EvoSafeHarness: Evolving Model- and Domain-Specific Harnesses for Securing Agents

• URL: http://arxiv.org/abs/2609.05903v1
• Date: Sat, 05 Sep 2026 05:56:41 GMT
• Title: EvoSafeHarness: Evolving Model- and Domain-Specific Harnesses for Securing Agents
• Abstract summary: EvoSafeHarness is a safety-specific optimization framework that synthesizes a deployable harness for a frozen model in a target domain.<n>It reduces average attack success rate from 45.6% to 10.0% at a 3.3-point utility cost.<n>It also achieves the best score on Agent-SafetyBench for every victim.
• Score: 74.54929751174289
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large Language Model (LLM) agents are turning language into real-world effects, making safety necessary against both indirect prompt injections and direct harmful requests. System-level safety harnesses add an enforcement layer beyond model-level defenses, but existing harnesses are usually designed once by experts and applied across heterogeneous models and domains. Effective protection is deployment-dependent: models differ in how much enforcement they need before utility declines, while domains differ in the effects, state, and action sequences that must be governed. A harness that is strict enough for one model may over-block another, and a policy that transfers across domains may miss application-specific safety relations. We present EvoSafeHarness, a safety-specific optimization framework that synthesizes a deployable harness for a frozen model in a target domain. It jointly searches a natural-language policy and executable code logic, guided by model behavior, domain specifications, and fresh-context adversarial review to reject benchmark-specific rules. Across four agent benchmark families, EvoSafeHarness achieves a stronger safety-utility frontier than fixed expert-designed defenses. On DecodingTrust-Agent, it reduces average attack success rate from 45.6% to 10.0% at a 3.3-point utility cost and achieves the best score in 14 of 15 cells. On AgentDojo, it reaches 82.8% utility at 0.0% ASR, twice CaMeL's utility at the same operating point, and transfers unchanged to unseen AgentDyn suites. It also achieves the best score on Agent-SafetyBench for every victim and keeps mean ASR below 20% under adaptive PAIR attacks with a refinement budget of 16. Analysis shows that domain semantics determine which safety relations and trajectory state are needed, while model and runtime behavior determine how and where those relations should be enforced.

Related papers

• BLINDSPOT: A Benchmark for Safety and Refusal Calibration in Long-Horizon Tool-Using Agents [20.217145643468598]
  We introduce Blindspot, a benchmark for trajectory-level safety calibration of long-horizon tool-using agents.<n>Blindspot evaluates complete user-agent-environment trajectories through adaptive adversarial interaction, stateful tool execution, and execution-grounded adjudication.<n>We evaluate 13 proprietary and open-weight LLMs using eight metrics covering unsafe completion, appropriate refusal, benign utility, over-refusal, repeated-run robustness, and post-refusal failure.
  arXiv  Detail & Related papers  (2026-09-14T20:12:27Z)
• HazardAuditor: From Executable Threats to Safer Computer-Use Agents [50.06351661912873]
  Existing guard models target static prompts and responses.<n>Existing safety platforms produce evaluation verdicts rather than the normalized supervision a guard model needs to learn.<n>We introduce HazardAuditor, an execution-grounded framework that closes both gaps.
  arXiv  Detail & Related papers  (2026-09-14T07:09:33Z)
• HarnessRisk: A Lifecycle-Oriented Benchmark for Agent Harness Safety [45.163264704448125]
  We present a lifecycle oriented benchmark that organizes agent safety into six operational phases.<n>We evaluate each trajectory using Utility, Attack Success Rate, Persistence, and Detection.<n>We find that explicit risk recognition does not reliably lead to safe action, as some configurations detect risks in more than 90% of runs.
  arXiv  Detail & Related papers  (2026-08-18T10:03:58Z)
• Selective Safety Steering via Value-Filtered Decoding [54.87935112120107]
  Large language models (LLMs) are trained to align with human values, but their generations may still violate safety constraints.<n>Existing decoding-time steering methods often intervene unnecessarily, modifying generations that would have been safe under the base model.<n>We propose a new test-time steering method designed to reduce such unnecessary interventions while improving the safety of unsafe responses.
  arXiv  Detail & Related papers  (2026-05-14T12:13:08Z)
• On-Policy Self-Evolution via Failure Trajectories for Agentic Safety Alignment [54.30690671490447]
  Existing safety-alignment signals are largely response-level or off-policy.<n>FATE transforms verifier-scored failures into repair supervision without expert demonstrations.<n>FATE reduces attack success rate by 33.5%, harmful compliance by 82.6%, and improves external trajectory-safety diagnosis by 6.5%.
  arXiv  Detail & Related papers  (2026-05-12T09:56:28Z)
• Spider-Sense: Intrinsic Risk Sensing for Efficient Agent Defense with Hierarchical Adaptive Screening [23.066685616914807]
  We argue that effective agent security should be intrinsic and selective rather than architecturally decoupled and mandatory.<n>We propose Spider-Sense framework, which allows agents to maintain latent vigilance and trigger defenses only upon risk perception.<n>Spider-Sense achieves competitive or superior defense performance, attaining the lowest Attack Success Rate (ASR) and False Positive Rate (FPR)
  arXiv  Detail & Related papers  (2026-02-05T07:11:05Z)
• ReasAlign: Reasoning Enhanced Safety Alignment against Prompt Injection Attack [52.17935054046577]
  We present ReasAlign, a model-level solution to improve safety alignment against indirect prompt injection attacks.<n>ReasAlign incorporates structured reasoning steps to analyze user queries, detect conflicting instructions, and preserve the continuity of the user's intended tasks.
  arXiv  Detail & Related papers  (2026-01-15T08:23:38Z)
• SaFeR-VLM: Toward Safety-aware Fine-grained Reasoning in Multimodal Models [66.71948519280669]
  Multimodal Large Reasoning Models (MLRMs) demonstrate impressive crossmodal reasoning but often amplify safety risks under adversarial prompts.<n> Existing defenses mainly act at the output level and do not constrain the reasoning process, leaving models to implicit risks.<n>We propose SaFeR-VLM, which integrates four components and supports dynamic and interpretable safety decisions beyond surface-level filtering.
  arXiv  Detail & Related papers  (2025-10-08T10:39:12Z)

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.