2026 · Medium

Medium-severity vulnerabilities disclosed in 2026

Medium-rated CVEs published in 2026, with SEC.co remediation and prioritization guidance.

4010 published vulnerabilities · page 38 of 41

• CVE-2026-42568MEDIUM 4.3

  Yamcs, a mission control framework used in aerospace and satellite operations, contains a flaw in how it validates user login attempts against LDAP directories. An authenticated user can inject specially crafted characters into their username to bypass normal search restrictions, potentially gaining unauthorized access to directory information. The vulnerability requires valid credentials to exploit, limiting immediate risk, but organizations relying on LDAP-backed Yamcs deployments should update promptly.

• CVE-2026-4298MEDIUM 4.3

  The DSGVO All in one for WP WordPress plugin contains an authorization flaw that allows any authenticated user with Subscriber-level access or higher to reset critical privacy policy settings. An attacker could silently revert your organization's customized cookie notices, Google Analytics consent language, and social media policies back to defaults—potentially disrupting privacy compliance and user trust without requiring administrative credentials or special permissions.

• CVE-2026-43708MEDIUM 4.3

  A cross-origin data exfiltration vulnerability in Apple's WebKit rendering engine affects Safari and multiple Apple operating systems. A malicious website can extract user data that should remain isolated to the user's own origin, bypassing the browser's same-origin policy. Apple has patched this issue across Safari, iOS, iPadOS, macOS, tvOS, visionOS, and watchOS with improved input validation to enforce proper origin boundaries.

• CVE-2026-44041MEDIUM 4.3

  UltraVNC versions through 1.8.2.2 contain a flaw in how the software processes text encoding conversion. When converting wide-character strings to multibyte format, the code reads from a caller-supplied buffer without first checking its bounds. If that buffer is missing a proper null terminator, the read operation continues beyond the buffer's legitimate memory, exposing data from adjacent memory regions or potentially crashing the application. This vulnerability requires an unusual programming pattern to trigger and is not currently known to be actively exploited.

• CVE-2026-44169MEDIUM 4.3

  MariaDB's access control for stored routines has a flaw that leaks routine definitions to users who shouldn't see them. If a user receives EXECUTE permission on a stored routine through a database role, they can view the routine's source code without having explicit SHOW CREATE ROUTINE privilege. This bypasses the intended separation between execution rights and visibility rights, allowing unauthorized code inspection. The issue affects MariaDB 11.4, 11.8, and 12.3 releases across a specific version range and has been patched.

• CVE-2026-44731MEDIUM 4.3

  OpenProject's meetings filter feature contains a user enumeration vulnerability that allows attackers with login credentials to discover which user accounts exist in the system and learn their full names. By testing different user IDs and analyzing how the application responds, an attacker can build a complete roster of valid accounts—information that's typically kept private. The vulnerability affects OpenProject versions before 17.3.2 and 17.4.0.

• CVE-2026-44732MEDIUM 4.3

  OpenProject, an open-source project management platform, contains an authorization bypass vulnerability in its document management functionality. An authenticated user can modify or move documents belonging to other projects even if they lack the necessary permissions, by exploiting a timing issue where attribute changes are applied before access controls are checked. The flaw affects versions prior to 17.3.2 and 17.4.0.

• CVE-2026-44750MEDIUM 4.3

  SAP MDG's Review Match Groups Application fails to enforce proper authorization checks for logged-in users. This means a user with basic system access can perform actions they shouldn't be able to, effectively gaining privileges beyond their intended role. The vulnerability is localized to data integrity risks—attackers cannot access confidential information or disrupt system availability, but they can modify data they're not authorized to change.

• CVE-2026-44755MEDIUM 4.3

  SAP Business Objects Business Intelligence Platform contains a vulnerability that allows authenticated users to send spoofed emails by manipulating email parameters that the system fails to properly validate. An attacker with legitimate access to the platform could craft emails that appear to come from different senders, potentially damaging organizational trust or enabling social engineering attacks. The vulnerability is limited to integrity concerns—system confidentiality and availability are not compromised.

• CVE-2026-44779MEDIUM 4.3

  Discourse, a popular open-source discussion platform, contains a flaw in bot debug endpoints that unintentionally exposes sensitive whisper translation audit logs. Whispers are private messages intended only for specific users or moderators, and their translation records should not be accessible. Authenticated users can access these debug endpoints and retrieve information about whisper translations that they shouldn't see, potentially revealing private conversations and moderation activities. The vulnerability affects multiple active release branches and has been patched.

• CVE-2026-44780MEDIUM 4.3

  Discourse has a privilege-bypass vulnerability where category moderators can read the full source of emails submitted to the platform, even if they lack explicit permission to view raw email content. When posts arrive via incoming email integration, a serialization component unintentionally exposes complete email headers, metadata, and body to anyone reviewing flagged posts—sidestepping the normal access controls that restrict raw email viewing. This leaks potentially sensitive information embedded in inbound emails to a wider set of moderators than intended.

• CVE-2026-44782MEDIUM 4.3

  Discourse, an open-source discussion platform, contains a logic error in how it handles user name visibility settings. Due to a misnamed predicate in the GroupPostSerializer component, user names are always included in API responses regardless of the site administrator's configuration to disable them. An authenticated user can exploit this to retrieve user name information that should be hidden, even when the 'enable_names' setting is turned off. The issue affects multiple version branches and has been patched.

• CVE-2026-44785MEDIUM 4.3

  Discourse, an open-source discussion platform, contains an authorization flaw in its AI-powered "explain" feature. When a user asks the system to explain a reply to a post, the feature only checks whether the user can see the reply itself—not whether they should have access to the parent post being discussed. This oversight allows any authenticated user with access to the AI helper to read hidden or restricted parent posts by simply requesting an explanation of a public reply to them. The vulnerability affects multiple version branches and has been patched.

