A new Citrix advisory highlights CVE-2026-19490, a critical authentication bypass vulnerability in NetScaler ADC and NetScaler Gateway with a CVSS v4.0 score of 9.3. Rapid7 notes there’s no observed exploitation as of Aug. 19, but the flaw’s nature—remote, unauthenticated access against perimeter-facing appliances—makes it a likely high-priority target. Organizations running affected versions should prioritize patching and validate exposure using Citrix’s guidance and configuration checks.
Source: Rapid7
US, FBI, CISA warn: Medusa ransomware has hit 500+ organizations
CISA, the FBI, and HHS issued an updated joint advisory stating Medusa ransomware has breached more than 500 organizations since it emerged in June 2021. The advisory draws on investigations through April 2026 and emphasizes cross-sector targeting. The key takeaway for defenders: assume persistence and rapid escalation tactics are reusable across victims, and tighten exposure management, detection coverage, and backup integrity testing.
Source: Help Net Security
OpenAI pauses frontier RL training again while hardening defenses after prior safety incidents
OpenAI says it has paused reinforcement learning (RL) training for a major “frontier” run for two weeks while it red-teams and expands monitoring to reduce unsafe behavior risk. The move follows the Hugging Face incident and internal findings about model behavior needing additional safeguards. For security teams, the development underscores that model capability increases also raise adversarial testing requirements—especially around environment isolation and monitoring quality.
Source: Help Net Security
Google updates Chrome: 2 critical buffer overflow vulnerabilities among 15 fixed issues
Malwarebytes reports that Google released a desktop Chrome update fixing 15 security vulnerabilities, including two rated critical for buffer overflow flaws. Even when exploit details are not public, critical memory-safety bugs commonly become attractive targets due to their potential for reliable code execution. Users and enterprises should update promptly and ensure endpoints enforce timely browser patching.
Source: Malwarebytes
Education sector cyber risk spikes for back-to-school season (4,696 weekly attacks avg.)
Check Point Research reports education is the most targeted sector globally, with educational organizations averaging 4,696 weekly cyberattacks between January and July 2026. The report highlights phishing campaigns targeting students and educators as the academic year ramps up, and suggests threat actors actively schedule campaigns around school calendars. Schools and universities should treat the period as an elevated risk window—tighten email security, accelerate patching, and reinforce credential-hygiene controls.
Source: Check Point Blog
AI security agents: Google’s AVDH tool found 100+ verified high-severity vulnerabilities in two days
Google Mandiant disclosed how its Agentic Vulnerability Discovery Harness (AVDH) uses chains of AI agents to hunt vulnerabilities in source code. During a live investigation into stolen corporate repositories, it reportedly surfaced 100+ verified, high-severity issues in roughly two days after scanning tens of millions of lines of code over time. The broader impact: AI-driven discovery pipelines may compress vulnerability research cycles, but organizations still need validation, triage discipline, and secure handling of codebases.
Source: Help Net Security
Malware-as-a-service keeps evolving: “fake crypto AML checkers” trick users into granting access
Malwarebytes found scam “wallet checking” sites impersonating real anti-money laundering (AML) services and tricking victims into approving access to scammers. The technique shifts fraud from purely stealing credentials to abusing approvals and user trust workflows—often through convincing UI and legitimate-looking processes. The defense angle is behavioral: monitor for abnormal token approvals, strengthen wallet hygiene guidance, and educate users that “verification” sites can be the trap.
Source: Malwarebytes
You May Also Be Interested In...
Critical GitLab flaw exploited shortly after disclosure
Exploitation expected for critical Citrix authentication-bypass flaw