판 이력 — CAITLYN: Can LLM Agents Autonomously Synthesize Defenses against Emerging Injection Attacks? | AIChainDay