rl-reward-curriculum-design

What problem does it solve?

This legacy wrapper keeps older reward, curriculum, and domain-randomization prompts working by directing them to the narrower RLXP skills that handle each task safely and consistently.

Core Features & Use Cases

• Dispatches bounded reward tuning to the reward-tuning skill.
• Sends reward code changes to reward-engineering only when evidence and contract approval justify it.
• Covers curriculum design, DR probing, reward review, and candidate validation without expanding scope.
• Use Case: A researcher asks for help adjusting reward terms for a new behavior, and the skill routes the request to the right intervention path while preserving guardrails.

Quick Start

Ask for legacy RL reward or curriculum help, and let the skill route your request to the correct RLXP intervention path.