Spider-Sense: Intrinsic Risk Sensing for Efficient Agent Defense with Hierarchical Adaptive Screening
This addresses security inefficiencies for LLM agents, offering a selective defense approach that is incremental over mandatory checking methods.
The paper tackles the security challenges of LLM-based autonomous agents by proposing Spider-Sense, an intrinsic risk sensing framework that triggers defenses only upon risk perception, achieving the lowest Attack Success Rate and False Positive Rate with only 8.3% latency overhead.
As large language models (LLMs) evolve into autonomous agents, their real-world applicability has expanded significantly, accompanied by new security challenges. Most existing agent defense mechanisms adopt a mandatory checking paradigm, in which security validation is forcibly triggered at predefined stages of the agent lifecycle. In this work, we argue that effective agent security should be intrinsic and selective rather than architecturally decoupled and mandatory. We propose Spider-Sense framework, an event-driven defense framework based on Intrinsic Risk Sensing (IRS), which allows agents to maintain latent vigilance and trigger defenses only upon risk perception. Once triggered, the Spider-Sense invokes a hierarchical defence mechanism that trades off efficiency and precision: it resolves known patterns via lightweight similarity matching while escalating ambiguous cases to deep internal reasoning, thereby eliminating reliance on external models. To facilitate rigorous evaluation, we introduce S$^2$Bench, a lifecycle-aware benchmark featuring realistic tool execution and multi-stage attacks. Extensive experiments demonstrate that Spider-Sense achieves competitive or superior defense performance, attaining the lowest Attack Success Rate (ASR) and False Positive Rate (FPR), with only a marginal latency overhead of 8.3\%.