Skip to content
Open Aghata

Work

Sandboxed code and generated files

Where code runs, what it can touch, and where the files it produces live.

When a task needs computation — parsing exports, reconciling datasets, generating a chart, assembling a document — Aghata writes and runs code in an isolated sandbox.

An ephemeral, isolated environment created for the session. It can compute over the files in that session; it is not your machine, has no access to your device, your other sessions, or your credentials, and disappears when the work is done.

Files. Anything the code produces that matters — the cleaned dataset, the report, the chart — is saved as a generated file attached to the session and stored in your Library. The code that produced a file is visible in the session, so a number in a report can be traced to the computation that made it.

Code steps appear like any other step: the code, its output or error, and the files it wrote. You don’t need to read the code — but you always can, and for anything you’d bet money on, you should be able to. That’s the point.

  • No network access from code except through Aghata’s own tool layer — a script can’t quietly call out to services you haven’t connected.
  • Long computations count toward your plan’s usage like any other work.
  • The sandbox is not a place to run your production jobs; it’s where task work happens. For recurring computation, wrap it in a Routine or Schedule.