Just published.
The newest CVEs, straight from CVE.org and NVD.
-
CVE-2026-49405
OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before 0.56.0, the local server exposes the server-sent events endpoint at /events without bearer authentication, permits cross-origin browser access, and uses only a caller-supplied client_id to select WebChannelEvent objects from the process-wide event bus. An unauthenticated local client, or an attacker-hosted web page opened by the victim, can subscribe to another client's stream by supplying a known or guessed identifier; the fixed system identifier makes the proactive-message stream predictable. The exposed events can include complete assistant responses, streamed message content, tool names, tool arguments, and raw tool outputs intended for the local user interface. This issue is fixed in version 0.56.0.
-
CVE-2026-49407
OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before version 0.57.44, its local /ws/dictation WebSocket endpoint lacks client authentication and browser-origin validation. An unauthenticated client, including a malicious web page able to reach the endpoint, can submit audio and trigger resource-intensive speech-to-text processing. Earlier affected versions also accumulate audio in an unbounded buffer, allowing continuous streaming to exhaust process memory. These conditions can degrade performance, exhaust resources, or crash the application. This issue is fixed in version 0.57.44.
-
CVE-2026-49393
OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before 0.56.0, the core server protects HTTP POST requests to /rpc with bearer middleware but exposes Socket.IO rpc:request and chat:start handlers without equivalent transport or handler-level authentication. A local unauthenticated client can connect to the Socket.IO namespace and dispatch requests through the same protected controller path that rejects a missing bearer on /rpc. Depending on the selected method, the attacker can read configuration or health data, start agent flows, interact with integrated account APIs, alter local data, or stop and restart services. This issue is fixed in version 0.56.0.
-
CVE-2026-123151
A flaw was found in the GNU debugger (GDB). When GDB connects to a remote debug target, a malicious target can cause GDB to overflow a buffer on the stack while parsing the target's replies. This can crash GDB or potentially allow the target to execute arbitrary code in the GDB client process with the privileges of the user running GDB.
-
CVE-2026-121997
A heap buffer overflow in gdbserver's QTDPsrc packet handler allows an unauthenticated remote attacker to achieve native code execution in the gdbserver process. The vulnerability results from a mismatch between a packet's declared source length and the actual data length, causing hex2bin to write beyond the allocated heap buffer.
-
CVE-2026-121994
A stack buffer overflow in the GDB client's remote protocol handler allows an attacker-controlled debug target to execute arbitrary code in the GDB process. The vulnerability exists in fetch_register_using_p, which decodes a stub's register reply into an alloca-sized buffer without bounding writes by the register width.
-
CVE-2026-121995
A stack buffer overflow in the GDB client's tracepoint upload handler allows an attacker-controlled debug stub to corrupt stack memory automatically during connection setup. The overflow results from GDB reading beyond the current reply boundary into a stale prior reply in the shared packet buffer.
-
CVE-2026-121996
A length validation flaw in gdbserver's agent-expression parser allows an unauthenticated remote attacker to read uninitialized heap memory from the gdbserver process and cause a denial of service. Disclosed heap content includes addresses suitable for defeating ASLR in further exploitation.
-
CVE-2026-111253
A flaw was found in dnf5daemon-server. Session methods on dynamically exported D-Bus session objects do not verify that the caller is the unique sender that created the session. A local attacker can modify another user's live package-management session and replace its resolved transaction plan. If the legitimate session party later authorizes execution, the root daemon may apply the attacker-chosen package operations.
-
CVE-2026-102739
ESPectre is a Wi-Fi motion sensing platform for ESP32. Prior to 3.0.0-rc1, Micro-ESPectre analysis tools, including tools/6_optimize_filter_params.py, tools/1_analyze_raw_data.py, and tools/10_train_ml_model.py, load NPZ datasets with NumPy pickle support enabled through micro-espectre/tools/csi_utils.py and a direct loader. A malicious dataset contributor or pull-request author can place an object-typed array in a .npz file under micro-espectre/data, causing pickle-controlled code to execute as the maintainer or developer when a documented analysis tool accesses the field. This issue is fixed in version 3.0.0-rc1.
-
CVE-2026-123158
sshd in OpenSSH before 10.6p2 on Linux, when musl libc is used without PAM, lacks a workaround for a password authentication bypass (related to misinterpretation of whether a string is intended to be a password hash). For some configurations of accounts, any password between 256 and 1024 characters succeeds. An example of a platform using musl libc without PAM is Alpine Linux.
-
CVE-2026-123095
mpv 0.29.0 before 0.42.0 contains a playlist injection vulnerability in the bundled ytdl_hook script that allows remote attackers to add unchecked entries by embedding newlines in playlist URLs. Attackers controlling a feed or page can bypass url_is_safe() to open file:// or av:// inputs, reading local files or loading shared objects via ladspa for code execution.
-
CVE-2026-122328
SDL_mixer 3.1.2 through 3.2.4 contains a stack-based buffer overflow vulnerability in FetchXLaw() of src/decoder_wav.c when decoding A-law or mu-law WAV files. Attackers can supply a crafted WAV file whose doubly-offset float expansion overwrites about 1 KB of adjacent stack memory, causing crashes or potentially code execution.