• CVE-2026-44957MEDIUM 4.3

  Revive Adserver versions 6.0.6 and earlier contain a flaw in their XML-RPC API that fails to properly verify user permissions when reassigning entities to different parent entities. An attacker with low-level API access cannot exploit this vulnerability alone; they would need either a separate critical vulnerability (CVE-2026-34917) or a custom third-party extension to reach the vulnerable code. When those conditions align, an attacker could reassign entities without authorization, creating inconsistent ownership relationships that may allow lateral privilege expansion or data manipulation.

• CVE-2026-45264MEDIUM 4.3

  Nextcloud versions spanning 17.0.0 through 21.0.3 contain a permission bypass vulnerability that allows users with read and create access—but explicitly not update access—to rename files within team folders. This unintended capability undermines the granular permission model Nextcloud enforces, potentially enabling unauthorized modification of file metadata and organizational disruption. The issue affects a broad version range and has been patched across all active release lines.

• CVE-2026-45286MEDIUM 4.3

  An authenticated user on a Nextcloud instance can discover other users' identities by abusing the Calendar app's attendee-suggestion feature. The vulnerability exists because this endpoint bypasses the access controls that Nextcloud applies elsewhere. An attacker already logged into the system can systematically enumerate valid usernames, potentially laying groundwork for targeted attacks like password spraying or social engineering. The flaw affects Nextcloud versions 5.5.13 through 5.5.16 and 6.2.0 through 6.2.2.

• CVE-2026-45544MEDIUM 4.3

  Nextcloud Tables, a collaborative document and data management component, contains an information disclosure vulnerability affecting versions 0.8.0 through 1.0.3. The issue allows users with read-only access to view filter criteria—potentially including sensitive column names, field definitions, or search logic—that should have been restricted to higher-privilege users. This represents a low-severity data exposure that could leak operational details about table structure and content filtering strategies. The vulnerability has been resolved in Nextcloud Tables versions 1.0.4 and 2.0.0.

• CVE-2026-45563MEDIUM 4.3

  Roxy-WI, a popular web management interface for load balancers and web servers, contains a flaw that allows any logged-in user to view detailed audit trails of other users' administrative actions. Even a guest-level user in one department can see which servers another user has accessed, what configuration changes they deployed, and what services they restarted. The vulnerability affects Roxy-WI versions 8.2.6.4 and earlier. While this doesn't grant direct control over infrastructure, it exposes sensitive operational history that should remain confidential.

• CVE-2026-45650MEDIUM 4.3

  CVE-2026-45650 is a user interface spoofing vulnerability in Microsoft Bing that allows attackers to misrepresent critical information to users over the network. An attacker can craft a malicious link or interaction that tricks Bing's UI into displaying false or misleading content, potentially leading users to believe they are viewing legitimate search results, advertisements, or information when they are not. This requires user interaction to succeed, meaning a victim must click a link or engage with the spoofed element.

• CVE-2026-45729MEDIUM 4.3

  ThorVG, a vector graphics rendering engine, contains a flaw that can crash applications using it when they process malicious SVG files. An attacker can craft a specially formatted SVG document as small as 6 bytes that triggers an application crash when the affected code path is invoked. This is a denial-of-service vulnerability affecting availability but not data confidentiality or integrity. The issue was introduced in earlier versions and is resolved in ThorVG 1.0.5.

• CVE-2026-45776MEDIUM 4.3

  OpenXDMoD is an open-source framework used by HPC (high-performance computing) centers to collect and monitor system performance metrics. Versions before 11.0.3 contain a session-handling flaw that allows an authenticated attacker to manipulate authorization checks. If an installation includes the optional Job Performance (SUPReMM) module, an attacker could view other users' job efficiency data they shouldn't have access to. The vulnerability requires an existing login but does not require admin privileges.

• CVE-2026-46542MEDIUM 4.3

  Nimiq, a Rust-based blockchain implementation, contains a vulnerability in how it validates cryptographic keys used for multisig (multi-signature) operations. When the system processes Ed25519 public keys, it only checks that the key is 32 bytes long—the correct size—but doesn't verify that those bytes actually represent a valid point on the Ed25519 elliptic curve. An attacker can send a specially crafted, mathematically invalid public key that passes the length check. When the Nimiq node attempts to process this malformed key during multisig delinearization, the code crashes due to an uncaught exception, causing a denial-of-service condition. This affects Nimiq versions before 1.4.0.

• CVE-2026-46548MEDIUM 4.3

  NocoDB, a database-as-spreadsheet platform, contains a server-side request forgery (SSRF) vulnerability in its webhook notification system. An authenticated user with permissions to create webhooks can craft malicious requests to the Slack, Discord, Mattermost, or Teams webhook plugins that bypass intended network protections, allowing them to reach internal hosts that should be off-limits. The flaw stems from improper configuration of HTTP agent settings in four webhook integrations. This issue is resolved in version 2026.04.1 and later.

• CVE-2026-46605MEDIUM 4.3

  Apache ActiveMQ has an authorization flaw that allows authenticated users to delete message queues and topics they shouldn't be able to modify. An attacker with valid credentials to your messaging system could disrupt operations by removing critical destinations, even if permission controls suggest they shouldn't have that ability. This affects ActiveMQ versions before 5.19.7 and 6.0.0 through 6.2.5.

• CVE-2026-46645MEDIUM 4.3

  SQLAdmin, a popular admin interface library for SQLAlchemy, contains an access control bypass vulnerability in its ajax_lookup feature. When developers restrict access to certain database models using the library's access control mechanism, authenticated users can circumvent those restrictions by sending requests directly to the ajax_lookup endpoint. This allows them to retrieve data from models they shouldn't have permission to access. The vulnerability was fixed in version 0.25.1.

