Papers for

it security teams

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

AI security incident reports need new details and protections

Beyond Predictable Paths: Redefining AI Security Incident Reporting for Agents

Abstract: AI agents are being deployed rapidly, accompanied by a growing number of AI-specific attacks and corresponding incidents. As incident reporting becomes increasingly important for legal compliance, governance, accountability, and security; current frameworks must be adapted to the unique characteristics of AI agents. In this paper, two editorial authors compare AI systems and AI agents and, drawing on input from 23 experts in academia and industry, identify the information required for reporting incidents where the security of AI agents is harmed. %involving AI agents. Potential reporting elements include, for example, agent memory and memory accesses, actual and potential levels of autonomy, and tool usage. Based on these findings, we identify several open research questions, including how to efficiently record incidents and how to determine whether vulnerabilities and incidents generalize. Expert feedback also highlighted potential reporting weaknesses, such as risks of data leakage and attacks targeting the reporting infrastructure itself, creating additional research needs. Lastly, we summarize privacy requirements and outline research directions for the secure and trustworthy deployment of AI agents.

Mon 21 SeptCryptography and SecurityArtificial Intelligence
The gist
AI agents are computer programs that can act on their own, but they can be attacked in ways that are different from usual software. The paper looks at how to report security problems with these AI agents by talking to experts. It suggests including new types of information in reports, like what the AI remembers or what tools it uses. The authors also point out challenges such as keeping report data safe from leaks or attacks.
Open 2609.24515v1

On-premise gpu pipeline speeds root cause analysis in kubernetes clusters

TriFleetRCA: On-Premise LLM Root Cause Analysis for Kubernetes

Abstract: Root cause analysis at a remote site is slow: evidence is scattered across pod logs, Kubernetes events and cluster-level objects, and many operators cannot send production logs to a hosted model at all. On-premise inference removes the second constraint but raises a question live-cluster benchmarks have not addressed: when one workstation GPU fixes both the model and the context budget, how should evidence be retrieved, and what happens when the runbooks the model consults have been tampered with? We present TriFleetRCA, a pipeline running entirely on one on-premise GPU that collects evidence at one of three scopes (pod, namespace, cluster), ranks it by template de-duplication then BM25, filters runbooks through an ingest guard, and returns a root cause with the evidence lines supporting it. We evaluate on a live Kubernetes cluster into which we inject four faults, so ground truth is known by construction, across 100 analyses with Qwen2.5-14B-Instruct at temperature 0. The hit rate was 0.85, 0.90 and 0.95 at pod, namespace and cluster scope; intervals overlap, but the whole scope effect comes from the one fault whose cause is a cluster-level object, and cluster scope costs 55% more tokens. De-duplication before ranking raised the hit rate from 0.75 to 0.90 at equal token cost. A poisoned runbook telling the model to delete the namespace was rejected by the guard every run; with the guard disabled the model declined to follow it in all 20 analyses, making the guard defence in depth rather than the sole barrier. Separating citation quality from accuracy proved informative: one fault was diagnosed correctly and cited incorrectly every trial, a failure mode accuracy conceals. Median latency was 1.6 s at 2,200 prompt tokens. We release the pipeline, the fault injector and all records.

Sun 20 SeptCryptography and SecurityArtificial IntelligenceDistributed, Parallel, and Cluster Computing
The gist
Finding the cause of problems in Kubernetes clusters can be slow because the evidence is scattered and some operators cannot send logs to cloud models. The authors built TriFleetRCA, a system that runs on one local GPU and collects, ranks, and filters evidence from various cluster levels to identify root causes quickly and safely. They tested it by injecting faults and found it can accurately pinpoint issues with good speed and guardrails against dangerous instructions. This approach helps teams troubleshoot Kubernetes issues without needing to send sensitive data externally.
Open 2609.23766v1