CVE Feed

ExLlamaV3’s EXL3 loader uses an untrusted, model-supplied index to select a CUDA kernel without first verifying that the index is within bounds. A malicious model can force the code to read and invoke an out‑of‑range function pointer, resulting in a reproducible crash. Under specific conditions involving certain host memory layouts, it may be possible for an attacker to exploit this vulnerability to achieve controlled indirect kernel dispatch during model load.

A flaw has been found in Edimax BR-6208AC 1.32. Affected by this issue is some unknown functionality of the file /goform/formWlSiteSurvey. This manipulation of the argument ifname causes buffer overflow. Remote exploitation of the attack is possible. The exploit has been published and may be used. The vendor was contacted early about this disclosure but did not respond in any way. This vulnerability only affects products that are no longer supported by the maintainer.

Progress MarkLogic Server 12.x releases before 12.1.0 and 11.x releases before 11.3.7 do not sufficiently sanitize input used by the Query Console file-exploration endpoint, which could allow HTTP response splitting and unintended client-side response handling.

Progress MarkLogic Server 12.x releases before 12.1.0 and 11.x releases before 11.3.7 may use cleartext transport for Amazon S3 connections by default. A network-positioned attacker could intercept data or credentials transmitted through an affected connection.

Progress MarkLogic Server 12.x releases before 12.1.0 and 11.x releases before 11.3.7 do not consistently validate TLS certificate hostnames for certain outbound connections. A network-positioned attacker could exploit this weakness to intercept or modify communications with an affected service.

Progress MarkLogic Server 12.x releases before 12.1.0 may use cleartext transport for certain cluster-management communications. A network-positioned attacker could intercept or modify information transmitted through an affected connection.

A security flaw has been discovered in Netcore NR268 1.7.121109. Affected by this issue is some unknown functionality of the component CGI Endpoint. Performing a manipulation results in cross-site request forgery. The attack may be initiated remotely. The exploit has been released to the public and may be used for attacks. The vendor was contacted early about this disclosure but did not respond in any way.

Improper Handling of Highly Compressed Data (Data Amplification) vulnerability in hexpm hex_core allows an untrusted repository mirror, or a server answering hex_core's API requests, to exhaust the memory of the Erlang VM running hex_core. The same code is vendored in Hex (mix_hex_repo, mix_hex_api) and rebar3 (r3_hex_repo, r3_hex_api). hex_repo:get_names/1, hex_repo:get_versions/1, hex_repo:get_package/2 and hex_repo:get_policy/2 decompress the registry response with zlib:gunzip/1 and decode the signed envelope before verifying its signature. A mirror doesn't have the repository's private key, but it chooses the compressed bytes, so a small response decompresses into an envelope of any size before the signature check rejects it. hex_api decodes responses with the application/vnd.hex+erlang content type through hex_safe_binary_to_term:safe_binary_to_term/1, which calls binary_to_term(Binary, [safe]). The safe option prevents atom creation, but it accepts compressed terms and doesn't limit the size of the decoded term, and the check that rejects functions and ports runs after the term has been built. The fix for CVE-2026-21619 added the safe option and that check without bounding the decoded size. This issue affects rebar3: from 3.9.1 before 3.27.3; hex: from 2.1.0 before 2.5.2; hex_core: from 0.1.0 before 0.20.0.

Progress MarkLogic Server 12.x releases before 12.1.0 and 11.x releases before 11.3.7 do not sufficiently verify the identity of a Kerberos authentication service in certain password-based authentication configurations. A network-positioned attacker could impersonate trusted authentication infrastructure, potentially resulting in unauthorized access or modification of protected data.

Progress MarkLogic Server 12.x releases before 12.1.0 and 11.x releases before 11.3.7 assign overly permissive operating-system permissions to certain log files, which could allow a local user to read or alter log data. The corrected default permissions apply to newly created and rotated logs; existing logs retain their current permissions until remediated.

Progress MarkLogic Server 12.x releases before 12.1.0 and 11.x releases before 11.3.7 enable XML external entity processing by default in multiple server paths, which could allow crafted XML to cause server-side requests or disclosure of local files.

