In a few clicks we can analyze your entire application and see what components are vulnerable in your application, and suggest you quick fixes.
Test your applicationsUpgrade @anthropic-ai/sandbox-runtime to version 0.0.16 or higher.
@anthropic-ai/sandbox-runtime is an Anthropic Sandbox Runtime (ASRT) - A general-purpose tool for wrapping security boundaries around arbitrary processes
Affected versions of this package are vulnerable to Protection Mechanism Failure due to improper enforcement of network sandboxing in the sandboxing logic. An attacker can bypass intended network restrictions by executing code that initiates network requests to unauthorized domains.
Note:
This is only exploitable if the sandbox policy does not configure any allowed domains.