The GeoDirectory WordPress plugin before 2.8.194 does not perform any status or ownership check when rendering a listing's map-popup through one of its single-item REST routes, allowing unauthenticated users to read the title, permalink and card fields of non-public (draft, pending, private and trashed) listings by supplying the listing ID.
Prior to version 26.10.0rc1, the Twisted framework's DNSDatagramProtocol component accepts UDP DNS responses based solely on matching the 16-bit transaction ID without verifying the source host and port. This allows an attacker who can send packets to the resolver's listening port with the correct transaction ID to inject forged DNS answers. Such forgery can redirect hostname resolution, potentially enabling mail interception, server-side request forgery, or man-in-the-middle attacks. The issue is fixed in version 26.10.0rc1.
A vulnerability in the Zephyr DNS stub resolver causes it to use a fixed UDP source port per configured server for DNS queries, violating RFC 5452 requirements. This allows an off-path attacker with UDP reachability to the device to predict the source port and attempt to spoof DNS responses by guessing only the 16-bit transaction ID. Successful spoofing can cause the application to resolve attacker-chosen addresses, potentially poisoning the DNS cache if enabled. The vulnerability is mitigated by a fix that randomizes the source port per query, increasing the guessing space.
CVE-2026-79980 is a medium-severity vulnerability in the DHCPv4 client of the Zephyr project. The client used a predictable transaction identifier that was incremented or reused rather than randomly generated per transaction as required by RFC 2131. This flaw allows an attacker who learns or predicts the transaction ID to forge DHCP replies that can redirect network traffic by altering gateway and DNS settings. The issue affects Zephyr versions from 1.7.0 up to but not including 4.5.0. A fix is available in version 4.5.0, which ensures fresh random transaction IDs are generated for each DHCP transaction phase.
CVE-2026-55222 is a vulnerability in ilbers isar versions 0.11 through 1.1 where the /etc/apt/trusted.gpg.d directory is created with ownership of the calling user instead of root during bootstrap setup. This allows an unprivileged user inside the image with a matching user ID to alter or add apt keys, potentially undermining apt repository integrity. The issue does not affect images with immutable root filesystems or those that never invoke apt. Exploitation requires additional steps such as manipulating a remote repository or intercepting insecure communications. The vulnerability has been patched in a commit identified as 29e84ed.
CVE-2026-114357 is a vulnerability in the DJI CrystalSky ZS600C firmware versions 0.3.9.7 through 0.6.6.9. It involves missing authentication in the Android Debug Bridge daemon (adbd), allowing attackers to obtain an unauthenticated shell with user ID 2000. The device ships with ADB host key authorization disabled and USB ADB enabled by default. If ADB over TCP is enabled, the unauthenticated shell can be accessed remotely over the network. This vulnerability has a high severity rating with a CVSS 4.0 base score of 7.
A vulnerability in the Linux kernel's gpio exar driver was fixed by ensuring the device ID is initialized before registering its cleanup action. The flaw caused improper handling of device IDs during failure in action registration, potentially leading to ID leaks or incorrect freeing of IDs belonging to other devices.
CVE-2026-98940 is a vulnerability in the Linux kernel's ASoC Qualcomm lpass-cpu driver where the code incorrectly uses the DAI ID as an index into the DAI driver table. This leads to out-of-bounds access when the DAI IDs are sparse and do not correspond to their position in the driver table. The issue is fixed by using the loop index for table access instead of the DAI ID.
A vulnerability in the Linux kernel's xen/gntalloc component was fixed by validating the grant count before allocation. The flaw involved allocating a grant-id array without first ensuring the requested count was within the global grant limit, potentially leading to improper memory allocation. The fix adds a subtraction-based check under a mutex to prevent integer wraparound and ensures 64-bit arithmetic is used for size calculations.
Capcom suffered a data breach impacting approximately 11.7 million email addresses from its CAPCOM ID service. The breach also involved unauthorized changes to login passwords for 196 accounts. Additional personal information such as dates of birth, gender, country/region, and nicknames were compromised for 43 users. Capcom has reset affected passwords and notified impacted users. The company reported the incident to authorities and continues to investigate while maintaining service availability.
Authlib versions prior to 1.9.0 contain a vulnerability in the RFC 9101 JWTAuthenticationRequest implementation where the unauthenticated, client-controlled request_uri parameter is passed to a hook before destination validation. This can allow a remote attacker with a valid public client ID to induce the authorization server to fetch unintended internal or link-local URLs if the integrator's hook performs unrestricted outbound fetches. The issue is fixed in version 1.9.0.
Authlib versions prior to 1.9.0 contain a vulnerability in OpenIDMixin.parse_id_token() where the nonce_supported decision is read from an untrusted ID token. This flaw allows an attacker with a captured validly signed token to bypass nonce-to-session binding and replay the token into an attacker-controlled session in implicit and hybrid flows. The vulnerability does not affect the plain authorization code flow and requires the token to pass signature, issuer, and audience validation. The issue is fixed in version 1.9.0.
A vulnerability in Authlib versions prior to 1.9.0 allows a registered client that is permitted an ID token response type but not the openid scope to add openid to the request, obtain user consent, and receive a signed ID token containing the user's subject and allowed profile claims. This occurs because OpenIDImplicitGrant.validate_authorization_request() improperly checks the client-controlled raw request.payload.scope instead of the filtered request.scope. Authorization code flow and deployments that allow openid for every client permitted an ID token response are not affected. The issue is fixed in version 1.9.0.