Progress MarkLogic Server 12.x releases before 12.1.0 and 11.x releases before 11.3.7 contain an out-of-bounds write vulnerability in certain language-processing functionality. A remote authenticated attacker could exploit this vulnerability to cause server instability or limited modification of data.

Progress MarkLogic Server 12.x releases before 12.1.0 contain an out-of-bounds read vulnerability in certain query-processing functionality. A remote authenticated attacker could exploit this vulnerability to cause server instability or disclose limited information from process memory.

Progress MarkLogic Server 12.x releases before 12.1.0 and 11.x releases before 11.3.7 contain a cross-site scripting vulnerability that could allow a remote attacker to execute script in the context of an authenticated user's browser session, which could result in unauthorized actions or disclosure of information accessible to that user.

Missing Authorization vulnerability in Arraytics Booktics booktics allows Exploiting Incorrectly Configured Access Control Security Levels.This issue affects Booktics: from n/a through 1.0.28.

Progress MarkLogic Server 12.x releases before 12.1.0 and 11.x releases before 11.3.7 do not sufficiently restrict access to certain security-sensitive information. A remote authenticated attacker with limited management permissions could obtain this information, which could result in unauthorized access to protected resources.

A vulnerability was identified in Netcore NR268 1.7.121109. Affected by this vulnerability is the function eval of the file /router/l7_call_board_user_set.cgi of the component Web UI. Such manipulation of the argument title/note/URL/content leads to cross site scripting. The attack can be launched remotely. The exploit is publicly available and might be used. The vendor was contacted early about this disclosure but did not respond in any way.

Progress MarkLogic Server 12.x releases before 12.1.0 and 11.x releases before 11.3.7 contain memory-safety defects in the JavaScript engine integration that could result in memory corruption, controlled process failure, or server instability.

Progress MarkLogic Server 12.x releases before 12.1.0 and 11.x releases before 11.3.7 allow an authenticated Query Console user to supply unsafe path values to a privileged filesystem helper, which could result in disclosure of arbitrary server files.

Progress MarkLogic Server 12.x releases before 12.1.0 and 11.x releases before 11.3.7 include deprecated OpsDirector functionality with weaknesses that could allow an unauthenticated remote attacker to make unauthorized requests to internal services or forge trusted requests, potentially resulting in unauthorized access to protected information.

A flaw was found in Hibernate Validator starting with version 5.4.1. This vulnerability allows an unauthenticated remote attacker to cause a Denial of Service (DoS) by submitting specially crafted input containing unescaped message parameters. Improper escaping allows user-supplied data to be re-evaluated during Expression Language (EL) processing, leading to excessive resource consumption and application unresponsiveness.

A vulnerability was determined in Netcore NR268 1.7.121109. Affected is the function user_ok of the component CGI Execution. This manipulation causes improper authentication. The attack can be initiated remotely. The exploit has been publicly disclosed and may be utilized. The vendor was contacted early about this disclosure but did not respond in any way.

Progress MarkLogic Server 12.x releases before 12.1.0 and 11.x releases before 11.3.7 contain an improper authentication vulnerability in SAML processing. In deployments that use SAML authentication, an unauthenticated remote attacker could cause the affected server to accept an assertion that should be rejected, potentially resulting in unauthorized access.

On Azure deployments, Progress MarkLogic Server 12.x releases before 12.1.0 and 11.x releases before 11.3.7 may record security-sensitive authentication information in diagnostic logs under certain configurations. An attacker with access to the affected logs could use this information to obtain unauthorized access to protected cloud resources.

Progress MarkLogic Server 12.x releases before 12.1.0 and 11.x releases before 11.3.7 use predictable CSRF tokens in Query Console. A remote attacker who lures an authenticated Query Console user to a malicious page could submit unauthorized requests and execute server-side code with the victim's permissions.

Progress MarkLogic Server 12.x releases before 12.1.0 and 11.x releases before 11.3.7 do not consistently record successful authentication events in audit logs, which could limit customers' ability to detect and investigate unauthorized access or credential misuse.

