Claude Skill
5
🐀 Fuzz your verifier before an RL agent does. Static + dynamic LLM security auditor to detect reward-hacking in RL post-training environmen...
Python MIT Updated 3d ago
freakyady
0
0