It would be extremely valuable to support a sandboxed Python/JavaScript execution step within Port's automation/workflow engine. Many automation use cases—such as entity reconciliation, API enrichment, pagination, and multi-step business logic—cannot be implemented with JQ alone and currently require external infrastructure (CI pipelines, cron jobs, cloud functions, etc.).

A lightweight code step that receives the automation payload and a scoped Port API client would allow users to keep custom logic native to Port, making automations easier to build, maintain, observe, and version alongside their Port resources.

Having this capability within a self-service action input form would allow large organizations with large amounts of data to pull in targeted, up-to-date information from elsewhere just-in-time within a self-service action. This type of data is usually not relevant enough to be contained within the software catalog, and is therefore not a good candidate for an entity selection dropdown. The type of data that might be helpful to fetch quickly within a self-service action input would be a filtered subset of a ServiceNow table, for example, that would be otherwise simply taking up space in the catalog (since it would not require additional properties or ownership information, and would be used only during the course of a self-service action run).