AgentRedBench: Dynamic Redteaming and Integration-Aware Defense for LLM Agents over SaaS Integrations

• URL: http://arxiv.org/abs/2606.02240v2
• Date: Tue, 02 Jun 2026 15:23:51 GMT
• Title: AgentRedBench: Dynamic Redteaming and Integration-Aware Defense for LLM Agents over SaaS Integrations
• Abstract summary: Indirect prompt injection in tool-use agents is a concrete production threat.<n>AGENTREDBENCH is a benchmark of 215 subtle underspecified authorization scenarios.<n>AGENTREDGUARD is a guard trained on an integration-diverse corpus of adversarial tool-response content.
• Score: 0.0954904463032233
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Indirect prompt injection in tool-use agents is a concrete production threat: LLM agents read from integrations (third-party services such as Gmail, Salesforce, or Jira accessed through tool calls) whose response content the user neither writes nor controls. Existing benchmarks under-measure the threat: most cover only a handful of integrations with the same attack payload replayed across runs, and open-source guards are trained on chat-style data rather than tool-response content. We introduce AGENTREDBENCH, a dynamic LLM-driven redteaming benchmark of 215 subtle underspecified authorization (attacks at the boundary of what the user's request authorises) scenarios across 24 enterprise integrations in nine functional families and five attack types. Across an eight-model panel (Anthropic, OpenAI, Google), no-guard ASR (attack success rate) ranges from 32% (Claude Sonnet 4.6) to 81% (Gemini 3 Flash). To keep the scenario set out of training corpora and preserve headline ASR meaning over time, we release the codebase, integration schemas, and AGENTREDGUARD model openly; the canonical scenarios are evaluated through a maintainer-mediated channel with immutable versioning. We release AGENTREDGUARD alongside the benchmark: a guard trained on an integration-diverse corpus of adversarial tool-response content. AGENTREDGUARD cuts panel ASR from 69.9% to 2.4% at 0.37% false-positive rate, outperforming every open-source baseline with non-trivial detection (Llama Guard, PromptGuard 2, ProtectAI) on both axes. Cross-integration and cross-attack type holdouts both confirm the gain transfers beyond the training subset.

Related papers

• SIR: Self-improving Red-teaming for Compute Use Agents [71.71987044117371]
  Computer use agents (CUAs) are vision-language models that perceive a screen and act on a real operating system through mouse, keyboard, and terminal.<n>Because they can be exposed to untrusted content while operating, they are vulnerable to indirect prompt injection (IPI)<n>We present SIR, a black box IPI attack that composes stealthy injections from a small library of reusable principles stated in plain language.
  arXiv  Detail & Related papers  (2026-08-31T03:39:43Z)
• LMSM: LLM Security Framework Inspired by Linux Security Modules [56.93644694627783]
  Large language models (LLMs) are increasingly deployed with layered defenses, yet malicious prompts can still bypass them.<n>We present Language Model Security Modules (LMSM), a security framework that adapts the separation behind Linux Security Modules (LSM) to LLM serving.
  arXiv  Detail & Related papers  (2026-08-26T12:13:05Z)
• Autonomous LLM Agent Worms: Cross-Platform Propagation, Automated Discovery and Temporal Re-Entry Defense [10.105892979517622]
  We present the first systematic framework for automated analysis of persistent worm propagation in file-backed multi-agent LLM ecosystems.<n>We demonstrate zero-click autonomous propagation, 3-hop cross-platform transmission without platform-specific adaptation, inter-agent privilege and data exfiltration.
  arXiv  Detail & Related papers  (2026-05-04T16:49:29Z)
• Semantic Intent Fragmentation: A Single-Shot Compositional Attack on Multi-Agent AI Pipelines [0.9772267314090434]
  We introduce Semantic Intent Fragmentation (SIF) as an attack class against LLM orchestration systems.<n>SIF exploits four mechanisms: bulk scope, silent data exfiltration, embedded trigger deployment, and quasi-identifier aggregation.<n>A GPT-20B orchestrator produces policy-violating plans in 71% of cases while every subtask appears benign.
  arXiv  Detail & Related papers  (2026-04-08T18:19:03Z)
• ClawsBench: Evaluating Capability and Safety of LLM Productivity Agents in Simulated Workspaces [17.202580606345666]
  Large language model (LLM) agents are increasingly deployed to automate productivity tasks.<n>Existing benchmarks rely on simplified environments and fail to capture realistic, multi-service environments.<n>We introduce ClawsBench, a benchmark for evaluating and improving LLM agents in realistic productivity settings.
  arXiv  Detail & Related papers  (2026-04-06T21:09:06Z)
• 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)
• BackdoorAgent: A Unified Framework for Backdoor Attacks on LLM-based Agents [58.83028403414688]
  Large language model (LLM) agents execute tasks through multi-step workflow that combine planning, memory, and tool use.<n>Backdoor triggers injected into specific stages of an agent workflow can persist through multiple intermediate states and adversely influence downstream outputs.<n>We propose textbfBackdoorAgent, a modular and stage-aware framework that provides a unified agent-centric view of backdoor threats in LLM agents.
  arXiv  Detail & Related papers  (2026-01-08T03:49:39Z)
• Breaking the Code: Security Assessment of AI Code Agents Through Systematic Jailbreaking Attacks [11.371490212283383]
  Code-capable large language model (LLM) agents are embedded into software engineering where they can read, write, and execute code.<n>We present JAWS-BENCH, a benchmark spanning three escalating workspaces that mirror attacker capability.<n>We find that under prompt-only conditions in JAWS-0, code agents accept 61% of attacks on average; 58% are harmful, 52% parse, and 27% run end-to-end.
  arXiv  Detail & Related papers  (2025-10-01T18:38:20Z)
• A Multi-Agent LLM Defense Pipeline Against Prompt Injection Attacks [1.1435139523855764]
  This paper presents a novel multi-agent defense framework to detect and neutralize prompt injection attacks in real-time.<n>We evaluate our approach using two distinct architectures: a sequential chain-of-agents pipeline and a hierarchical coordinator-based system.
  arXiv  Detail & Related papers  (2025-09-16T19:11:28Z)
• AgentVigil: Generic Black-Box Red-teaming for Indirect Prompt Injection against LLM Agents [54.29555239363013]
  We propose a generic black-box fuzzing framework, AgentVigil, to automatically discover and exploit indirect prompt injection vulnerabilities.<n>We evaluate AgentVigil on two public benchmarks, AgentDojo and VWA-adv, where it achieves 71% and 70% success rates against agents based on o3-mini and GPT-4o.<n>We apply our attacks in real-world environments, successfully misleading agents to navigate to arbitrary URLs, including malicious sites.
  arXiv  Detail & Related papers  (2025-05-09T07:40:17Z)
• 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)

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.