To run user and AI generated code safely, we take appropriate security
precautions. These are:
Your code is run on an isolated worker within a sandbox. This isolates
your code from other users' code.
Every automation is started in a new sandbox.
Your code do not have access to a file system.
Every execution is subject to resource limits which prevents malicious
code from doing extended harm, as well as limiting the amount of
billed minutes that a single execution can use.
Automation inputs and outputs are not logged unless specified by
your code.
