arXiv:2610.06966v1 Announce Type: new Abstract: Indirect prompt injection (IPI) hides adversarial instructions in content that…
1 min Open reportarXiv:2610.06994v1 Announce Type: new Abstract: Backdoor-defense leaderboards print a clean-accuracy drop and read it as removal…
1 min Open reportarXiv:2610.07009v1 Announce Type: new Abstract: Image-to-text jailbreaks place harmful intent in text, image content, or the…
1 min Open reportarXiv:2610.07089v1 Announce Type: new Abstract: LLM agents increasingly act in multi-actor environments, exposing them to misuse…
1 min Open reportarXiv:2610.07125v1 Announce Type: new Abstract: While open-weight models have enjoyed steady progress in capabilities and wide…
1 min Open reportarXiv:2610.07256v1 Announce Type: new Abstract: AI agents powered by large language models (LLMs) can perform complex tasks but…
1 min Open reportarXiv:2610.07258v1 Announce Type: new Abstract: Enterprise AI agents that share a memory store face two unaddressed risks…
1 min Open reportarXiv:2610.07282v1 Announce Type: new Abstract: Protecting critical infrastructure increasingly depends on continuously verifying…
1 min Open reportarXiv:2610.07298v1 Announce Type: new Abstract: Cyber threat analysis increasingly depends on evidence distributed across vendor…
1 min Open reportarXiv:2610.07310v1 Announce Type: new Abstract: Security operations centres and national incident-response teams defending…
1 min Open reportarXiv:2610.07345v1 Announce Type: new Abstract: IAM policy analysis typically emphasizes the authorization capabilities encoded…
1 min Open reportarXiv:2610.07351v1 Announce Type: new Abstract: Extremely Lossy Functions (ELFs) are a standard model primitive that captures…
1 min Open reportarXiv:2610.07386v1 Announce Type: new Abstract: Network traffic analysis is central to network security, spanning tasks from…
1 min Open reportarXiv:2610.07489v1 Announce Type: new Abstract: Increasing connectivity to the outside world and the lack of inbuilt security…
1 min Open reportarXiv:2610.07510v1 Announce Type: new Abstract: Developers can build LLM agents by adapting third-party models through benign…
1 min Open reportarXiv:2610.07531v1 Announce Type: new Abstract: A person who answers an unexpected call claiming to come from a bank has no way…
1 min Open reportarXiv:2610.07532v1 Announce Type: new Abstract: LLMs have advanced rapidly, raising growing concerns about their safety. Recent…
1 min Open reportarXiv:2610.07635v1 Announce Type: new Abstract: Compiler-introduced security bugs (CISBs) arise when an optimization, lowering…
1 min Open reportarXiv:2610.07639v1 Announce Type: new Abstract: Coding agent harnesses mediate tool use and authorize actions, yet their security…
1 min Open reportarXiv:2610.07645v1 Announce Type: new Abstract: Self-improving LLM agents increasingly distill successful experiences into…
1 min Open report