-
CVE-2026-119344
Cocos AI is a confidential computing system for AI. In version 0.8.2 and earlier, Cocos AI accepts trusted execution environment attestation Evidence when firmware patch levels and software measurements match, without binding the Evidence to an allowed platform owner or unique machine identity and without enforcing revocation. A remote attacker who controls or compromises another platform trusted by the same verifier can substitute valid Evidence from that platform and cause authorization of the wrong platform. This vulnerability is addressed in 0.9.0. However, this vulnerability can't be patched in source code alone, since the vulnerability arises from the confidential computing threat model itself.
-
CVE-2026-119345
Cocos AI is a confidential computing system for AI. Prior to 0.9.0, Cocos AI performs attestation only once when an attested TLS connection is established. An attacker who can modify the attester's workload code, memory-backed state, configuration, or attached devices after that check can keep the existing connection open while the peer continues trusting stale initial Evidence. This gap is especially significant for long-lived connections because the verifier has no mechanism to detect the changed state. This issue is fixed in version 0.9.0.
-
CVE-2026-115739
Hard-coded Cryptographic Key vulnerability exists in the DLP on-prem extensions version 11.13.1 and earlier allows a remote attacker to decrypt credentials and evidence files.
-
CVE-2026-115740
Server side request forgery vulnerability exists in the DLP on-prem extensions version 11.13.1 and earlier allows the file Prefix parameter in the OpenEventEvidenceFile.do endpoint accepts arbitrary UNC paths without validation. The server-side code attempts to open the specified file path, causing Windows to resolve the hostname and initiate SMB authentication to the attacker-controlled host.
-
CVE-2026-119744
Cendio ThinLinc contains an improper authentication vulnerability in its browser-based web client (tlwebaccess) that allows an authenticated attacker to impersonate any other user, including root, by exploiting flaws in multi-step authentication flows such as one-time passwords, login banners, or expired password prompts. Attackers can leverage this authentication bypass during the multi-step login process to gain unauthorized access to other user accounts, including administrative privileges, in versions from 4.1.0 through 4.20.0.
-
CVE-2026-123032
A security flaw has been discovered in OpenBMB Internet of Agents up to 9d78e88e47160eb42ee1ee8538c05e1c2c47877a. This affects the function ConnectionManager.connect of the file im_server/app.py of the component WebSocket Connection Manager. The manipulation of the argument agent_name results in missing authentication. It is possible to launch the attack remotely. The exploit has been released to the public and may be used for attacks. This product takes the approach of rolling releases to provide continious delivery. Therefore, version details for affected and updated releases are not available. The project was informed of the problem early through an issue report but has not responded yet.
-
CVE-2026-106716
ReelShort for Android version 3.8.55 loads an attacker-controlled activityUrl supplied through its cmsvictor://open deep link in a WebView that exposes the Android and Offerwall JavaScript bridges without validating the page origin. After a user opens the crafted link, the page can invoke getUserInfo, getCredentials, and initUser to obtain user metadata and an encrypted credential payload.
-
CVE-2026-115738
SQL injection vulnerability in DLP on-prem extensions version 11.13.1 and earlier allows a remote authenticated attacker to inject SQL queries without sanitization or parameterized queries.
-
CVE-2026-106713
An issue in FreeReels for Android version 2.3.11 allows a remote attacker to obtain sensitive information via the com.dramawave.feature.web.WebPageActivity component
-
CVE-2026-106714
Chewy for Android version 26.23.0 does not adequately validate the host of a product-details deep link before loading it in a Hybrid WebView. An attacker-controlled page with a /dp/<number> path can access the injected nma_petmd_bridge JavaScript interface and invoke getPetProfiles, disclosing a signed-in user's pet profile information after the user opens the crafted link.
-
CVE-2026-106715
NewsBreak for Android version 26.21.0 treats data:text/html; URLs as trusted in its WebView JavaScript bridge origin check. An attacker can supply such a page through the newsbreak://openweb deep link and invoke NBJS.callNative with the getLoginInfo action, disclosing a signed-in user's account information and session token after the user opens the crafted link.
-
CVE-2026-104720
Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal') vulnerability in Kobikom Telecommunications and Informatics Services Inc. Kobikom allows Path Traversal. This issue affects Kobikom: before KobiPBX-FOP3-SEC-r1.
-
CVE-2026-105321
Improper neutralization of special elements used in an SQL command ('SQL injection') vulnerability in Kobikom Telecommunications and Informatics Services Inc. Kobikom allows SQL Injection. This issue affects Kobikom: before KobiPBX-FOP3-SEC-r1.