• CVE-2026-46700MEDIUM 4.3

  Actual, a personal finance management tool, contains an access control flaw in its sync-server component prior to version 26.6.0. The GET endpoint used to retrieve secrets—such as bank integration credentials—only checks that a user has a valid session, but does not verify admin privileges. This means any authenticated non-admin user in OpenID-enabled multi-user deployments can query the secrets store and discover which external banking services (SimpleFin, Pluggy, GoCardless) have been configured by administrators, potentially exposing the names and availability of sensitive integrations. The flaw is limited to information disclosure; attackers cannot modify or delete secrets directly through this endpoint.

• CVE-2026-46747MEDIUM 4.3

  SINEC INS, Siemens' network security appliance, contains a path traversal vulnerability in its file upload API endpoint. An authenticated user can craft malicious directory paths to access files outside their intended scope on the server. The vulnerability affects all versions prior to V1.0 SP2 Update 6 and requires valid login credentials to exploit, limiting but not eliminating risk in environments where account compromise or insider threats are concerns.

• CVE-2026-46764MEDIUM 4.3

  Apache Airflow contains an authorization bypass flaw in its audit-log API endpoints. An authenticated user with read access to audit logs for one workflow (Dag) can bypass per-Dag scoping restrictions and view audit-log entries from any other Dag in the same Airflow deployment by directly requesting specific event log IDs. The vulnerability stems from inconsistent permission enforcement: the collection endpoint properly restricts results by Dag, but the detail endpoint applies only a generic audit-log permission check without verifying the requester has access to the specific Dag whose logs are being retrieved. This allows low-privileged users to enumerate and read sensitive audit trails across Dags they should not be able to access.

• CVE-2026-47224MEDIUM 4.3

  NanaZip, a Windows-native fork of the popular 7-Zip archiver, contains a memory safety flaw in how it processes LVM (Logical Volume Manager) disk images. When a user opens a specially crafted LVM disk image file, the application can read beyond the boundaries of allocated memory. This weakness could cause NanaZip to crash or, in theory, leak sensitive data from memory. The issue affects NanaZip versions 3.0.1000.0 through 6.0.1697.0. Users should upgrade to version 6.0.1698.0 or later to resolve the issue.

• CVE-2026-47236MEDIUM 4.3

  Solidtime, an open-source time-tracking application, contains an authorization bypass vulnerability in its team management interface. Prior to version 0.12.2, the web page that displays team members and pending invitations fails to properly enforce permission checks, allowing any employee in the organization to view sensitive information—including pending invitation email addresses and member lists—even though the same data is correctly restricted in the official API. The vulnerability stems from incomplete permission validation in the Jetstream page handler, which checks only basic team membership rather than the required invitations:view and members:view permissions.

• CVE-2026-47263MEDIUM 4.3

  Discourse versions released between early 2026 and mid-2026 contain an access control flaw in their webhook event redelivery system. When Discourse attempts to resend failed webhook notifications, it broadcasts internal event data to a world-readable channel without properly restricting who can see it. Any logged-in user—or any visitor on instances without login requirements—can discover and read webhook events by guessing sequential webhook IDs. This exposes sensitive data that webhook endpoints are designed to receive privately, such as user activity, content moderation events, or custom application logic tied to your Discourse instance.

• CVE-2026-47675MEDIUM 4.3

  Hono, a JavaScript web framework, contains a flaw in how it sanitizes cookie options. While the framework validates certain cookie parameters (domain and path) to prevent malicious characters from breaking the Set-Cookie header, it fails to apply the same checks to sameSite and priority options. If an application passes user-controlled input directly into these parameters, an attacker could inject additional cookie attributes into the response header, potentially manipulating cookie behavior or setting unintended security policies.

• CVE-2026-47696MEDIUM 4.3

  WWBN AVideo, an open-source video hosting platform, contains a payment processing vulnerability in versions 29.0 and earlier. When both the AuthorizeNet and YPTWallet plugins are active, any logged-in user can artificially inflate their account wallet balance without actually paying. The vulnerable endpoint accepts a user-supplied amount parameter and immediately credits the wallet without verifying that a real payment transaction occurred through Authorize.Net. This is a financial manipulation flaw that bypasses payment authentication entirely.

• CVE-2026-47991MEDIUM 4.3

  Adobe Experience Manager contains a flaw that allows attackers to craft deceptive URLs that redirect users to attacker-controlled websites. If a victim clicks such a link, they may be taken to a fake login page or other malicious site where their credentials could be stolen, leading to account compromise. This vulnerability affects multiple AEM versions and requires user interaction—the attacker must convince someone to click the malicious link.

• CVE-2026-48092MEDIUM 4.3

  7-Zip versions 9.34 through 26.00 contain a flaw in how they process SquashFS archive files that can leak sensitive data from memory when extracting files. The vulnerability exists only in 32-bit builds of 7-Zip and requires an attacker to craft a malicious archive with specially modified metadata. When a user extracts such an archive, heap memory contents that should remain private are instead written into the extracted file, potentially exposing passwords, encryption keys, or other sensitive information stored in memory. The issue stems from integer arithmetic wrapping that bypasses safety checks. Users on 64-bit systems are not affected.

• CVE-2026-48103MEDIUM 4.3

  7-Zip versions 9.34 through 26.00 contain a memory read vulnerability in the WIM (Windows Imaging Format) archive handler. When processing specially crafted WIM files, the software reads a small amount of data just beyond an allocated memory region due to an off-by-one error in bounds checking. This occurs automatically in the file manager when listing directory contents, requiring only that a user open or preview a malicious WIM file. The practical impact is limited to crashes or potential minor leaks of adjacent memory; no file corruption or system compromise is possible through this flaw alone.

