USENIX Security 26 - Context Contamination in LLM Analysis of Network Security Logs: Poison with P… @UsenixOrg
USENIX Security 26 - Context Contamination in LLM Analysis of Network Security Logs: Poison with P…  @UsenixOrg
Uploaded October 2026 | Updated November 2026, 5 days ago
USENIX Security '26 - Context Contamination in LLM Analysis of Network Security Logs: Poison with Passive Prompt Injection and Mitigation Evaluation

Rabimba Karanjai, PayPal / University of Houston; Yang Lu, University of Houston; Hemanth Hegadehalli Madhavarao, PayPal; Lei Xu, Kent State University; Weidong Shi, University of Houston

Large Language Models are increasingly deployed in Security Operations Centers for log analysis tasks including summarization, alert triage, and threat investigation. These systems ingest logs from external-facing services and process network logs as natural language contexts to generate security insights. We demonstrate that this architectural pattern introduces a critical vulnerability: adversaries can embed prompt injection payloads in log-generating fields that persist in storage and are executed when analysts query the LLM, achieving what we term passive prompt injection. We present LogInject, a systematic framework for evaluating these threats. Using LogInject-1.0, a benchmark of 12,847 log entries including 2,569 adversarial samples, we evaluate three production LLMs across four attack objectives: activity concealment, false positive generation, information exfiltration, and output hijacking. Our findings reveal an up to 88.2% attack success rate (83.4% average across models) under the baseline conditions. We introduce Context Stitching, a novel technique that fragments payloads across multiple log entries to evade stateless filters while exploiting LLM long-context reasoning, achieving a 76.4% success rate. As mitigation, we evaluate layered defenses by combining input filtering, prompt hardening, and output validation, demonstrating a 90.4% attack reduction, although 8.4% residual vulnerability persists. Our results establish that LLM-based log analysis creates an inherent confused deputy vulnerability where untrusted data and trusted instructions compete indistinguishably for model attention, requiring defense in-depth architectures and continued human oversight for security-critical decisions.

View the full USENIX Security '26 program at usenix.org/conference/usenixsecurity26/technical-sessions
USENIX Security 26 - Context Contamination in LLM Analysis of Network Security Logs: Poison with P…WOOT 26 - LIMA: Defining, Benchmarking and Detecting Cross-Layer Vulnerabilities in LLM...NSDI 26 - BURST: Seeking High-performance, Interoperability and Scalability in Soft-RDMAVehicleSec 26 - RVDebloater: Mode-based Adaptive Firmware Debloating for Robotic VehiclesSREcon26 Europe/Middle East/Africa - How We Troubleshoot Difficult Problems: Cognition and...USENIX Security 26 - Transparent Dictionaries from Polynomial CommitmentsSREcon26 Europe/Middle East/Africa - The Optimization We Didnt BuildSREcon26 Europe/Middle East/Africa - Pull Requests Are Dead, Long Live Peer ReviewNSDI 25 - Enabling Portable and High-Performance SmartNIC Programs with AlkaliUSENIX Security 26 (Enigma Track) - Youth Digital Immersion: Cyber Readiness Builds Mental...USENIX Security 26 - Five Queries Are Enough: Query-Efficient and Surrogate-Free Membership Infere…USENIX Security 26 - MATE: Policy-Aware Security Auditing for Mobile Agents via Synthesis-Driven T…
USENIX |

USENIX Security '26 - Context Contamination in LLM Analysis of Network Security Logs: Poison with P…

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER