Proposal: URML (substrate-neutral robot intent) Layer-2 / Layer-3 → SkiROS skill semantics with knowledge-graph substrate declaration — plus license-clarification ask

[idoco2003]

Hi @RobotLabLTH team,

Proposing a URML v0.1 capability-manifest mapping for SkiROS2 over RobotLabLTH/SkiROS2. URML (Apache-2.0) is a substrate-neutral spec for robot intent: typed primitive vocabulary + capability manifest + static validator.

SkiROS is structurally distinctive among URML's Move #12 robot-command-library engagements: where the behavior-tree engagements provide execution semantics, SkiROS adds a knowledge-graph substrate (world model) that grounds skill parameters against scene state. URML's primitive vocabulary maps cleanly onto SkiROS skill classes; the knowledge-graph substrate surfaces a novel URML manifest concern (skill-grounding-mode declaration) URML is queueing as a Spec RFC.

Two housekeeping notes: (1) GitHub reports licenseInfo: Other on the repo — URML's adapter cannot ship until OSI classification is confirmed. (2) URML's internal notes had this project at Aalborg / RVMI historically; engaging the active RobotLabLTH/SkiROS2 (Lund University Robotics Lab / Department of Automatic Control) as the canonical upstream.

This is proposal-only, part of URML's Move #12 outreach (16 RFCs covering speech / translation / robot-command-library substrates for URML's NL layer).

Full RFC with manifest mapping, four alternatives, and the knowledge-graph-substrate design discussion: https://github.com/URML-MARS/URML/blob/main/docs/rfcs/0163-skiros2-outreach.md

Questions worth maintainer input on:

1. License clarification. What is the explicit OSI license URML should cite? Historical SkiROS shipped BSD-3-Clause; is that current?
2. Repo-location confirmation. Is RobotLabLTH/SkiROS2 the canonical upstream for URML to engage going forward (vs. historical Aalborg / RVMI lineage)?
3. Knowledge-graph substrate declaration. URML proposes adding a knowledge_graph_substrate manifest field. Is this a useful abstraction for SkiROS-class deployments, or should the knowledge graph stay SkiROS-internal?
4. Skill-grounding-mode declaration. Is world_model \| runtime_query \| hybrid the right enumeration, or do you see other modes worth declaring?
5. World-model seed path. Is a static YAML seed the canonical convention, or do deployments commonly use other seed sources?
6. Adapter home. URML's reference/skill-bridge/, contributed example in SkiROS2/examples/, or external bridge repo?
7. Anything else.

Ido Yahalomi (URML maintainer, urml.dev, greenvh@gmail.com)

---

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