Sandboxie-Plus versions prior to 1.18.6 contain a vulnerability where a sandboxed process can send specific messages to the SYSTEM service (SbieSvc) to modify files outside the sandbox. This allows changes to file attributes, timestamps, writeability, or 8.3 short names. Exploiting short-name redirection may lead to SYSTEM code execution if a privileged process resolves an attacker-controlled short-name path accessible by SbieSvc. The issue is fixed in version 1.18.6.
CVE-2026-122192 is a high-severity vulnerability in AWS Firecracker affecting versions 0.23.0 through 1.16.1. It involves improper handling of a failed memory remap when restoring microVMs from snapshots using the File memory backend. A guest user with root privileges could exploit this flaw to read or modify the Firecracker VMM process memory via balloon or virtio-mem discard requests if the host fails to create the replacement mapping. AWS has fixed this issue in Firecracker version 1.16.2 and later.
A vulnerability in django-rest-framework versions 3.12.0 through 3.18.1 causes TokenProxy.delete() to fail silently when attempting to revoke tokens. This occurs because the deletion query uses the proxy's user-id primary key instead of the token-string primary key, resulting in zero deletions without error. Consequently, tokens may remain active and continue authenticating despite apparent revocation. The issue is fixed in version 3.18.2.
CVE-2026-111823 is a critical vulnerability in Mage OSS mage-ai versions up to 0.9.80. The issue involves the software proceeding with session creation even when encountering an unknown OAuth provider, which is a detection of an error condition without appropriate action (CWE-390). This flaw can lead to complete compromise of confidentiality, integrity, and availability of the affected system.
Discord libdave before 1.2.0 did not reject an MLS Welcome message when the resulting group roster contained an unrecognized participant. An attacker in control of the DAVE signaling path (the voice gateway, or an equivalent position able to add, alter, or withhold signaling messages to a client) could cause affected clients to accept an unauthorized member into the end-to-end encrypted media session, compromising the confidentiality and integrity of audio and video.
CVE-2026-104002 is a medium severity vulnerability in the AWS Powertools for AWS Lambda (Python) data masking utility. It involves a fail-open error handling issue that may allow sensitive field values intended to be masked to be read by unauthorized actors. The issue affects versions from 3.6.0 up to and including 3.34.0. AWS has released version 3.35.0 to remediate this vulnerability.
AsyncHttpClient versions from 3.0.8 up to but not including 3.0.12 contain an improper authentication vulnerability in the SCRAM and Digest authentication processing. The library logs mismatches in server signature verification but still accepts the response as authenticated, potentially allowing an unauthenticated peer to be accepted as the server on non-TLS or compromised transports. This issue is fixed in version 3.0.12.
A vulnerability in libssh affects the ProxyCommand feature, where an unchecked fork() failure can be recorded as process ID -1. During cleanup, this can cause signals to be sent across the caller's accessible process tree, resulting in a local denial of service. This issue impacts environments using ProxyCommand in libssh client configurations.
Detection of Error Condition Without Action vulnerability in Apache Tomcat when configuring CRLs for a FFM based connector.
This issue affects Apache Tomcat: from 11.0.0-M1 through 11.0.22, from 10.1.0-M7 through 10.1.55, from 9.0.83 through 9.0.118.
Users are recommended to upgrade to version 11.0.23, 10.1.56 or 9.0.119, which fixes the issue.
A vulnerability in pam_usb versions prior to 0.9.1 causes the software to silently ignore permission errors when accessing input device nodes, leading to a false negative detection of virtual input devices. This results in the authentication process continuing without denial when it should not, potentially allowing unauthorized access. The issue is fixed in version 0.9.1.
Red Hat Discovery, also known as Discovery, is an inspection and reporting tool that finds,
identifies, and reports environment data, or facts, such as the number of physical and virtual
systems on a network, their operating systems, and relevant configuration data stored within
them. Discovery also identifies and reports more detailed facts for some versions of key
Red Hat packages and products that it finds in the network.
Gitsign is a keyless Sigstore to signing tool for Git commits with your a GitHub / OIDC identity. From 0.4.0 to before 0.15.0, CertVerifier.Verify() in pkg/git/verifier.go unconditionally dereferences certs[0] after sd.GetCertificates() without checking the slice length. A CMS/PKCS7 signed message with an empty certificate set is a structurally valid DER payload; GetCertificates() returns an empty slice with no error, causing an immediate index-out-of-range panic. On the gitsign --verify code path (the GPG-compatible mode invoked by git verify-commit), the panic is silently recovered by internal/io/streams.go's Wrap() function, which returns nil instead of an error. main.go then exits with code 0, causing exit-code-only verification callers to interpret the failed verification as success. This vulnerability is fixed in 0.15.0.
Improper handling of error condition during host-induced faults can allow a local high-privileged attack to selectively drop guest DMA writes, potentially resulting in a loss of SEV-SNP guest memory integrity
Dell Alienware Command Center 6.x (AWCC), versions prior to 6.10.15.0, contain a Detection of Error Condition Without Action vulnerability. A low privileged attacker with local access could potentially exploit this vulnerability, leading to Arbitrary Code Execution.
CVE-2025-26465 is a vulnerability in OpenSSH affecting version 6.8p1 when the VerifyHostKeyDNS option is enabled. It allows a machine-in-the-middle attack due to improper handling of error conditions during host key verification. Successful exploitation requires the attacker to exhaust the client's memory resources, making the attack complexity high. A security update addressing this issue is available from Red Hat for their Enterprise Linux 8 distributions.
Envoy is a cloud-native, open source edge and service proxy. The HTTP/2 protocol stack in Envoy versions prior to 1.29.3, 1.28.2, 1.27.4, and 1.26.8 are vulnerable to CPU exhaustion due to flood of CONTINUATION frames. Envoy's HTTP/2 codec allows the client to send an unlimited number of CONTINUATION frames even after exceeding Envoy's header map limits. This allows an attacker to send a sequence of CONTINUATION frames without the END_HEADERS bit set causing CPU utilization, consuming approximately 1 core per 300Mbit/s of traffic and culminating in denial of service through CPU exhaustion. Users should upgrade to version 1.29.3, 1.28.2, 1.27.4, or 1.26.8 to mitigate the effects of the CONTINUATION flood. As a workaround, disable HTTP/2 protocol for downstream connections.