• CVE-2026-48111MEDIUM 4.3

  7-Zip versions 9.21 through 26.00 contain a boundary-checking flaw in their UEFI firmware image parser. When processing certain archive sections, the parser uses an incorrect comparison operator that allows a malicious opcode value to read data beyond an array's bounds. This can either crash the application when the out-of-bounds memory is invalid, or leak small amounts of adjacent string data into the archive's metadata. The flaw is triggered automatically when opening a specially crafted archive file, but the leaked information is limited and does not expose sensitive data or memory layout information.

• CVE-2026-48518MEDIUM 4.3

  MultiJuicer, a platform for running isolated Juice Shop instances on Kubernetes clusters, contains a cross-site request forgery (CSRF) vulnerability in its team join endpoint. Versions 8.0.0 through 10.0.0 accept requests regardless of content type, allowing attackers to craft web pages that silently add victims to attacker-controlled teams. When a victim visits an attacker's page, their browser automatically submits a form to the vulnerable endpoint, enrolling them in the attacker's team without their knowledge. This causes victims to unknowingly solve security challenges under the attacker's identity, inflating scores and exposing any sensitive data entered during exercises. The attack requires no prior authentication and bypasses standard CORS protections and SameSite cookie policies.

• CVE-2026-48789MEDIUM 4.3

  AnythingLLM versions prior to 1.13.0 contain a path traversal vulnerability on Windows systems that allows authenticated users to list and potentially access documents outside the intended documents directory. The vulnerability stems from incomplete validation of Windows-style directory paths. An attacker with valid credentials could craft specially encoded paths to escape the documents folder sandbox and view file listings in other locations on the Windows system.

• CVE-2026-48810MEDIUM 4.3

  FreeScout, a free help desk platform built on Laravel, contains an authorization flaw in version 1.8.220 and earlier. A user with conversation editing permissions who authored a message in one mailbox can edit that message's content even after an administrator removes them from that mailbox. The vulnerability exploits a gap in access controls: the system verifies the user created the message and has the global edit permission, but fails to confirm the user still belongs to the mailbox where the conversation lives. This allows former mailbox members to alter thread history and potentially mislead team members or customers.

• CVE-2026-48811MEDIUM 4.3

  FreeScout, an open-source helpdesk and shared inbox platform, contains a flaw that allows former team members to permanently delete internal notes—even after their access to the mailbox has been revoked. A non-admin user who previously created private threads in a conversation can return and destroy those notes without authorization, because the system fails to verify whether the user still belongs to the mailbox. This affects FreeScout versions before 1.8.221.

• CVE-2026-4888MEDIUM 4.3

  Everest Forms, a popular WordPress form-building plugin, contains a security flaw that allows low-privilege logged-in users to send emails from your website to anyone they choose. Any user with Subscriber access or higher can exploit this by calling an internal email-testing function without proper permission checks. This doesn't require clicking malicious links or advanced technical skills—just authenticated access to your WordPress admin panel.

• CVE-2026-48891MEDIUM 4.3

  Apache Airflow contains an information disclosure vulnerability in its web UI's dependency graph viewer. Authenticated users with permission to read certain workflows can inadvertently discover the names and identifiers of other workflows they should not have access to by examining trigger and sensor dependency relationships displayed in the graph. This happens because the filtering logic that restricts which workflows a user can see was incompletely applied—it hides the top-level workflow information but still leaks workflow identifiers in the detailed dependency connections. Organizations relying on workflow-level access controls to keep workflow names confidential across teams are at risk.

• CVE-2026-48934MEDIUM 4.3

  Node.js contains a flaw in how it validates TLS certificates during secure connections. An authenticated attacker could potentially bypass certificate verification, allowing them to intercept or impersonate HTTPS traffic in environments where they have network or application access. This is not a remote unauthenticated attack—it requires the attacker to already have some level of access to the system or network.

• CVE-2026-49140MEDIUM 4.3

  Nanobot versions before 0.2.1 have a denial-of-service flaw in how they handle media downloads from Matrix chat rooms. An authenticated user in a room can deliberately send specially crafted media events with missing or wrong size information, causing the system to download large files without properly checking their declared sizes first. By sending many of these malicious requests at once, an attacker can force the Nanobot process to consume excessive memory and bandwidth until the service becomes slow or unresponsive. The attacker must already be a member of the room to exploit this.

• CVE-2026-49288MEDIUM 4.3

  Statamic, a Laravel-based content management system, contained an authorization flaw that allowed authenticated users with Control Panel access to view content and metadata they weren't supposed to see. An attacker with valid login credentials could browse restricted entries, assets, user profiles, roles, and groups—exposing titles, custom field values, entry content, asset metadata, and the mere existence of sensitive organizational structures. The vulnerability is read-only; attackers could not modify data. Fixed in Statamic 5.73.23 and 6.20.0.

• CVE-2026-49322MEDIUM 4.3

  The 2025 Indian Motorcycle Scout Bobber + Tech model contains a flaw in its wireless control system that allows someone with access to the motorcycle's internal network to steal the owner's PIN unlock code by observing just a single authentication attempt. Instead of using proper cryptographic security, the system performs simple mathematical operations that can be reversed to recover the PIN, completely bypassing the bike's primary security lock.

• CVE-2026-49323MEDIUM 4.3

  The 2025 Indian Motorcycle Scout Bobber + Tech model contains a flaw in how its wireless control module authenticates with the engine control module. An attacker positioned on the vehicle's internal network can intercept a single authentication exchange and reverse-engineer the motorcycle's immobilizer secret—the cryptographic key that prevents unauthorized engine starts. Once recovered, the attacker can bypass the immobilizer entirely and start the engine without the key fob.

• CVE-2026-49337MEDIUM 4.3

  libde265, an open-source H.265 video decoder library, contains a memory management flaw that allows attackers to trigger unbounded memory growth on a victim's system. By sending a specially crafted sequence of video data packets, an attacker can cause the decoder to accumulate slice header data in memory without ever releasing it. This happens during normal video playback and can eventually exhaust available memory (denial of service). The vulnerability requires user interaction—someone must open or play a malicious video file—but does not require any special privileges or network access beyond the ability to deliver the crafted video.

