The Agent Skills Directory

[SAFE]: The skill provides documentation and implementation examples for the official Anthropic code execution beta tool. All described behaviors, such as file uploads via the Files API and Python execution in a hosted sandbox, are standard features of the platform.
[SAFE]: The documentation includes security best practices, such as advising users to scrub outputs for secrets and noting the isolated nature of the sandbox.
[SAFE]: Code examples demonstrate legitimate use of established libraries (pandas, matplotlib, numpy) and standard API integration patterns without any malicious patterns or unauthorized access attempts.

code-execution