Stop Fixating on Prompts: Reasoning Hijacking and Constraint Tightening for Red-Teaming LLM Agents

With the widespread application of LLM-based agents across various domains, their complexity has introduced new security threats. Existing red-team methods mostly rely on modifying user prompts, which lack adaptability to new data and may impact the agent's performance. To address the challenge,...

Hackers or Hallucinators? A Comprehensive Analysis of LLM-Based Automated Penetration Testing

The rapid advancement of Large Language Models LLMs has created new opportunities for Automated Penetration Testing AutoPT, spawning numerous frameworks aimed at achieving end-to-end autonomous attacks. However, despite the proliferation of related studies, existing research generally lacks...

PYSEC-2026-2298

vLLM is an inference and serving engine for large language models LLMs. From 0.1.0 to before 0.19.0, a Denial of Service vulnerability exists in the vLLM OpenAI-compatible API server. Due to the lack of an upper bound validation on the n parameter in the ChatCompletionRequest and CompletionReques...

PYSEC-2026-2298

vLLM is an inference and serving engine for large language models LLMs. From 0.1.0 to before 0.19.0, a Denial of Service vulnerability exists in the vLLM OpenAI-compatible API server. Due to the lack of an upper bound validation on the n parameter in the ChatCompletionRequest and CompletionReques...

EUVD-2026-19351

vLLM is an inference and serving engine for large language models LLMs. From 0.1.0 to before 0.19.0, a Denial of Service vulnerability exists in the vLLM OpenAI-compatible API server. Due to the lack of an upper bound validation on the n parameter in the ChatCompletionRequest and CompletionReques...

CoopGuard: Stateful Cooperative Agents Safeguarding LLMs against Evolving Multi-Round Attacks

As Large Language Models LLMs are increasingly deployed in complex applications, their vulnerability to adversarial attacks raises urgent safety concerns, especially those evolving over multi-round interactions. Existing defenses are largely reactive and struggle to adapt as adversaries refine...

LLM-Enabled Open-Source Systems in the Wild: An Empirical Study of Vulnerabilities in GitHub Security Advisories

Large language models LLMs are increasingly embedded in open-source software OSS ecosystems, creating complex interactions among natural language prompts, probabilistic model outputs, and execution-capable components. However, it remains unclear whether traditional vulnerability disclosure...

Combating Data Laundering in LLM Training

Data rights owners can detect unauthorized data use in large language model LLM training by querying with proprietary samples. Often, superior performance e.g., higher confidence or lower loss on a sample relative to the untrained data implies it was part of the training corpus, as LLMs tend to...

PT-2026-29877

Name of the Vulnerable Software and Affected Versions vLLM versions 0.5.5 through 0.17.999 Description vLLM, an inference and serving engine for large language models LLMs, exhibits an inconsistency in audio processing. Versions 0.5.5 through 0.17.999 utilize numpy.mean for mono downmixing via...

CVE-2026-27893

CVE-2026-27893 affects vLLM’s inference/serving engine. From version 0.10.1 up to (but not including) 0.18.0, two model implementation files hardcode trust_remote_code=True when loading sub-components, bypassing the user’s --trust-remote-code=False security opt-out. This enables remote code execu...

Towards Leveraging LLMs to Generate Abstract Penetration Test Cases from Software Architecture

Software architecture models capture early design decisions that strongly influence system quality attributes, including security. However, architecture-level security assessment and feedback are often absent in practice, allowing security weaknesses to propagate into later phases of the software...

CVE-2026-32114

Discourse (open‑source discussion platform) contains an Insecure Direct Object Reference (IDOR) vulnerability. Prior to versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2, any authenticated user can access metadata about AI personas, features, and LLM models by supplying their identifiers. This m...

CVE-2026-32114 Discourse's unscoped status lookups leak restricted metadata

Discourse is an open-source discussion platform. Prior to versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2, there is an Insecure Direct Object Reference IDOR vulnerability that allows any authenticated user to access metadata about AI personas, features, and LLM models by providing their...

The vulnerability of the PyNcclPipe class in the library for working with Large Language Models (LLMs) like vLLM allows a hacker to execute arbitrary code.

The vulnerability of the PyNcclPipe class in the library for working with Large Language Models LLMs like vLLM is related to deficiencies in the deserialization mechanism. Exploiting this vulnerability allows a remote attacker to execute arbitrary code...

Measuring and Exploiting Confirmation Bias in LLM-Assisted Security Code Review

Security code reviews increasingly rely on systems integrating Large Language Models LLMs, ranging from interactive assistants to autonomous agents in CI/CD pipelines. We study whether confirmation bias i.e., the tendency to favor interpretations that align with prior expectations affects LLM-bas...

Security Assessment and Mitigation Strategies for Large Language Models: A Comprehensive Defensive Framework

Large Language Models increasingly power critical infrastructure from healthcare to finance, yet their vulnerability to adversarial manipulation threatens system integrity and user safety. Despite growing deployment, no comprehensive comparative security assessment exists across major LLM...

The vulnerability of the pythonic_tool_parser.py script from the library for working with large language models (LLMs), vLLM, allows a perpetrator to trigger a service failure.

The vulnerability of the pythonictoolparser.py script, a library for working with large language models LLMs, is related to the use of a regular expression with inefficient computational complexity. Exploiting this vulnerability could allow a malicious actor to cause service failures...

PISmith: Reinforcement Learning-Based Red Teaming for Prompt Injection Defenses

Prompt injection poses serious security risks to real-world LLM applications, particularly autonomous agents. Although many defenses have been proposed, their robustness against adaptive attacks remains insufficiently evaluated, potentially creating a false sense of security. In this work, we...

FalconEYE 2.1.0

FalconEYE represents a paradigm shift in static code analysis. Instead of relying on predefined vulnerability patterns, it leverages large language models to reason about your code the same way a security expert would, understanding context, intent, and subtle security implications that tradition...

Why LLMs Fail: A Failure Analysis and Partial Success Measurement for Automated Security Patch Generation

Large Language Models LLMs show promise for Automated Program Repair APR, yet their effectiveness on security vulnerabilities remains poorly characterized. This study analyzes 319 LLM-generated security patchesacross 64 Java vulnerabilities from the Vul4J benchmark. Using tri-axis evaluation...