openclaw-sandboxing

Manage Docker-based sandboxing for tool execution with configurable isolation modes and network policies.

2|Updated Mar 7, 2026
One-click install
npx skills add https://github.com/TheophilusChinomona/claude-openclaw-plugin --skill openclaw-sandboxing
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: openclaw-sandboxing
Source: https://github.com/TheophilusChinomona/claude-openclaw-plugin/tree/main/plugins/openclaw/skills/openclaw-sandboxing
Command: npx skills add https://github.com/TheophilusChinomona/claude-openclaw-plugin --skill openclaw-sandboxing

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill addresses the security risks associated with running arbitrary code or tools by providing robust sandboxing capabilities, ensuring that tool execution is isolated from the host system and other sensitive environments.

Core Features & Use Cases

  • Containerized Execution: Run tools within isolated Docker containers to minimize the blast radius of potential errors or malicious actions.
  • Configurable Isolation: Customize sandbox modes (off, non-main, all), scope (session, agent, shared), and workspace access (none, ro, rw).
  • Network Control: Define network access for sandboxes, from completely isolated (none) to bridged (bridge) for package installations.
  • Browser Sandboxing: Configure isolated browser environments for web-based tasks.
  • Use Case: When executing a script that downloads and runs external code, configure the sandbox to use mode: "all", scope: "session", and docker.network: "none" to ensure maximum security and prevent any unintended network access or host system interference.

Quick Start

Configure the sandbox to use all sessions, agent scope, and read-write workspace access with a bridge network for development.

Frequently Asked Questions about openclaw-sandboxing

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I sandbox tool execution using Docker containers?

You can sandbox tool execution by running tools within isolated Docker containers, configuring isolation modes, scopes, and network policies to prevent host system interference and minimize blast radius.

What Docker network policies can I set for isolated container execution?

Network policies for isolated container execution range from completely isolated with no network access to bridged networks, allowing you to control whether sandboxes can reach external resources for tasks like package installations.

Can I configure read-write workspace access for Docker sandboxes?

Yes, Docker sandboxes support configurable workspace access including none, read-only, and read-write modes, letting you precisely control how isolated containers interact with your host files and directories via bind mounts.

What is the best way to securely run downloaded external code in an isolated environment?

The best way to securely run downloaded external code is configuring a Docker sandbox with mode set to all, scope to session, and network to none, ensuring maximum security and preventing unintended network access.

Does Docker sandboxing support custom image builds for specific tool execution requirements?

Yes, Docker sandboxing supports custom Docker image builds, allowing you to create controlled environments tailored to specific tool execution requirements while maintaining strict isolation from the host system.

When should I use agent scope versus session scope for containerized tool execution?

Use session scope for isolated single-session tool execution, or agent scope when sandboxed tools need shared access across an agent's workflow, with options to configure shared scope for broader collaborative environments.