Quarkus REST Data Panache provides a way to automatically generate REST endpoints for data entities. A flaw was found where certain endpoints do not properly check the sort parameter provided by users. This could allow an attacker to learn information they should not have access to or slow down the system by forcing the database to perform complex tasks, potentially causing a denial of service.

ZITADEL 3.0.0 before 4.19.2 contains a privilege escalation vulnerability in OIDC token exchange that allows end-user impersonators to impersonate administrators including IAM_OWNER by bypassing the admin.impersonation check. Attackers holding IAM_END_USER_IMPERSONATOR or ORG_END_USER_IMPERSONATOR roles can sign in to relying applications as administrators with their roles and claims.

ZITADEL 3.0.0 before 4.19.2 contains an authentication bypass vulnerability in the Login V2 UI that allows unauthenticated attackers to obtain other users' session tokens by forging unsigned session cookies. Attackers can guess time-ordered session IDs around a victim's sign-in and supply them with empty tokens to receive rotated tokens, taking over accounts including MFA-completed administrator sessions.

ZITADEL 4.x before 4.19.2 and 3.x through 3.4.15 contains an authentication bypass in the SAML identity-provider intent callback, which verifies the assertion signature using the provider named in the request URL but resolves the user from the login intent without checking they match. Attackers with organization-administrator permission can register a SAML provider with their own signing key to obtain authenticated sessions as any user linked to another SAML provider on the instance, including in other organizations.

ZITADEL 3.0.0 before 4.18.0 contains a missing authentication vulnerability in the hosted Login V1 UI, where the change-username handler acts on the account bound to a login flow before any factor is verified. Unauthenticated attackers can submit a victim's login name and then post a new username to rename any active human user, including administrators, locking them out.

Zebra before 6.4.0 contains an improper integrity validation vulnerability in the inbound gossip download path that allows unauthenticated peers to force dropping of new blocks without penalty. Attackers can serve a real block with a coinbase scriptSig rewritten to height 1, which keeps the ZIP-244 block hash, triggering a behind-tip drop before consensus validation and delaying tip discovery for mining backends.

Zebra before 6.4.0 contains an uncontrolled resource consumption vulnerability that allows unauthenticated peers to force full proof verification by serving forged block bodies with valid header hashes. Attackers can flip authorizing-only bytes such as a coinbase scriptSig, wasting Sapling, Orchard, and Ironwood proof verification without penalty and stalling sync for 67 seconds.

Zebra 2.3.0 before 6.4.0 contains a protection mechanism failure that allows unauthenticated peers to evade bans because bans and misbehavior scores are keyed on exact IP addresses. Attackers can reconnect from fresh IPv6 addresses in their /64 allocation or flush the 20,000-entry FIFO ban list to restore access for banned peers.

ZcashFoundation Zebra 6.4.0 before 6.4.2 parses version 6 transactions carrying a pre-NU6.3 consensus branch ID, but the size computation during UnminedTx construction treats re-serialization of the Orchard bundle as infallible, so the resulting error trips an expect that aborts zebrad. Unauthenticated peers can send a single pre-handshake tx message to crash the node, and can repeat this after each restart.

Heym v0.0.5 before v0.0.113 contains a broken access control vulnerability that allows former collaborators to keep executing JWT-auth workflows by reusing scoped workflow:execute tokens after share removal. Attackers can mint a token with a TTL up to 10 years, which the owner cannot list or revoke, and repeatedly run the workflow consuming shared credentials.

MONAI 1.2.0 before 1.6.1 contains an insecure deserialization vulnerability in nnUNetV2Runner that unpickles the postprocessing_file path read from inference_information.json without confining it to the results directory. Attackers who can modify inference_information.json can point it at a malicious pickle elsewhere on the filesystem to execute code when predict_ensemble_postprocessing runs.

MONAI 0.4.0 before 1.6.1 contains a remote code execution vulnerability that allows on-path attackers to execute code by tampering with SENet pretrained weights fetched over plaintext HTTP. Attackers can substitute a crafted .pth file that load_state_dict_from_url unpickles without weights_only=True, executing code when SEResNet50 or other SENet models are constructed with pretrained=True.

