Agent Skills: fine-tuning-with-trl

TRL: SFT, DPO, PPO, GRPO, reward modeling for LLM RLHF.