URML (open robot intent language): a safety envelope as an Ogma requirement input, request for comment

[idoco2003]

Hi Ogma maintainers,

URML (urml.dev) is a small, Apache-2.0 language for describing robot intent: it validates a request statically against a capability manifest and a safety envelope, then dispatches. Ogma is the closest existing tool I have found to what URML's envelope wants downstream: it turns formal requirements into runtime-monitoring applications, generates Copilot monitors, and already integrates with frameworks including ROS, which URML targets too.

Nothing here asks Ogma to change or maintain anything. This is a request for comment, and an exciting one from URML's side.

The shape: a URML safety envelope is a declared set of properties; Ogma is a path from declared requirements to a generated, framework-integrated monitor. Two real questions. First, what input does Ogma expect (FRET-style structured requirements, a specification file), and could a URML envelope be lowered to that form? Second, given Ogma's ROS integration, how should a generated monitor's verdict feed back into a URML-governed system, so a runtime violation becomes a first-class signal rather than a side channel?

Full write-up: https://github.com/URML-MARS/URML/blob/main/docs/rfcs/0365-ogma-outreach.md

Thanks for Ogma; closing the gap from requirements to deployed monitors is exactly the hard part.

Ido Yahalomi (URML, greenvh@gmail.com)

AI-assisted prose, maintainer-reviewed before posting (see VIBE.md). Human-only correspondence available on request.

[ivanperez-keera]

Thanks, @idoco2003 for the post. I'll transfer it to a discussion for now; we can create tickets later is we want to change something in Ogma to make a connection to URML easier.

We can also chat via email if preferable. My email is ivan.perezdominguez@nasa.gov.