Contrast 1.8.0 before 1.24.1 contains an improper verification vulnerability in Coordinator peer recovery that fails to confirm the serving peer uses the same manifest generation. Attackers with write access to manifest history can replay a superseded history head so withdrawn workloads obtain trusted mesh certificates, or with the secret seed disclose the Mesh CA key.

Edgeless Systems Contrast 1.8.0 before 1.24.1 contains an incorrect permissions vulnerability where the node installer writes the containerd config world-writable (0666). Local unprivileged users on the node can modify the containerd config, for example overwriting the runc location, to escalate privileges.

LibreNMS before 26.9.0 contains an LDAP filter injection in ldap-authorization and ad-authorization modes because usernames are inserted unescaped into LDAP search filters and authentication only checks user existence. When REMOTE_USER is not forwarded, unauthenticated attackers can send a wildcard username via HTTP Basic Authorization (PHP_AUTH_USER) to match all users and groups and gain admin access.

LibreNMS before 26.9.0 contains a missing authorization vulnerability in PortsController::purge that allows global-read users to physically delete ports without delete privileges. Attackers with read-only accounts can send DELETE requests to /ports/purge, using purge=all to irreversibly erase every deleted port on all visible devices.

LibreNMS before 26.9.0 contains a stored cross-site scripting vulnerability in get_oxidized_nodes_list() that renders Oxidized API fields and SNMP sysName unescaped. Attackers controlling the Oxidized response or a device's SNMP sysName can inject script that executes in browsers of users viewing the Tools Oxidized page.

LibreNMS before 26.9.0 contains a path traversal vulnerability in nfsen_channel_common.inc.php that allows authenticated users to read unauthorized RRD data by injecting ../ sequences into the channel parameter. Attackers with access to one NFSen-enabled device can request graph.php with a traversed channel value to view traffic and NetFlow graphs of devices outside their authorized scope.

LibreNMS before 26.9.0 contains an SQL injection vulnerability that allows unprivileged users to inject SQL by supplying an array-valued notification_id parameter to notifications.inc.php. Attackers can perform cross-table reads of users.password hashes, with results persisted in notifications_attribs and readable through the interface.

LibreNMS before 26.9.0 contains a stored cross-site scripting vulnerability in includes/html/pages/device/wireless.inc.php, which echoes wireless sensor sensor_descr values, populated from SNMP DisplayString fields during discovery, into HTML without escaping. Attackers controlling a monitored device's SNMP agent can set values such as AI-AP-MIB::aiSSID to malicious JavaScript that executes when authenticated users view the device's Wireless tab, enabling session hijacking.

Gitoxide gix-pack before 0.75.0 contains an uncontrolled memory allocation vulnerability that allows remote attackers to abort the process by supplying a huge pack-header object count. Malicious remotes can serve a 12-byte pack header claiming 0xFFFFFFFF objects, forcing Tree::with_capacity to reserve hundreds of gigabytes and crash clone or fetch operations.

Gitoxide gix-pack before 0.75.0 contains a denial of service vulnerability that allows malicious remotes to crash fetch operations by sending crafted thin-pack OFS_DELTA entries. Attackers can serve a thin pack with a REF_DELTA referencing an existing object followed by an OFS_DELTA whose base_distance exceeds pack_offset, triggering a panic in LookupRefDeltaObjectsIter that aborts the fetch process.

gitoxide gix-pack before 0.75.0 contains a reachable assertion in BytesToEntriesIter::new_from_header: the pack header decoder accepts version 3, but a release-enabled assert_eq! requires version 2. A malicious remote can send a 12-byte pack header declaring version 3 to panic the client and abort the clone or fetch.

Gitoxide gix-mailmap before 0.35.0 contains a denial of service vulnerability that allows attackers to exhaust CPU through quadratic-time parsing in Snapshot::from_bytes. Attackers can commit a crafted .mailmap with descending-ordered entries forcing repeated front inserts, blocking workers running log, blame, or shortlog for minutes.