A flaw in All-in-One WP Migration and Backup, a WordPress plugin installed on millions of sites, can let an unauthenticated attacker take over a site. Tracked as CVE-2026-19949, it is a second-order SQL injection caused by incorrect handling of escaped characters when the plugin rewrites database content during a restore. An attacker plants crafted data through WordPress trackbacks, which triggers when an administrator exports and imports the site, both routine plugin operations. The injection can leak the plugin's secret import key through a public comment, letting the attacker import a malicious backup archive containing executable code and seize full control. ServMask fixed it in version 7.110, but many sites remain unpatched.
Attackers are actively probing LiteLLM deployments for an authorization flaw that turns a low-privilege account into full control of the AI gateway. Tracked as CVE-2026-35029 and affecting versions before 1.83.0, the flaw is a missing permission check on the configuration-update endpoint, so a read-only user can change settings reserved for administrators. LiteLLM sits between applications and model providers and stores provider API keys, database details, and admin credentials, making it a rich target. By abusing configuration writes, an attacker can extract secrets from server environment files and even reset the dashboard login to seize admin access. Researchers recorded thousands of probing requests, some directly attempting to read known secret files.
Days after disclosure, attackers are exploiting a critical authentication-bypass flaw in JFrog Artifactory, the widely used repository manager for binaries, packages, containers, and build artifacts. Tracked as CVE-2026-82329 and scored 9.8, the flaw lets an unauthenticated attacker with network access gain administrative privileges under Artifactory's default configuration. Researchers at watchTowr observed exploitation beginning September 1, with attackers minting admin tokens for themselves and enumerating users, groups, and credentials. Because Artifactory sits at the center of software supply chains and CI/CD pipelines, admin access lets attackers tamper with build pipelines, poison trusted dependencies, and push malicious code downstream to customers. JFrog patched it in version 7.161.20 on August 28.
VulnCheck reported active exploitation of two critical flaws, one in the AI workflow builder Langflow and one in Ruby on Rails. The Langflow bug, CVE-2026-0768, scored 9.8, lets an attacker run arbitrary Python code as root through improper input validation. The Rails bug, CVE-2026-66066 and nicknamed KindaRails2Shell at 9.5, lets an unauthenticated attacker read arbitrary files by uploading a crafted image that exploits a mismatch between Active Storage and the libvips image library, leaking secrets like the Rails master key and cloud credentials and ultimately enabling code execution. Detections jumped from about 50 to 360 within a day, with attackers querying environment variables for OpenAI and AWS keys and probing SSH access.
Researchers at Wordfence and Patchstack disclosed a cluster of critical WordPress vulnerabilities, most scored 9.8, that let unauthenticated attackers take over sites or run code. In the WPMU DEV Dashboard plugin, CVE-2026-76581 is a single-sign-on authentication bypass that can hand an attacker an administrator session. The Avada theme's CVE-2026-18431 allows arbitrary file writes that lead to remote code execution. In the Pods plugin, CVE-2026-19598 lets an attacker escalate to administrator or overwrite any user's password, while TranslatePress's CVE-2026-19632 exposes the raw administrator password-reset link. Each independently enables full site compromise, and a separate GiveWP flaw in the same batch was covered earlier.
CISA added a critical ownCloud flaw to its exploited-vulnerabilities catalog after attackers used it to steal data from a research organization. Tracked as CVE-2023-49105 and scored 9.8, the WebDAV authentication-bypass bug lets an unauthenticated attacker who knows a victim's username read, modify, or delete that user's files when no signing key is configured, which is the platform's default. Disclosed back in November 2023, it affects ownCloud Server core versions 10.6.0 through 10.13.0 and was fixed in 10.13.1, yet unpatched instances remain exposed nearly two years later. Public exploit code exists, and CISA set a short deadline for federal agencies, underscoring that long-standing self-hosted flaws keep getting weaponized.
A critical flaw in GiveWP, a WordPress donation and fundraising plugin installed on more than 100,000 sites, lets an unauthenticated attacker run commands on the hosting server. Tracked as CVE-2026-82222, it chains three weaknesses: an unsafe PHP deserialization helper, a donation flow that stores attacker-controlled serialized objects, and a gadget chain in bundled libraries that turns that into system command execution. Although exploitation normally needs an account, an exposed registration action lets an attacker create one even when registration is disabled, making it effectively unauthenticated. It affects versions up to 4.16.7.1 and is fixed in 4.16.7.2, which blocks serialized data during donation processing.
A researcher disclosed two root remote code execution chains in the Unitree G1 humanoid robot, one reachable over Bluetooth from nearby without any pairing. Tracked as CVE-2026-76639 and CVE-2026-76640, the Bluetooth chain abuses a gap in Unitree's cloud service, which handed over another robot's key material to any free account without checking ownership, then used a buffer overflow in the Wi-Fi provisioning code to run code as root on the robot's control computer. The researcher demonstrated that a compromised robot can spread the exploit to another within Bluetooth range, making it wormable. There is no confirmed fixed firmware for the on-robot flaws, though the cloud ownership check was tightened.
ServiceNow patched four flaws in its widely used AI Platform, three of them scored 10.0 and exploitable by an unauthenticated attacker with no user interaction. The first, CVE-2026-18885, is a code injection in the GraphQL Composite Data API that allows arbitrary code execution and access to instance data. The second, CVE-2026-18886, is an access-control flaw in the configuration image-upload processor that lets an attacker create or modify data and escalate privileges. The third, CVE-2026-74820, is a SQL injection allowing arbitrary queries against the instance database. ServiceNow fixed hosted instances itself, but self-managed customers must apply the updates. No exploitation of these three has been reported yet.
PaperCut warned that attackers are actively exploiting vulnerabilities in all versions of its widely deployed NG and MF print-management software, and confirmed real customer incidents. Two flaws are involved: CVE-2026-82078, unsafe dynamic class loading in the database connection utilities that lets an attacker run arbitrary Java bytecode, and CVE-2026-81578, an access-control flaw in the web management interface that lets an unauthenticated attacker change system configuration. PaperCut released emergency out-of-cycle patches for its version 25 and 26 branches, with a version 24 build still in progress, and later issued a hardened second release. PaperCut servers have a history of being targeted by ransomware crews.