-
CVE-2026-105330
Use of Hard-coded Cryptographic Key in JWT token authentication in Apache OFBiz 24.09.06 and 24.09.07 on all platforms allows an unauthenticated, network-reachable attacker to obtain full administrative access via a forged JWT signed with the factory-default security.token.key value that CVE-2026-31986 removed from newly generated configurations, on a deployment that upgraded from a release predating that fix while retaining its existing security.properties with security.internal.sso.enabled set to true. Users are recommended to upgrade to version 24.09.08. Upgrading alone does not replace an already-configured key: deployments that carried framework/security/config/security.properties forward from a release predating the CVE-2026-31986 fix must also run './gradlew rotateSecretKeys' after upgrading, since the existing generateSecretKeys task leaves an already-set value untouched.
-
CVE-2026-105461
Missing check for ORDERMGR_UPDATE in ordermgr allows user with only ORDERMGR_VIEW permission to update orders. Affected versions before : 24.09.08 Users are recommended to upgrade to version 24.09.08 which fixes this issue.
-
CVE-2026-95652
Execution with unnecessary privileges vulnerability in TÜBİTAK BİLGEM Software Technologies Research Institute pardus-usb-formatter allows Privilege Abuse. This issue affects pardus-usb-formatter: from 0.7.0 before 0.7.2.
-
CVE-2026-118869
An OS command injection vulnerability exists in SHIRASAGI, which may allow an authenticated attacker to execute arbitrary OS commands on the system running the product.
-
CVE-2026-123029
Wowza Streaming Engine before 4.12.2 processes XML external entities, which allows remote attackers to read local files with the privileges of the Wowza Streaming Engine process.
-
CVE-2026-123033
Authentication bypass in the HA heartbeat origin check in middlewared in TrueNAS SCALE before 25.10.8 (and in release trains before their respective fixes) allows an attacker on the same network segment to achieve unauthenticated full-administrator API access and execute commands as root. middlewared registers the HA peer-authentication hook (ha_permission) unconditionally, with no failover-licence or heartbeat-interface check. An unauthenticated connection whose TCP peer address is a hardcoded HA heartbeat address (169.254.10.1 or 169.254.10.2) and whose peer port is <= 1024 is granted a node credential that is authorized for every API method and treated as full administrator. Both conditions are properties of the attacker's own socket, and thus an attacker on the same layer-2 segment satisfies them without forging anything. The hook is present on non-HA, unlicensed, single-node Community Edition installations. The fix additionally requires that the connection terminate on the HA heartbeat listener (a local address/port the attacker cannot control).
-
CVE-2026-123017
GodotSteam 4.17 through 4.23.2 trusts the peer value of 1 to mean that a player is the server. A grant_gold(1000000) call can be forged.
-
CVE-2026-123014
iTerm2 before 3.7.7beta2 has an argument injection vulnerability in the x-man-page: URL handler, allowing an attacker to execute arbitrary commands with the privileges of the logged-in user by inducing the user to open a crafted URL in iTerm2. Percent-encoded NUL characters survive input sanitization and are subsequently converted to whitespace during command-line parsing, allowing injection of additional arguments to /usr/bin/man. An attacker can inject the -P option to select an attacker-controlled pager command. NOTE: this issue exists because of an incomplete fix for CVE-2023-46321.
-
CVE-2026-123011
STMicroelectronics st-mems-android-linux-sensors-hal before 7bdf334 does not check whether the allowed number of channels is exceeded.
-
CVE-2026-69005
Integer overflow in NPU driver parameter processing due to insufficient bounds checking.
-
CVE-2026-69006
Insufficient validation of length fields in the GNSS firmware loader can result in kernel‑memory corruption.
-
CVE-2026-69025
In the Exynos NPU driver, improper validation of user input in the ioctl handler can lead to an out-of-bounds write, which causes a denial of service
-
CVE-2026-69043
In the Samsung Exynos GPU driver, improper synchronization during buffer object creation and closing can lead to memory corruption
-
CVE-2026-69044
In the Samsung GPU driver, improper reference count handling during command submission can lead to memory corruption.
-
CVE-2026-69050
Out-of-bounds memory access in a device driver due to improper resource state tracking.
-
CVE-2026-69066
Memory corruption in GPU driver due to improper validation of VMID reservation.
-
CVE-2026-69068
Insufficient length validation in the GNSS firmware loader may lead to corruption of kernel memory.
-
CVE-2026-68984
In the Samsung Exynos NPU kernel driver, an integer overflow allows an unprivileged application to trigger memory corruption.
-
CVE-2026-68985
In the Samsung Exynos NPU driver, improper validation of user input in the ioctl handler can lead to an out-of-bounds write, which causes a denial of service.
-
CVE-2026-68990
In the Samsung Exynos H-Arx DRM plugin, improper input validation during buffer operations can cause memory corruption, potentially leading to denial of service or privilege escalation.
-
CVE-2026-68991
In the Samsung Exynos ENN NeuralNetworks HAL, improper input validation during request processing can cause memory corruption and lead to a denial of service condition.
-
CVE-2026-68993
A double free vulnerability in NPU driver due to improper pointer management (940)
-
CVE-2026-68994
A double free vulnerability in NPU driver due to improper pointer management (941)
-
CVE-2026-69002
Incorrect authorization in Samsung framework services due to missing permission checks
The newest published CVE records from CVE.org and NVD, refreshed every few minutes, with severity, exploit odds and fix status.