Hierarchical Reinforcement Learning with Runtime Safety Shielding for Power Grid Operation

Create a 'Safety Shield' wrapper for RL agents that enforces strict operational constraints at runtime. This library should act as a hard-coded gatekeeper that sits between the agent's policy output and the environment actuator.

Suggested repo: shieldrl

"The first RL-agent wrapper that physically cannot break your system."

Estimated effort: 40h