• CVE-2026-49355MEDIUM 4.3

  OpenProject versions before 17.4.0 contain an information disclosure vulnerability in the meeting agenda API endpoint. An authenticated attacker can view private work package details that are linked to meeting agenda items, even when those work packages belong to projects the attacker cannot normally access. This leaks sensitive project information to users who should not have visibility into it.

• CVE-2026-49369MEDIUM 4.3

  JetBrains YouTrack versions before 2026.1.13162 contained a flaw that allowed authenticated users to access sensitive information about other users and groups they shouldn't be able to see. The vulnerability is limited to the Users and Groups administrative pages and requires valid login credentials to exploit. This is a straightforward authorization issue where the application failed to properly restrict who could view certain user and group data.

• CVE-2026-49377MEDIUM 4.3

  JetBrains TeamCity contains a configuration flaw where default agent parameters inadvertently expose sensitive data to authenticated users. An attacker with valid login credentials can access information through TeamCity's agent configuration that should remain restricted. This is a network-accessible issue affecting TeamCity deployments before version 2025.11.2, though the vulnerability requires prior authentication to exploit.

• CVE-2026-49378MEDIUM 4.3

  JetBrains TeamCity contained a vulnerability where stored credentials could be inadvertently exposed through the parameter autocompletion feature. When users typed in parameter fields, the system would suggest previously stored credential values, potentially revealing sensitive authentication data to anyone with access to the TeamCity interface. This issue affects TeamCity versions prior to 2026.1 and requires an authenticated user to interact with the affected feature. The exposure is limited to local disclosure within the TeamCity environment rather than remote exfiltration.

• CVE-2026-49482MEDIUM 4.3

  ClipBucket v5 has a vulnerability in its subtitle editing feature that allows authenticated users to corrupt subtitle data across multiple videos at once. By sending a specially crafted request with a wildcard character, an attacker with valid credentials can overwrite all subtitle titles for any video they control in a single action. This is a data integrity issue rather than a data exposure risk, and affects versions before 5.5.3 - #141.

• CVE-2026-49848MEDIUM 4.3

  FreeSWITCH versions prior to 1.11.1 contain a flaw in the Verto module's authentication handler that allows user-supplied session variables to be written into connection state before password validation occurs. Because these writes are append-only and failed authentication attempts don't close the WebSocket connection, variables from a bad login attempt persist and carry over into a subsequent successful login on the same connection. This creates a window for session state manipulation through repeated authentication attempts.

• CVE-2026-50569MEDIUM 4.3

  Fission, an open-source Kubernetes serverless framework, contains a validation bypass in its HTTP trigger configuration. Prior to version 1.25.0, two URL-related fields—RelativeURL and Prefix—were not properly validated when HTTPTrigger resources were created directly through kubectl or the Kubernetes REST API. While these fields were validated at the CLI level, the validation logic was absent from the core application and its API server rules. This means an attacker with Kubernetes API access could bypass URL restrictions by creating or modifying HTTPTrigger resources directly, potentially routing traffic to unintended functions or exposing restricted endpoints.

• CVE-2026-50739MEDIUM 4.3

  A flaw in Revive Adserver allows low-privileged users to associate their tracking tools with advertising campaigns owned by other managers on the same server. This happens because a security fix applied to one direction of the linking process was not extended to the reverse operation, creating an inconsistency that attackers can exploit. The vulnerability affects Revive Adserver version 6.0.7 and earlier.

• CVE-2026-50744MEDIUM 4.3

  Revive Adserver 6.0.7 contains a flaw in its XML-RPC authentication system. When an administrator login attempt is made via the ox.login API method, the server returns an error message—as intended—but inadvertently creates a valid session ID and sends it back in the HTTP response headers anyway. An attacker with basic API access can intercept this session ID and use it to bypass the admin-only restriction, gaining unrestricted access to sensitive API functions without needing actual admin credentials.

• CVE-2026-5137MEDIUM 4.3

  The RTMKit plugin for WordPress (used to manage themes for Elementor) has a vulnerability that allows certain WordPress users to load and run unauthorized PHP code on a website. Specifically, a user with Contributor access or higher can manipulate a request to the plugin's template rendering feature to include files they shouldn't be able to access, then execute whatever code is inside those files. This requires an attacker to already have a low-level account on the WordPress site.

• CVE-2026-5138MEDIUM 4.3

  Foreman, Red Hat's infrastructure management platform, contains a flaw that allows authenticated users with basic host-editing permissions to bypass authorization controls and view sensitive network configuration data from other organizations and locations. An attacker could extract subnet layouts, IP address ranges, gateway configurations, DNS server details, and VLAN assignments belonging to infrastructure they should not have access to. The vulnerability requires valid credentials and existing permissions to exploit, limiting its immediate blast radius, but it creates a significant cross-tenant data leakage risk in multi-tenant deployments.

• CVE-2026-52795MEDIUM 4.3

  Gogs, a self-hosted Git service, contains a logic error in its Watch API that allows any authenticated user to monitor private repositories they should not have access to. The vulnerability stems from an inverted access check—the API returns an error when a user CAN read the repository, rather than when they CANNOT. Once a user watches a private repository, their dashboard reveals sensitive information including commit messages, branch names, issue titles, and pull request details. If email notifications are enabled, attackers also receive email digests containing issue and comment content from repositories they should be excluded from.

• CVE-2026-53422MEDIUM 4.3

  A flaw in Erlang OTP's SFTP server module allows authenticated users to discover whether files and directories exist outside their authorized access area. When a user sends a specially crafted request to the server's path-resolution function, the server's response reveals whether a path exists on the system—even if the user shouldn't have permission to see that location. An attacker can use this to map out sensitive filesystem structures, such as identifying the presence of critical files or system directories, without actually reading any contents. This is an information-disclosure issue that requires valid SFTP credentials to exploit.

