sandbox-manager

Audit and tune Claude Code sandbox settings to minimize permission prompts.

Updated Nov 2, 2025
One-click install
npx skills add https://github.com/SelrahcD/claude-marketplace --skill sandbox-manager-selrahcd
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: sandbox-manager
Source: https://github.com/SelrahcD/claude-marketplace/tree/main/selrahcd-dot-claude/skills/sandbox-manager
Command: npx skills add https://github.com/SelrahcD/claude-marketplace --skill sandbox-manager-selrahcd

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Interactively audit and tune Claude Code sandbox settings to minimize permission prompts while maintaining security. Every change must cite its source and explain the trade-off.

Core Features & Use Cases

  • Identify permission friction and read relevant settings files to diagnose why prompts occur.
  • Propose source-backed sandbox adjustments (allowWrite, allowedDomains, allowUnixSockets) with trade-offs and documentation references.
  • Apply changes after user approval, targeting user-global, project, or local settings to reduce prompts without compromising security.

Quick Start

Read current sandbox configuration, identify friction points, and propose precise, source-backed changes to reduce prompts.

Frequently Asked Questions about sandbox-manager

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

FAQPage Schema
How do I reduce frequent permission prompts in Claude Code sandbox?

Reduce frequent Claude Code permission prompts by reading settings files to diagnose friction points and applying source-backed sandbox adjustments like allowWrite or allowedDomains while maintaining security.

How does Claude Code sandbox configuration work across different scopes?

Claude Code sandbox configuration operates across user-global, project, and local scopes by reading respective settings files to apply targeted permission changes and reduce prompts without compromising security.

Can I configure allowed domains for tools like docker and kubectl in the sandbox?

Yes, you can configure allowed domains for tools like docker, kubectl, terraform, and gh by proposing source-backed sandbox adjustments to the settings files to minimize tool-specific permission prompts.

What is the best way to diagnose why Claude Code keeps asking for permissions?

Diagnose recurring Claude Code permission requests by reading the user-global, project, and local settings files to identify friction points and propose documented sandbox configuration changes with clear trade-offs.

Are there security trade-offs when adding allowed domains to Claude Code sandbox?

Adding allowed domains to the Claude Code sandbox involves security trade-offs, requiring every proposed change to cite its source and explain the trade-off to ensure permissions are minimized without compromising security.