the following command in DeepSeek Harness:

dsh plugin install tensorlakeai/dsh-tensorlake-sandbox

Paste the following prompt into your AI chat to install this plugin:

In DeepSeek Harness, use the command dsh plugin install tensorlakeai/dsh-tensorlake-sandbox to install the plugin from the full source address https://github.com/tensorlakeai/dsh-tensorlake-sandbox.

About this plugin

When using DeepSeek Harness for model inference, file operations and subprocess executions typically run locally, which can pose security risks and environment inconsistencies, such as compromising the host system with untrusted code. The dsh-tensorlake-sandbox plugin addresses these challenges by moving DeepSeek Harness's file, subprocess, Bash, terminal, and LSP operations into a short-lived Tensorlake microVM sandbox, providing a secure remote environment that ensures reproducible execution and resource control. Users can configure sandbox resources like CPU, memory, and timeout to suit different needs. This plugin is ideal for developers and teams using DeepSeek Harness, especially those requiring portable, isolated execution environments to enhance security and streamline workflows, allowing them to focus on model inference tasks.

Use Cases

  • Securely execute DeepSeek Harness file and Bash operations in a Tensorlake microVM.
  • Provide reproducible isolated development environments for model inference tasks.
  • Manage subprocess and terminal operations via remote sandboxes.

Best For

  • Developers and engineers using DeepSeek Harness for model inference.
  • AI research teams requiring secure remote execution environments.
  • Software developers seeking workflow isolation and reproducibility.