TRACES: Proactive Safety Auditing for Multi-Turn LLM Agents via Trajectory-State Modeling

• URL: http://arxiv.org/abs/2605.27690v1
• Date: Tue, 26 May 2026 21:11:02 GMT
• Title: TRACES: Proactive Safety Auditing for Multi-Turn LLM Agents via Trajectory-State Modeling
• Abstract summary: LLM agents increasingly operate through multi-turn tool use and environment interaction.<n>Post-hoc diagnosis frequently misses the chance to flag risks while they are unfolding.<n>We propose TRACES, a representation-based proactive auditor.
• Score: 18.58206519209223
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: LLM agents increasingly operate through multi-turn tool use and environment interaction, where safety risks often emerge from intermediate steps long before they surface in the final outcome. Reactive auditing is therefore insufficient: post-hoc diagnosis frequently misses the chance to flag risks while they are unfolding. We propose TRACES, a representation-based proactive auditor that learns prefix-level trajectory risk states from the hidden representations of an observer LLM. TRACES induces latent mechanism features from step representations and models their temporal evolution to estimate whether a partial trajectory is drifting toward unsafe behavior. To sidestep the cost and ambiguity of step-level risk annotation, TRACES is trained with weak trajectory-level supervision while still producing dense prefix-level risk estimates. Across multiple agent safety benchmarks, TRACES improves both full-trajectory safety prediction and proactive risk discrimination. Our analyses further suggest that these risk states can help train a safer agent, highlighting the broader potential of proactive auditing for long-horizon agent safety.