• CVE-2026-53436MEDIUM 4.3

  Jenkins contains a validation flaw in its login redirect mechanism that allows attackers to craft phishing URLs appearing to come from a legitimate Jenkins instance. When users log in, Jenkins is supposed to redirect them to internal pages, but the vulnerability allows attackers to redirect users to external malicious sites by exploiting how the application handles relative path segments (like `./` or `../`). An attacker would need to trick a user into clicking a specially crafted link, but the exploit itself is straightforward and doesn't require special technical skills.

• CVE-2026-53437MEDIUM 4.3

  Jenkins versions 2.567 and earlier (LTS 2.555.2 and earlier) contain a flaw in how they validate redirect URLs after user login. An attacker can craft a malicious redirect URL that appears to point to a legitimate Jenkins instance by inserting tab or newline characters between the `//` protocol separator, causing the validation to pass. When a user clicks such a link after logging in, they may be redirected to an attacker-controlled site while believing they're staying within Jenkins, enabling credential harvesting or other phishing attacks.

• CVE-2026-53438MEDIUM 4.3

  Jenkins versions 2.567 and earlier (or LTS 2.555.2 and earlier) contain a permission bypass flaw that allows authenticated users holding the Item/Cancel permission to cancel build queue items without requiring Item/Read permission. This means an attacker with limited cancellation rights can disrupt builds they shouldn't be able to view or access, effectively using one permission to circumvent another. The vulnerability is not actively exploited in the wild and requires authenticated access, making it a moderate risk in most deployments.

• CVE-2026-53439MEDIUM 4.3

  Jenkins versions 2.567 and earlier (LTS 2.555.2 and earlier) contain a permission bypass vulnerability that allows low-privileged attackers to discover sensitive information about other users. Specifically, attackers who have been granted the basic Overall/Read permission can view other users' timezone settings and enumerate the names of views in other users' private "My Views" sections. This is an information disclosure issue that could support reconnaissance or social engineering attacks, though it does not enable direct system compromise.

• CVE-2026-53440MEDIUM 4.3

  Jenkins versions 2.567 and earlier (LTS 2.555.2 and earlier) contain a flaw in their "Delegate to servlet container" security realm that fails to validate redirect destinations after user login. An attacker can craft a malicious link that redirects authenticated users to an attacker-controlled website, enabling phishing attacks that steal credentials or distribute malware while appearing to come from a legitimate Jenkins instance.

• CVE-2026-53463MEDIUM 4.3

  ImageMagick, a widely-used open-source image editing library, has a vulnerability in its distort operation that can cause the application to crash when given malformed parameters. The issue stems from improper handling of null pointers—a memory safety problem that occurs before the software attempts to process the distort command. Users who interact with untrusted images or accept image files from external sources face potential denial-of-service risk, though the vulnerability does not compromise data confidentiality or integrity.

• CVE-2026-53634MEDIUM 4.3

  Sharp, a Laravel-based content management framework, contains an authorization bypass affecting its Quick Creation Command feature in versions 9.0.0 through 9.22.2. An authenticated user without permission to create records in a specific entity could still use the Quick Creation feature to view creation forms and submit new records, provided that entity had a Quick Creation Command handler configured. This represents a privilege escalation flaw where the access control layer was not properly enforced on certain endpoints. The vulnerability has been resolved in version 9.22.3.

• CVE-2026-53675MEDIUM 4.3

  BuddyPress 14.4.0 has a flaw in its friends REST API that allows any logged-in user to view another user's complete friend list without permission. The vulnerability exists because the API endpoint checks only that someone is authenticated, not whether they should have access to the specific friend list being requested. An attacker with any user account can enumerate and collect the private social connections of any other user on the platform.

• CVE-2026-53736MEDIUM 4.3

  Easy Twitter Feeds versions before 1.2.13 have a cross-site request forgery (CSRF) vulnerability in their post duplication feature. An attacker can craft a malicious link that, when clicked by a logged-in user, will automatically duplicate posts without the user's knowledge or consent. The vulnerability exists because the plugin fails to validate that duplication requests actually came from the user, making it trivial for attackers to abuse the feature via social engineering or injected links.

• CVE-2026-53739MEDIUM 4.3

  Yoast Duplicate Post, a popular WordPress plugin for content management, contains a cross-site request forgery (CSRF) flaw in versions through 4.6. The vulnerability allows an attacker to craft a malicious webpage or email that, when visited by a logged-in WordPress administrator, automatically suppresses admin notices across the entire network without the admin's knowledge or consent. No special technical skill is required to exploit this—only the ability to trick an admin into clicking a link or visiting a page.

• CVE-2026-53781MEDIUM 4.3

  Summarize, a podcast and media management CLI tool, contains a vulnerability that allows attackers to exhaust a system's disk space. If you use Summarize before version 0.17.0, an attacker who controls a podcast feed or media URL can force the application to download an extremely large (or infinite) file to your disk. The vulnerability exists because Summarize doesn't properly validate file sizes when the server doesn't send proper size headers or uses certain types of data streaming. This can render your system unusable by filling up its storage.

• CVE-2026-53826MEDIUM 4.3

  OpenClaw versions prior to 2026.4.26 inadvertently leak sensitive workspace path information when child AI model sessions are spawned from sandboxed parent environments. An authenticated attacker can trigger this disclosure by creating child sessions, allowing them to discover the real underlying workspace location or related memory context that should remain isolated. The vulnerability requires valid credentials and does not involve network-level attacks or severe data manipulation, but it does undermine the security boundary that sandboxing is designed to maintain.

