Source Intelligence

DisclaimerUnofficial, and not affiliated with Anthropic. Nearly all of this is read straight out of what ships: npm bundles, captured prompts, published docs. Anthropic's own notes go in verbatim, marked as theirs. The rest is my reading, and every entry carries the strings behind it. If one looks wrong, vote it down and say why.

All of v2.0.30 Home All releases olderv2.0.29 v2.0.31newer
Claude Code v2.0.30

Enhanced Sandbox Security Policy

What: A new configuration option allowUnsandboxedCommands that provides policy-level control over the dangerouslyDisableSandbox parameter.

How to use:

// In your settings
{
  "sandbox": {
    "allowUnsandboxedCommands": false  // Enforce sandbox-only mode
  }
}

Details:

  • Default: true (maintains backward compatibility)
  • When set to false: All commands run in sandbox mode regardless of the dangerouslyDisableSandbox parameter
  • Enforcement: The security check at Q9A() at line 249229 validates both the parameter AND the policy setting
  • User communication: Claude receives different system prompt instructions based on this setting (lines 208919-208942)
  • Evidence: Configuration schema at line 473000, getter function XP8() at line 208521

Related

Other releases about the same thing. Found by shared names or similar wording; neither means one caused the other.

See this entry in the whole of v2.0.30 →