ThreatsDay Bulletin: Hybrid P2P Botnet, 13-Year-Old Apache RCE and 18 More Stories

Thursday. Another week, another batch of things that probably should've been caught sooner but weren't. This one's got some range — old vulnerabilities getting new life, a few "why was that even possible" moments, attackers leaning on platforms and tools you'd normally trust without thinking twic...

Follow My Eyes: Backdoor Attacks on VLM-Based Scanpath Prediction

Scanpath prediction models forecast the sequence and timing of human fixations during visual search, driving foveated rendering and attention-based interaction in mobile systems where their integrity is a first-class security concern. We present the first study of backdoor attacks against VLM-bas...

PT-2026-31758

OpenClaw before 2026.3.22 contains an improper authentication verification vulnerability in Google Chat app-url webhook handling that accepts add-on principals outside intended deployment bindings. Attackers can bypass webhook authentication by providing non-deployment add-on principals to execut...

PraisonAI 信息泄露漏洞

PraisonAI is a low-code multi-agent collaboration framework developed by Mervin Praison. Versions of PraisonAI prior to 4.5.128 contained an information leakage vulnerability. This vulnerability stemmed from the AgentOS deployment platform not implementing authentication, and the default CORS...

The software for deploying and executing AI models by NVIDIA Triton Inference Server has vulnerabilities related to integer overflow. This allows attackers to trigger a service failure.

The vulnerability of NVIDIA Triton Inference Server’s deployment and AI model execution software is related to a numerical overflow issue. Exploiting this vulnerability can allow an attacker to cause service failures...

The software for deploying and executing AI models in NVIDIA Triton Inference Server is vulnerable due to incorrect conversion of numerical value types. This vulnerability allows attackers to trigger service failures.

The vulnerability of NVIDIA Triton Inference Server’s deployment and AI model execution software is related to incorrect conversion of numerical value types. Exploiting this vulnerability can allow a remote attacker to cause service failures...

New Chaos Variant Targets Misconfigured Cloud Deployments, Adds SOCKS Proxy

Cybersecurity researchers have flagged a new variant of malware called Chaos that's capable of hitting misconfigured cloud deployments, marking an expansion of the botnet's targeting infrastructure. "Chaos malware is increasingly targeting misconfigured cloud deployments, expanding beyond its...

PT-2026-31345

Name of the Vulnerable Software and Affected Versions Kibana affected versions not specified Description An authenticated user with access to the automatic import feature can submit specially crafted requests with excessively large input values. When multiple such requests are sent concurrently,...

CVE-2026-35043

BentoML is a Python library for building online serving systems optimized for AI apps and model inference. Prior to 1.4.38, the cloud deployment path in src/bentoml/internal/cloud/deployment.py was not included in the fix for CVE-2026-33744. Line 1648 interpolates systempackages directly into a...

EUVD-2026-19869

OpenObserve is a cloud-native observability platform. In 0.70.3 and earlier, the validateenrichmenturl function in src/handler/http/request/enrichmenttable/mod.rs fails to block IPv6 addresses because Rust's url crate returns them with surrounding brackets e.g. "::1" not "::1". An authenticated...

Information Exposure

Overview openclaw is a 🦞 OpenClaw — Personal AI Assistant Affected versions of this package are vulnerable to Information Exposure via the connect process. An attacker can obtain sensitive host filesystem paths and deployment metadata by making authenticated requests as a non-admin client...

CVE-2026-28810 Predictable DNS Transaction IDs Enable Cache Poisoning in Built-in Resolver

Generation of Predictable Numbers or Identifiers vulnerability in Erlang/OTP kernel inetres, inetdb modules allows DNS Cache Poisoning. The built-in DNS resolver inetres uses a sequential, process-global 16-bit transaction ID for UDP queries and does not implement source port randomization...

LanG -- a Governance-Aware Agentic AI Platform for Unified Security Operations

Modern Security Operations Centers struggle with alert fatigue, fragmented tooling, and limited cross-source event correlation. Challenges that current Security Information Event Management and Extended Detection and Response systems only partially address through fragmented tools. This paper...

Can Drift-Adaptive Malware Detectors Be Made Robust? Attacks and Defenses under White-Box and Black-Box Threats

Concept drift and adversarial evasion are two major challenges for deploying machine learning-based malware detectors. While both have been studied separately, their combination, the adversarial robustness of drift-adaptive detectors, remains unexplored. We address this problem with AdvDA, a rece...

PYSEC-2026-158

BentoML is a Python library for building online serving systems optimized for AI apps and model inference. Prior to 1.4.38, the cloud deployment path in src/bentoml/internal/cloud/deployment.py was not included in the fix for CVE-2026-33744. Line 1648 interpolates systempackages directly into a...

PYSEC-2026-158

BentoML is a Python library for building online serving systems optimized for AI apps and model inference. Prior to 1.4.38, the cloud deployment path in src/bentoml/internal/cloud/deployment.py was not included in the fix for CVE-2026-33744. Line 1648 interpolates systempackages directly into a...

CVE-2026-35043

BentoML is a Python library for building online serving systems optimized for AI apps and model inference. Prior to 1.4.38, the cloud deployment path in src/bentoml/internal/cloud/deployment.py was not included in the fix for CVE-2026-33744. Line 1648 interpolates systempackages directly into a...

CVE-2026-35043 BentoML: command injection in cloud deployment setup script (deployment.py)

BentoML is a Python library for building online serving systems optimized for AI apps and model inference. Prior to 1.4.38, the cloud deployment path in src/bentoml/internal/cloud/deployment.py was not included in the fix for CVE-2026-33744. Line 1648 interpolates systempackages directly into a...

CVE-2026-35043 BentoML: command injection in cloud deployment setup script (deployment.py)

BentoML is a Python library for building online serving systems optimized for AI apps and model inference. Prior to 1.4.38, the cloud deployment path in src/bentoml/internal/cloud/deployment.py was not included in the fix for CVE-2026-33744. Line 1648 interpolates systempackages directly into a...

CVE-2026-35043

CVE-2026-35043 affects BentoML prior to 1.4.38. The cloud deployment path in bentoml/_internal/cloud/deployment.py interpolates system_packages directly into a shell command in the generated setup.sh, enabling remote code execution on the CI/CD cloud build infrastructure during deployment. The is...