sue-config

What problem does it solve?

This Skill eliminates the risk of misalignment and failure across DeepResearch ML/HPC scale-up experiment workflows caused by inconsistent, invalid, or fragmented workspace configuration in the runtime.yaml contract.

Core Features & Use Cases

• Single Contract Normalization: Combines reusable policy/schema and concrete project values into a single, structured runtime.yaml file with strict separation between the two namespaces.
• Workspace Setup & Migration: Handles new workspace initialization, legacy config migration, and repair of missing or invalid configuration keys to meet DeepResearch contract requirements.
• Validation & Guardrails: Enforces DeepResearch root path validation, inode-light output rules, sandbox backend checks, and policy/concrete separation rules to prevent downstream workflow errors.
• Use Case: A research team running large-scale language model training on the Snellius supercomputer can use this skill to set up a valid runtime.yaml contract so all subsequent SUE steps (script writing, dryrun, fullrun, result summarization) use consistent, validated configuration without manual rework.

Quick Start

Use the sue-config skill to normalize or repair the runtime.yaml configuration contract for your DeepResearch scale-up experiment workspace.