HARDE Framework Optimizes Agent Harnesses for Risk Detection

HARDE introduces a two-stage optimization framework for LLM agent harnesses to mitigate injection and safety risks. The system uses a trigger-monitor-feedback loop to enable targeted interventions during runtime without disrupting benign tasks.