Huntress is tracking a large automated password-spray campaign against Microsoft 365 that has made more than 81 million login attempts through the Azure CLI in two weeks and broken into 78 accounts across 64 organizations. The attackers replay old username and password pairs from breach data against an authentication flow that sends credentials straight to the token endpoint without triggering interactive multi-factor authentication, so weak or reused passwords give them direct access. Several victims had MFA, but it was scoped only to admins, only to certain apps, or only to untrusted locations, and so did not cover this path. The traffic comes from infrastructure whose address ranges trace back to China.
A critical flaw in Progress Kemp LoadMaster lets an unauthenticated attacker run commands as root on the appliance by sending a crafted request to its API. Rated 9.8, the bug (CVE-2026-8037) sits in a function meant to sanitize input before it reaches a shell command, and LoadMaster's position as an edge load balancer and application delivery controller makes a pre-authentication flaw especially dangerous, since it can turn a protective choke point into a direct foothold. Progress patched it in early June, and researchers at watchTowr published a full technical write-up with a working proof-of-concept on June 29. No exploitation has been reported yet, but Progress also makes MOVEit, a past mass-exploitation target.
CISA has updated its Known Exploited Vulnerabilities catalog to warn that ransomware gangs are now exploiting BlueHammer, a Microsoft Defender privilege-escalation flaw. The bug (CVE-2026-33825) lets a local attacker who already has a foothold escalate to SYSTEM by abusing Defender's file-remediation logic, giving them access to password hashes and the control needed to disable defenses and prepare systems for encryption. It was leaked with proof-of-concept code by a researcher in early April as a protest over Microsoft's disclosure process, exploited as a zero-day, then patched on April 14. It cannot be used for remote compromise on its own, but it strengthens attackers after initial access.
Citrix has released fixes for six vulnerabilities in NetScaler ADC and NetScaler Gateway, including a high-severity memory-disclosure flaw that researchers place in the same class as the 2023 CitrixBleed bug. That flaw (CVE-2026-8451, rated 8.8) leaks small amounts of memory through malformed SAML requests and shares a root cause with an earlier NetScaler bug that was exploited within days of disclosure. The bulletin also covers an unauthenticated arbitrary file read and several denial-of-service issues, with CVSS scores from 6.9 to 8.8. No exploitation has been reported yet, but NetScaler appliances have drawn more than 20 entries on CISA's exploited-vulnerabilities list in three years, several used in ransomware.
Aflac Life Insurance Japan, a subsidiary of the US insurance giant Aflac, says attackers broke into its policyholder portal and stole personal data belonging to about 4.38 million customers and agents. The intruders accessed systems repeatedly between June 15 and June 25, when the breach was detected through a surge in traffic, and the company suspended affected systems in response. Exposed data includes names, addresses, phone numbers, dates of birth, gender, and insurance account details, plus premium payment account information for roughly 230,000 people; no credit card data was taken. Aflac says the incident is limited to its Japan systems and does not affect its US operations.
Microsoft is warning that attackers can hijack AI agents through poisoned tool descriptions, the plain-text notes that tell an agent what a tool does. Because agents connect to systems through the Model Context Protocol and read these descriptions to decide how to act, an attacker who updates a trusted third-party tool can bury a hidden instruction in its description, telling the agent to quietly collect and exfiltrate data on its next task. Many setups pick up description changes without re-approval, so the poisoned version goes live silently. Each step the agent takes looks legitimate and runs with the user's own permissions, so no alarm fires.
Researchers at LayerX detailed BioShocking, an attack that manipulates AI browser agents into ignoring their safety rules by convincing them they are inside a fictional game. Using a web page with a puzzle that rewards deliberately wrong answers, the attack gets the agent to accept a false reality, after which it treats a request to open a page and copy its contents as just another step. In the demonstration, that page redirected to the victim's work GitHub repository and the agent handed over SSH credentials, treating the theft as finishing the game. None of the six AI browser agents tested flagged it as a rule violation.
Attackers have begun exploiting a critical flaw in Oracle E-Business Suite, the financial and operations platform used by large enterprises, threat intelligence firm Defused reports. The bug (CVE-2026-46817), rated 9.8, sits in the File Transmission component of Oracle Payments and lets an unauthenticated attacker with HTTP access take over the system through a low-complexity attack. Oracle patched it in its May 2026 update, but exploitation began over the weekend despite no public proof-of-concept existing, meaning attackers built their own. Observed payloads attempt to read sensitive system files. Shadowserver tracks more than 450 EBS instances exposed online, many in North America and Asia, with unknown numbers still unpatched.
A public proof-of-concept has been released for a critical flaw in libssh2 (CVE-2026-55200), the client-side SSH library embedded in curl, Git, PHP, backup agents, firmware updaters, and countless appliances. A malicious or compromised SSH server can send a crafted packet that corrupts memory on the connecting client, with no credentials or user interaction needed, potentially leading to code execution. Rated 9.2, the bug affects all versions through 1.11.1. The fix was merged into the source on June 12, but no tagged release exists yet, so distributions are backporting it. The hardest part is that libssh2 is often statically bundled, so package updates miss those copies entirely.
Nissan has disclosed that current and former employees' data was stolen after attackers exploited a zero-day flaw in Oracle PeopleSoft, the software it uses to manage payroll, tax, and personnel records. In a filing with California's attorney general, Nissan said Oracle informed it that the personnel records of hundreds of companies may have been taken. The attacks, tied to the extortion group ShinyHunters, exploited PeopleSoft vulnerability CVE-2026-35273 as a zero-day between late May and early June, primarily hitting education organizations, before Oracle issued mitigations. ShinyHunters has begun leaking stolen data, with Nissan joining victims that include the University of Nottingham and a US insurance regulator group.
TrustStrike Labs