3+ day, 14+ hour ago   (419+ words) Anthropic has disclosed four cybersecurity-evaluation incidents in which Claude models reached and interacted with third-party systems after a configuration failure exposed test environments to the public internet. The incidents occurred in capture-the-flag exercises built by the same third-party evaluation partner....

4+ day, 12+ hour ago   (421+ words) A newly disclosed, now-remediated ChatGPT isolation flaw could have enabled attackers to access Gmail data, chat history, uploaded files, and other connected-app content from a victim’s session, then secretly relay it to a separate attacker-controlled account. The issue effectively created…...

6+ day, 16+ hour ago   (403+ words) CrowdStrike has launched SafeMind, a family of purpose-built cybersecurity models and agentic AI harnesses designed to help defenders identify, validate, and remediate threats inside the Falcon platform. Announced during Fal.Con 2026 in Las Vegas, SafeMind represents CrowdStrike’s move away from…...

1+ week, 4+ day ago   (387+ words) The acquisition, announced September 1, 2026, will deepen agentic AI capabilities within Palo Alto Networks’ Cortex security operations platform. The company said Console’s technology is intended to help security teams analyze alerts, resolve operational issues, and execute requests at machine speed through…...

1+ week, 4+ day ago   (365+ words) A newly disclosed vulnerability in the Hugging Face Transformers library can cause attacker-controlled Python code to be written to a local system before a user decides whether to trust and execute remote code. Tracked as CVE-2026-80047 and documented in CERT…...

1+ week, 5+ day ago   (287+ words) Group-IB has uncovered BraZetsu, a Python-based Windows malware framework designed to help cybercriminals gain, evaluate, and sell access to compromised corporate systems. The malware is linked with high confidence to the Brazilian threat actor Exilware and functions as the technical…...

1+ week, 5+ day ago   (292+ words) Threat actors are actively exploiting two critical remote code execution vulnerabilities affecting Langflow and Ruby on Rails. The flaw lies in the code-validation path, where insufficient validation of a user-supplied parameter before it is used to execute Python can allow…...

2+ week, 4+ day ago   (482+ words) A critical vulnerability in NVIDIA NemoClaw, tracked as CVE-2026-65105, could let attackers hijack AI agents after a victim visits a malicious website. The issue combines an insecure Ollama network configuration with DNS rebinding, allowing unauthenticated access to the local model…...

2+ week, 4+ day ago   (482+ words) A critical vulnerability in NVIDIA NemoClaw, tracked as CVE-2026-65105, could let attackers hijack AI agents after a victim visits a malicious website. The issue combines an insecure Ollama network configuration with DNS rebinding, allowing unauthenticated access to the local model…...

3+ week, 1+ day ago   (557+ words) A newly disclosed prompt-injection technique can reportedly turn a routine “summarize this page” request in xAI’s Grok web chat into a silent data-exfiltration attack, exposing a user’s name, approximate location, subscription tier, and active conversation history. Security researchers at Adversa…...