• CVE-2026-53835MEDIUM 4.3

  OpenClaw versions before 2026.5.6 have a flaw in how they enforce configuration rules for Feishu dynamic-agent bindings. An authenticated user can create or modify these bindings while bypassing the normal access controls that should restrict who can make such changes. This means someone with legitimate credentials could potentially create unauthorized connections between senders and agents that policy should have blocked.

• CVE-2026-53845MEDIUM 4.3

  OpenClaw versions before 2026.5.6 contain a vulnerability that allows authenticated users to bypass security hooks designed to audit and enforce policies on skill command execution. When skill commands are routed through a specific dispatch path, they skip the before-tool-call hooks that normally intercept and validate these operations. An attacker with valid credentials could exploit this to execute commands without triggering audit logs or policy checks.

• CVE-2026-53848MEDIUM 4.3

  OpenClaw before version 2026.5.26 has a security flaw that allows authenticated operators to bypass command allowlist restrictions. An attacker with valid operator credentials can craft specially formed requests that use transparent command wrappers to execute operations that should have been blocked by the allowlist. This affects the integrity of command execution controls but does not expose data or cause system availability issues.

• CVE-2026-53867MEDIUM 4.3

  Capgo versions prior to 12.128.2 do not properly clean up user profile images from storage when those images are replaced or deleted. This means old image files remain accessible on the backend server even after users think they've removed them. An attacker with a user account can leverage previously shared or discovered image URLs to retrieve these orphaned files, potentially accessing private user photos or other sensitive image data that should have been deleted.

• CVE-2026-53900MEDIUM 4.3

  Firefox for iOS was inadvertently preserving cookies from initial PDF requests even when those requests were redirected across different websites. An attacker could craft a malicious site that, when visited by a user, exploits this behavior to inject cookies into legitimate requests sent to an unrelated target domain—potentially allowing credential theft or session hijacking without the user's knowledge. Apple users running Firefox for iOS below version 152.0 are affected. Mozilla has patched this in Firefox for iOS 152.0.

• CVE-2026-53908MEDIUM 4.3

  MCO (MyComplianceOffice) contains a user enumeration vulnerability in its authentication workflows. When users attempt to reset passwords or retrieve usernames, the application responds differently depending on whether an account exists. An attacker with login access can exploit these timing or content differences to systematically discover valid usernames and associated email addresses—useful for follow-up social engineering, credential stuffing, or targeted account takeover attempts. The vendor has not been successfully contacted to confirm scope beyond version 25.3.3.1.

• CVE-2026-54006MEDIUM 4.3

  Open WebUI, a self-hosted AI platform, contains an authorization bypass vulnerability in its calendar event management API. A regular user can create an event in their own calendar and then move it into another user's calendar without proper permission checks. While the initial event creation correctly validates authorization, the move operation skips this validation entirely. This allows any authenticated user to access or modify calendar events belonging to other users by knowing their calendar ID—a common identifier in multi-user deployments.

• CVE-2026-54014MEDIUM 4.3

  Open WebUI, a self-hosted AI platform designed for offline operation, contains a path traversal flaw that allows authenticated users to read files outside the intended cache directory. The vulnerability stems from an incomplete validation check in the file serving logic that fails to properly isolate the cache folder from similarly-named sibling directories. An attacker with valid credentials can exploit directory name confusion to access restricted files. The issue has been patched in version 0.9.6.

• CVE-2026-54016MEDIUM 4.3

  Open WebUI versions before 0.9.6 contain an authorization bypass in the search_knowledge_files function. When a model lacks attached knowledge bases but function calling is enabled, an authenticated user can query file metadata from any knowledge base by guessing or knowing its identifier, even if they should not have access. The vulnerability leaks file names and structural information about private or restricted knowledge bases, but does not allow file content extraction or modification.

• CVE-2026-54259MEDIUM 4.3

  Wagtail, a Django-based content management system, contains an information disclosure vulnerability in its Documents and Images chooser feature. When an administrator or staff member accesses these chooser interfaces, the system incorrectly reveals filenames, names, and URLs of documents and images that the user should not have permission to browse. This happens despite the user lacking explicit 'choose' permissions for those collections. The vulnerability requires an attacker to already have Wagtail admin access—ordinary website visitors cannot exploit it. Torchbox has patched this issue across three supported version branches.

• CVE-2026-54260MEDIUM 4.3

  Wagtail, a Django-based content management system, contains a vulnerability in versions before 7.0.8, 7.3.3, and 7.4.2 that allows authenticated admin users to degrade system performance by uploading or processing images with specially crafted filter specifications. The vulnerability requires admin-level access to the Wagtail backend, so it cannot be exploited by regular website visitors. When triggered, the malicious filter specs force the system to perform expensive image rendition processing operations that consume significant server resources, potentially slowing or disrupting the CMS for all users.

• CVE-2026-54262MEDIUM 4.3

  Wagtail, a Django-based content management system, has a permission bypass vulnerability affecting versions before 7.0.8, 7.3.3, and 7.4.2. Users with the basic "Can submit translation" permission can circumvent access controls to create translations for any page in the system, regardless of whether they have permission to modify that content. This allows unauthorized users to introduce translations into restricted content areas. The vulnerability has been patched in the three specified releases.

• CVE-2026-54686MEDIUM 4.3

  Warp, a terminal-based development environment with AI capabilities, contains a vulnerability in how it processes shell integration commands. Between April 2021 and May 2026, Warp accepted certain terminal state-changing instructions from the PTY stream without confirming they originated from legitimate shell integration. An attacker could exploit this by crafting malicious terminal output that, when viewed by a victim in Warp, tricks the application into accepting false metadata—such as spoofed current working directories or falsified SSH session information. The vulnerability requires user interaction (viewing attacker-controlled terminal content) and has a relatively low severity impact focused on integrity rather than confidentiality or availability. It is resolved in version 0.2026.05.06.15.42.stable_01 and later.