Related papers

• AURA-Eval: Evaluation Framework for Acting Under Risk Awareness in LLM Agent Trajectories [12.694014584212695]
  We introduce AURA-Eval, a framework combining controlled augmentation with granular diagnosis of behavior in tool-use trajectories.<n>Using 157 sourced trajectories, we generate 1,249 evaluation items and evaluate 20 frontier and open-weight models.<n>Our results show that LLM agents engage in unsafe behavior more often when no safe path exists.
  arXiv  Detail & Related papers  (2026-09-06T18:58:56Z)
• Forecasting Trajectory-Level Safety Risks in Black-Box Multi-Turn Interactions [13.366980144986707]
  We propose Recast, a safety risk forecasting framework for large language models (LLMs)<n>Recast predicts 88.3% of future safety failures with an average lead time of 2.41 turns, while maintaining a false alarm rate of 12.3%.
  arXiv  Detail & Related papers  (2026-07-29T12:14:23Z)
• AgentLens: Interpretable Safety Steering via Mechanistic Subspaces for Multi-Turn Coding Agent [53.82005364479556]
  Coding agents based on large language models (LLMs) demonstrate remarkable autonomous capabilities.<n>Existing safety mechanisms mainly rely on external guardrails.<n>We propose AgentLens, a white-box defense framework that performs runtime safety detection and representation-level mitigation.
  arXiv  Detail & Related papers  (2026-06-21T21:13:55Z)
• Enhancing Agent Safety Judgment: Controlled Benchmark Rewriting and Analogical Reasoning for Deceptive Out-of-Distribution Scenarios [0.2578242050187029]
  ROME is a controlled benchmark-construction pipeline that rewrites known unsafe trajectories into deceptive evaluation instances.<n>ARISE is a retrieval-guided inference-time enhancement that retrieves ReAct-style analogical safety trajectories from an external analogical base and injects them as structured reasoning exemplars.
  arXiv  Detail & Related papers  (2026-05-05T00:21:00Z)
• DRAFT: Task Decoupled Latent Reasoning for Agent Safety [59.46137757545185]
  We propose DRAFT (Task Decoupled Latent Reasoning for Agent Safety), a latent reasoning framework that decouples safety judgment into two trainable stages.<n>An Extractor distills the full trajectory into a compact continuous latent draft, and a Reasoner that jointly attends to the draft and the original trajectory to predict safety.<n>DRAFT consistently outperforms strong baselines, improving accuracy from 63.27% (LoRA) to 91.18% averaged over benchmarks.
  arXiv  Detail & Related papers  (2026-02-11T07:45:14Z)
• SafePred: A Predictive Guardrail for Computer-Using Agents via World Models [12.569157125705052]
  We present SafePred, a predictive guardrail framework for Computer-using Agents (CUAs) in complex real-world environments.<n>Based on this approach, we present SafePred that establishes a risk-to-decision loop to ensure safe agent behavior.<n>Extensive experiments show that SafePred significantly reduces high-risk behaviors, achieving over 97.6% safety performance and improving task utility by up to 21.4% compared with reactive baselines.
  arXiv  Detail & Related papers  (2026-02-02T07:04:06Z)
• ToolSafe: Enhancing Tool Invocation Safety of LLM-based agents via Proactive Step-level Guardrail and Feedback [53.2744585868162]
  Monitoring step-level tool invocation behaviors in real time is critical for agent deployment.<n>We first construct TS-Bench, a novel benchmark for step-level tool invocation safety detection in LLM agents.<n>We then develop a guardrail model, TS-Guard, using multi-task reinforcement learning.
  arXiv  Detail & Related papers  (2026-01-15T07:54:32Z)
• Anchoring Refusal Direction: Mitigating Safety Risks in Tuning via Projection Constraint [52.878820730054365]
  Instruction Fine-Tuning (IFT) has been widely adopted as an effective post-training strategy to enhance abilities of Large Language Models (LLMs)<n>Recent research into the internal mechanisms of LLMs has identified the refusal direction (r-direction) in the hidden states, which plays a pivotal role in governing refusal behavior.<n>To mitigate such drift, our proposed ProCon method introduces a projection-constrained loss term that regularizes the projection magnitude of each training sample's hidden state onto the r-direction.
  arXiv  Detail & Related papers  (2025-09-08T15:24:33Z)
• LM Agents May Fail to Act on Their Own Risk Knowledge [15.60032437959883]
  Language model (LM) agents pose a diverse array of potential, severe risks in safety-critical scenarios.<n>While they often answer "Yes" to queries like "Is executing sudo rm -rf /*' dangerous?", they will likely fail to identify such risks in instantiated trajectories.
  arXiv  Detail & Related papers  (2025-08-19T02:46:08Z)
• AGENTSAFE: Benchmarking the Safety of Embodied Agents on Hazardous Instructions [64.85086226439954]
  We present SAFE, a benchmark for assessing the safety of embodied VLM agents on hazardous instructions.<n> SAFE comprises three components: SAFE-THOR, SAFE-VERSE, and SAFE-DIAGNOSE.<n>We uncover systematic failures in translating hazard recognition into safe planning and execution.
  arXiv  Detail & Related papers  (2025-06-17T16:37:35Z)
• AgentAuditor: Human-Level Safety and Security Evaluation for LLM Agents [48.925168866726814]
  AgentAuditor is a universal, training-free, memory-augmented reasoning framework.<n>ASSEBench is the first benchmark designed to check how well LLM-based evaluators can spot both safety risks and security threats.
  arXiv  Detail & Related papers  (2025-05-31T17:10:23Z)
• AGrail: A Lifelong Agent Guardrail with Effective and Adaptive Safety Detection [47.83354878065321]
  We propose AGrail, a lifelong guardrail to enhance agent safety.<n>AGrail features adaptive safety check generation, effective safety check optimization, and tool compatibility and flexibility.
  arXiv  Detail & Related papers  (2025-02-17T05:12:33Z)
• Preemptive Detection and Correction of Misaligned Actions in LLM Agents [58.39520480675366]
  InferAct is a novel approach to detect misaligned actions before execution.<n>It alerts users for timely correction, preventing adverse outcomes.<n>InferAct achieves up to 20% improvements on Marco-F1 against baselines in misaligned action detection.
  arXiv  Detail & Related papers  (2024-07-16T15:24:44Z)
• SMARLA: A Safety Monitoring Approach for Deep Reinforcement Learning Agents [7.33319373357049]
  This paper introduces SMARLA, a black-box safety monitoring approach specifically designed for Deep Reinforcement Learning (DRL) agents. SMARLA utilizes machine learning to predict safety violations by observing the agent's behavior during execution. Empirical results reveal that SMARLA is accurate at predicting safety violations, with a low false positive rate, and can predict violations at an early stage, approximately halfway through the execution of the agent, before violations occur.
  arXiv  Detail & Related papers  (2023-08-03T21:08:51Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.