Sandboxing - Claude Code
Filesystem and network isolation for Bash tool calls on Linux and macOS.
Sandboxing confines Bash tool execution to an allowlist of paths and hosts. It's how you let Claude run commands without granting the shell full reach into your machine.
What it does
On Linux and macOS, Claude Code can launch Bash calls inside a sandbox that restricts filesystem access to the project directory and network access to a configured allowlist. Commands that try to reach outside the allowed area fail fast with a clear error. This is the strongest containment option short of a full container.
When to use it
- Granting broad Bash access without worrying about rogue commands.
- Running untrusted or generated scripts safely.
- Multi-tenant environments where you can't trust every skill or prompt.
- Any setup where you want defense-in-depth on top of permission rules.
Gotchas
- Some tools need network access you might not have whitelisted. Watch for "blocked by sandbox" errors and allowlist specifically.
- Sandboxing has performance overhead - small for most work, noticeable for file-heavy tasks.
- Not available on Windows. Use WSL2 or a container instead.
Official docs: https://code.claude.com/docs/en/sandboxing.md
Technical content at the intersection of AI and development. Building with AI agents, Claude Code, and modern dev tools - then showing you exactly how it works.
Was this helpful?
Related Guides
Related Tools
Claude Code
Anthropic's agentic coding CLI. Runs in your terminal, edits files autonomously, spawns sub-agents, and maintains memory...
View Tool
Claude Agent SDK
Anthropic's SDK for building agents on the same tools, agent loop, and context management that power Claude Code. Python...
View ToolCodeburn
Interactive TUI dashboard that shows exactly where your Claude Code and Cursor tokens are going, in real time.
View ToolConductor
Mac app for running parallel Claude Code, Codex, and Cursor agents in isolated workspaces. Watch every agent work at onc...
View ToolRelated Videos
Generate Videos in Codex + Claude Code with This...
Check out HeyGen! https://heygen.1stcollab.com/developersdigest The video introduces HeyGen, an AI video generation platform with a rich API and a CLI that lets developers create, fetch, and manipula...
10x Design in Claude Code and Codex
Try Higgsfield: https://higgsfield.ai/s/higgsfield-general-campaign-developersdigest-juDMTi AI coding agents can build entire websites in minutes, but the default results often still look generic and...
Self Improving Applications with Claude Code & Codex
Build a Self‑Improving Next.js App: Supabase + Vercel Eve + GitHub Issues + Agent Loops Check out Supabase: https://supabase.plug.dev/1wWOTGS The video demonstrates how to build and deploy a self-im...
Related Posts
Claude Fable 5.1 in Claude Code: How to Switch, Pin, and Pay Less
Claude Code v2.1.257 points /model fable at Fable 5.1: 1M context, $10/$50 per MTok, $0.25 cache reads. Setup, gotchas,...
Claude Design, Slides and Docs in Any Chat: How to Use Them
Claude Design, Slides and Docs now open in any Claude chat, Claude Code session or the Artifacts tab, on every plan. Exa...
Claude for Startups Offer Paused: What Founders Do Now
Claude for Startups paused its free Claude Team year and $1,000 API credit offer at capacity. What the program page says...
Claude Advisor Tool vs a Bigger Model: When the Pairing Pays
Anthropic's own runs: a Fable 5.1 advisor added 1.7 points to Opus 5.5 for 2.1x the cost. When the Claude advisor tool p...
Claude Code Retry Watchdog: Keep Unattended Runs Alive
CLAUDE_CODE_RETRY_WATCHDOG=1 makes Claude Code wait out 429 and 529 errors instead of failing after 10 retries. When to...
Claude Code vs Codex vs Copilot CLI: Local Sandboxes Compared
Copilot CLI local sandboxing is GA. How it compares with Claude Code and Codex on defaults, network, credentials, Window...
