File-Based Output Mode Solves Large Step Data

output: mode is a hidden but vital tool for handling big data in step workflows. Right now, when step outputs exceed 1MB - like huge GitHub API responses or full log dumps - ${step_id.stdout} file paths are the only fallback, but they won’t work inside scripts or commands. That forces users into awkward workarounds, risking lost data or broken pipelines.

This file-based mode solves the problem by resolving output to a stable, full file path - directly usable as a source in later steps. It’s not just about storage; it’s about trust: large payloads stay intact, no truncation, no parsing headaches.

Culturally, this aligns with how US tech teams handle big data - preferring reliable access over shaky shortcuts. Think of it like moving from a thumb drive to a secure cloud folder: consistent, safe, and scalable.

But here is a catch: file paths are location-sensitive. Always validate the path before using it - missing files or permissions can break your step silently, even if output resolves correctly.

The bottom line: when large outputs dominate, switch to file mode. Your data remains whole, your pipelines stay robust. Do you trust your step outputs to stay accessible - even at scale?