rl-experiment-plan

What problem does it solve?

This Skill prevents premature or poorly scoped reinforcement-learning training by turning vague experiment ideas into an explicit task, metric, budget, and approval contract before any launch.

Core Features & Use Cases

• Contract Creation: Guides the creation or update of the active task/study contract that records scope, metrics, guardrails, budgets, hardware, and evaluation rules.
• Baseline Planning: Helps prepare baseline, smoke, and evaluation runs without changing reward logic, curriculum, domain randomization, or task code.
• Legacy Compatibility: Supports older prompts by dispatching work to narrower RLXP planning skills while keeping the training gate explicit.
• Use Case: A researcher with a new RL project can use this Skill to formalize the study before running any compute-heavy experiment.

Quick Start

Use this skill to convert the current reinforcement-learning project into a validated study contract and baseline plan before any training starts.