• CVE-2026-54886MEDIUM 4.3

  An authenticated user on an SFTP server powered by Erlang/OTP can send specially crafted network messages that cause the SFTP channel handler to enter an infinite loop. The vulnerable code path is triggered only when the server receives SSH extended data (a message type that should never appear in normal SFTP operations) with specific properties. Once triggered, that channel stops responding and consumes CPU continuously, though the overall server remains operational. An attacker with valid SFTP credentials can open many such channels to degrade performance across the system. No data theft or modification is possible; the impact is purely denial of service.

• CVE-2026-55517MEDIUM 4.3

  Deno, a modern JavaScript/TypeScript runtime, has a flaw in how it processes WebSocket connection responses. When a remote server sends back specially crafted response headers containing non-ASCII bytes, Deno crashes entirely rather than handling the malformed data gracefully. An attacker controlling a WebSocket server could exploit this to deny service to any Deno client that connects to it. The issue affects Deno versions before 2.7.5 and is resolved in that release.

• CVE-2026-55542MEDIUM 4.3

  Snipe-IT, an IT asset and license management platform, contains an authorization bypass in its signature image retrieval feature when deployed with S3 object storage. Authenticated users who can guess or discover a signature filename can obtain a temporary signed S3 URL without proper permission checks. This allows unauthorized access to signature images for 5 minutes. The vulnerability stems from the S3 code path returning a URL before performing the same authorization checks that protect local file storage. Version 8.6.1 addresses this issue.

• CVE-2026-55653MEDIUM 4.3

  A vulnerability exists in OpenSSH where a malicious SSH server can trigger a double-free memory error in the client during Diffie-Hellman Group Exchange (DH-GEX) parameter validation in FIPS mode. When an SSH client connects to a compromised or attacker-controlled server, the server can send specially crafted DH-GEX parameters that cause the client process to crash, resulting in a denial of service. This affects users and automated systems that initiate SSH connections to untrusted or compromised servers.

• CVE-2026-55838MEDIUM 4.3

  RustFS, a distributed storage system written in Rust, contains an authorization bypass in its metrics endpoint. The /rustfs/admin/v3/metrics endpoint fails to enforce admin-level IAM policy checks, allowing any authenticated user—even those with minimal permissions like bucket-only access—to view sensitive operational metrics about the entire cluster. This is a privilege escalation vulnerability; the system trusts that a user is authorized to access cluster-wide telemetry data without verifying their IAM role.

• CVE-2026-55873MEDIUM 4.3

  SeaweedFS versions 4.08 through 4.33 contain an authorization flaw in their S3Tables integration. When requests use AWS SigV4 signing for the S3Tables service, the system incorrectly maps low-privileged user identities to a shared administrator account instead of enforcing proper access controls. This misconfiguration allows authenticated users with basic S3 permissions to discover confidential information—specifically, the names and Amazon Resource Names (ARNs) of table buckets managed by administrators. The vulnerability requires valid AWS credentials to exploit and is limited to information disclosure; no data modification or system disruption is possible. SeaweedFS 4.34 and later versions resolve this issue.

• CVE-2026-56217MEDIUM 4.3

  Capgo versions before 12.128.2 contain a vulnerability that lets attackers with API access weaken the security of app updates. Specifically, attackers can bypass organization policies that require app updates to be encrypted by directly manipulating the database to remove encryption keys, allowing them to force apps to accept unencrypted updates. This is a policy-enforcement gap rather than a fundamental cryptographic break, but it does undermine intentional security controls that teams put in place.

• CVE-2026-56255MEDIUM 4.3

  Capgo versions before 12.128.2 contain a denial of service flaw in the demo application creation endpoint. An authenticated user with organization write permissions can abuse this endpoint to repeatedly create demo applications without any rate limiting. Each request triggers approximately 138 database operations, which can degrade system performance, inflate operational costs, and potentially cause service instability for all users.

• CVE-2026-56298MEDIUM 4.3

  Capgo versions before 12.128.2 do not remove EXIF metadata from images that users upload through the app information endpoint. EXIF data embedded in photos can reveal precise geolocation coordinates, timestamps, device information, and other sensitive details. An authenticated attacker could upload specially crafted images to extract this embedded metadata, potentially learning the physical locations where images were taken or other operational details about the target environment.

• CVE-2026-56307MEDIUM 4.3

  Cap-go versions before 12.128.12 contain a pagination flaw in the device-management API endpoint that allows authenticated users with read access to devices to encounter broken pagination loops. When requesting lists of devices, the cursor mechanism fails to advance properly, causing the same pages of results to repeat indefinitely while later data becomes unreachable. This can disrupt workflows that depend on iterating through device inventories and may cause duplicate processing of device records.

• CVE-2026-56310MEDIUM 4.3

  Cap-go before version 12.128.2 has a flaw in how it enforces API key permissions. Normally, an organization-limited API key should only access data from that specific organization. However, this vulnerability allows someone holding such a restricted key to read member information from organizations they should not have access to. The exposed data includes usernames, email addresses, profile images, roles, and temporary account flags. While the vulnerability requires valid API credentials to exploit, it represents a meaningful breach of the intended access control boundaries.

• CVE-2026-56319MEDIUM 4.3

  Capgo versions before 12.128.2 leak information that allows authenticated users with limited API access to discover app IDs they shouldn't know about. By observing how the server responds to requests, an attacker can figure out which other apps exist in the system—even apps they have no permission to access. This breaks the isolation between different customer accounts and their data.

• CVE-2026-56333MEDIUM 4.3

  Capgo versions prior to 12.128.2 contain a vulnerability that allows authenticated organization administrators to bypass security validation checks and set invalid security policies. An authenticated attacker with admin privileges can directly modify the organization database table to circumvent field-level validation, potentially setting insecure values for critical parameters like API key expiration periods. This is an internal privilege abuse risk rather than a remote, unauthenticated attack vector.