ActBench: Self-Evolving Benchmark of Behavioral Safety in Cowork Agents

• URL: http://arxiv.org/abs/2608.09476v1
• Date: Mon, 10 Aug 2026 11:45:03 GMT
• Title: ActBench: Self-Evolving Benchmark of Behavioral Safety in Cowork Agents
• Abstract summary: Cowork agents may complete benign tasks while disclosing protected data, manipulating unauthorized state, invocate unauthorized API.<n>We define behavioral safety and introduce ActBench, a benchmark that evaluates such behavior risk from execution trajectories rather than final responses.<n>ActBench contains 600 cases from 213 scenarios, spanning 15 risk behaviors, six execution spaces, and 48 web-service APIs.
• Score: 23.994328225666663
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Cowork agents may complete benign tasks while disclosing protected data, manipulating unauthorized state, invocate unauthorized API. We define behavioral safety and introduce ActBench, a self-evolving benchmark that evaluates such behavior risk from execution trajectories rather than final responses. Each case pairs a benign task with an adversarial variant that preserves its instruction, configuration, initial state, rating model, and trusted records while injecting a task-reachable payload. ActBench contains 600 cases from 213 scenarios, spanning 15 risk behaviors, six execution spaces, and 48 web-service APIs.To move beyond static payloads, we propose a reward-guided beam search method that jointly optimizes attack effectiveness and task utility, while reflection diagnoses failed execution checkpoint and guides payload revision. Besides, we propose a dual evidence verification mechanism that verifies agent execution safety and utility through log evidence and LLM-based trajectory evidence.We evaluate 15 LLMs and 6 open-source cowork agents over 24,000 trajectories. Under a fixed harness, attack success rates ranges from 10.1% to 94.4% across models, while under a fixed base model, they range from 73.7% to 94.4% across agents.These results show greater variation across models than agent harness, while attacks remain highly successful across all tested harnesses.Our benchmark is released at: https://github.com/zjuicsr/ActBench.

Related papers

• BenchShield: Formal Model-Backed Instrumentation for Reward Integrity in LLM-Agent Evaluation Infrastructure [47.61632934139226]
  LM-agent benchmarks increasingly function as interactive evaluation infrastructure.<n>This interactivity makes evaluations vulnerable to reward hacking.<n>This paper presents BenchShield, a model-backed instrumentation layer for reward integrity.
  arXiv  Detail & Related papers  (2026-09-10T03:10:58Z)
• EvoSafeHarness: Evolving Model- and Domain-Specific Harnesses for Securing Agents [74.54929751174289]
  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.
  arXiv  Detail & Related papers  (2026-09-05T05:56:41Z)
• StepGuard: Learning Step-Level Guardrails with Scalable Supervision and Safety-Utility Balancing [80.39644823192113]
  We propose StepGuard, a step-level guard model that can audit completed agent trajectories and check tool actions before they are executed.<n>StepGuard achieves the highest average accuracy among open-weight guard models, with performance comparable to GPT-5.4.<n>When used to guard agents on AgentDojo and AgentDyn, StepGuard reduces mean attack success rate by 77.3%.
  arXiv  Detail & Related papers  (2026-08-25T16:17:27Z)
• AgentLens: Revealing The Lucky Pass Problem in SWE-Agent Evaluation [11.272830796781925]
  We evaluate 2,614 OpenHands trajectories from eight model backends on 60 SWE-bench Verified tasks.<n>Among passing trajectories in this subset, 10.7% exhibit behavior we call a Lucky Pass.<n>We introduce AgentLens, a framework for process-level assessment of SWE-agent trajectories.
  arXiv  Detail & Related papers  (2026-05-13T03:00:57Z)
• 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)
• Behavioral Integrity Verification for AI Agent Skills [9.127363793428119]
  We formalize this as the behavioral integrity verification (BIV) problem.<n>On 49,943 skills from the OpenClaw registry, the deviation taxonomy reveals a pervasive description-implementation gap.<n>On a 906-skill malicious-skill detection benchmark, BIV reaches an F1 of 0.946, outperforming state-of-the-art rule-based and single-pass LLM baselines.
  arXiv  Detail & Related papers  (2026-05-12T08:41:09Z)
• SafetyDrift: Predicting When AI Agents Cross the Line Before They Actually Do [0.0]
  SafetyDrift models agent safety trajectories as absorbing Markov chains.<n>In communication tasks, agents that reach even a mild risk state have an 85% chance of violating safety within five steps.<n>A monitor built on these models detects 94.7% of violations with 3.7 steps of advance warning at negligible computational cost.
  arXiv  Detail & Related papers  (2026-03-28T05:52:04Z)
• Measuring and Exploiting Confirmation Bias in LLM-Assisted Security Code Review [6.417595678110472]
  We study whether confirmation bias affects LLM-based vulnerability detection, and whether this failure mode can be exploited in software supply-chain attacks.<n>Study 1 quantifies confirmation bias through controlled experiments on 250 CVE vulnerability/patch pairs evaluated across four state-of-the-art models under five framing conditions for the review prompt.<n>Study 2 evaluates exploitability in practice mimicking adversarial pull requests that reintroduce known vulnerabilities while framed as security improvements or urgent functionality fixes via their pull request metadata.
  arXiv  Detail & Related papers  (2026-03-19T10:40:27Z)
• 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)
• AI Security Beyond Core Domains: Resume Screening as a Case Study of Adversarial Vulnerabilities in Specialized LLM Applications [71.27518152526686]
  Large Language Models (LLMs) excel at text comprehension and generation, making them ideal for automated tasks like code review and content moderation.<n>LLMs can be manipulated by "adversarial instructions" hidden in input data, such as resumes or code, causing them to deviate from their intended task.<n>This paper introduces a benchmark to assess this vulnerability in resume screening, revealing attack success rates exceeding 80% for certain attack types.
  arXiv  Detail & Related papers  (2025-12-23T08:42:09Z)
• AegisLLM: Scaling Agentic Systems for Self-Reflective Defense in LLM Security [74.22452069013289]
  AegisLLM is a cooperative multi-agent defense against adversarial attacks and information leakage.<n>We show that scaling agentic reasoning system at test-time substantially enhances robustness without compromising model utility.<n> Comprehensive evaluations across key threat scenarios, including unlearning and jailbreaking, demonstrate the effectiveness of AegisLLM.
  arXiv  Detail & Related papers  (2025-04-29T17:36:05Z)
• SafeAgentBench: A Benchmark for Safe Task Planning of Embodied LLM Agents [58.65256663334316]
  We present SafeAgentBench -- the first benchmark for safety-aware task planning of embodied LLM agents in interactive simulation environments.<n>SafeAgentBench includes: (1) an executable, diverse, and high-quality dataset of 750 tasks, rigorously curated to cover 10 potential hazards and 3 task types; (2) SafeAgentEnv, a universal embodied environment with a low-level controller, supporting multi-agent execution with 17 high-level actions for 9 state-of-the-art baselines; and (3) reliable evaluation methods from both execution and semantic perspectives.
  arXiv  Detail & Related papers  (2024-12-17T18:55:58Z)

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.