HN Showcase
Back to browse
GitHub Repository

Run Claude Code autonomously with --dangerously-skip-permissions in an isolated Docker container.

6 starsShell

Run Claude Code with –dangerously-skip-permissions in a Docker sandbox

by sayil·Mar 19, 2026·2 points·2 comments
Visit ProjectView on HN

AI Analysis

●●SolidSolve My ProblemShip It

Docker sandbox for Claude Code's dangerous flag when Anthropic won't let you run it bare.

Strengths
  • •Non-root container user with credential passthrough from ~/.claude.json keeps host machine safe.
  • •One-command install and run from any project folder — zero configuration overhead.
  • •Fresh container per session ensures clean state and prevents credential leakage.
Weaknesses
  • •Thin wrapper around Docker — no novel architecture beyond mounting volumes and passing flags.
  • •Only supports Claude Code; doesn't generalize to other AI agents needing sandboxing.
Target Audience

Developers using Claude Code for autonomous agent workflows

Similar To

Dev Containers · GitHub Codespaces · Docker CLI wrappers

Post Description

Claude Code's --dangerously-skip-permissions flag lets agents run autonomously without interruption, but Anthropic warns you to only use it inside a sandboxed environment.

dangerously spins up an isolated Docker container, mounts your current project directory, and launches Claude Code inside it. This ensures file system changes are contained to your project.

One command to install, run it from any project folder: npm install -g dangerously dangerously

Similar Projects