🚨 CVE-2026-61560
`@zereight/mcp-gitlab` is a Model Context Protocol server for GitLab. Prior to version 2.1.27, the SSE transport mode (`SSE=true`) exposes all MCP tools without any authentication. The `upload_markdown` tool reads arbitrary files from the server's local filesystem via an unsanitized `file_path` parameter and uploads them to a GitLab project. Combined, any unauthenticated network-reachable attacker can read `/proc/self/environ` to steal the server's `GITLAB_PERSONAL_ACCESS_TOKEN` and achieve full GitLab account takeover. This is the default configuration for Docker deployments. Version 2.1.27 contains a patch.
🎖@cveNotify
`@zereight/mcp-gitlab` is a Model Context Protocol server for GitLab. Prior to version 2.1.27, the SSE transport mode (`SSE=true`) exposes all MCP tools without any authentication. The `upload_markdown` tool reads arbitrary files from the server's local filesystem via an unsanitized `file_path` parameter and uploads them to a GitLab project. Combined, any unauthenticated network-reachable attacker can read `/proc/self/environ` to steal the server's `GITLAB_PERSONAL_ACCESS_TOKEN` and achieve full GitLab account takeover. This is the default configuration for Docker deployments. Version 2.1.27 contains a patch.
🎖@cveNotify
GitHub
fix: make download and upload work for remote deployments by Killusions · Pull Request #482 · zereight/gitlab-mcp
This is currently both a security and operational issue.
Users can both download and upload files to the server's storage.
For upload_markdown, we can just only accept inline file contents ...
Users can both download and upload files to the server's storage.
For upload_markdown, we can just only accept inline file contents ...
🚨 CVE-2026-92248
A flaw was found in the file-psd plugin in GIMP. When generating a thumbnail preview for a specially crafted PSD (Photoshop Document) image file, an integer overflow occurs during the multiplication of values from an embedded JPEG header. This leads to an undersized heap allocation, resulting in a heap-based buffer overflow when the image data is decoded. This buffer overflow corrupts adjacent heap objects, allowing for a controlled memory write that can result in an application crash or arbitrary code execution.
🎖@cveNotify
A flaw was found in the file-psd plugin in GIMP. When generating a thumbnail preview for a specially crafted PSD (Photoshop Document) image file, an integer overflow occurs during the multiplication of values from an embedded JPEG header. This leads to an undersized heap allocation, resulting in a heap-based buffer overflow when the image data is decoded. This buffer overflow corrupts adjacent heap objects, allowing for a controlled memory write that can result in an application crash or arbitrary code execution.
🎖@cveNotify
Redhat
CVE-2026-92248 - Red Hat Customer Portal
CVE Details App
🚨 CVE-2026-73463
On affected platforms running Arista EOS, when multiple gRPC Network Security Interface (gNSI) transports are configured, a race condition in the gNSI Authz service may cause a policy rotation to fail silently. An authenticated user whose access was revoked by the new policy may retain unauthorized access to gRPC interfaces. This does not affect Bootz.
This issue was discovered internally by Arista, and the company is not aware of any malicious exploitation of this vulnerability in customer networks.
🎖@cveNotify
On affected platforms running Arista EOS, when multiple gRPC Network Security Interface (gNSI) transports are configured, a race condition in the gNSI Authz service may cause a policy rotation to fail silently. An authenticated user whose access was revoked by the new policy may retain unauthorized access to gRPC interfaces. This does not affect Bootz.
This issue was discovered internally by Arista, and the company is not aware of any malicious exploitation of this vulnerability in customer networks.
🎖@cveNotify
Arista Networks
Security Advisory 0169 - Arista
September 9, 2026 Race condition during authorization policy updates may temporarily allow unintended access.
🚨 CVE-2026-92081
fastify is a fast and low-overhead web framework for Node.js. In versions before 5.12.5, when a route registers a response trailer via reply.trailer() and is served over HTTP/2, fastify unconditionally sets the Transfer-Encoding: chunked header, which is forbidden on HTTP/2, so Node.js throws while serializing the response headers. The exception is not caught and becomes an uncaughtException, so a single unauthenticated HTTP/2 request to any route that uses trailers crashes the server process and drops all in-flight requests, and it can be repeated on every restart. The issue is fixed in fastify 5.12.5, and users should upgrade to 5.12.5 or later. As a workaround, avoid registering response trailers with reply.trailer() on routes served over HTTP/2 until upgrading.
🎖@cveNotify
fastify is a fast and low-overhead web framework for Node.js. In versions before 5.12.5, when a route registers a response trailer via reply.trailer() and is served over HTTP/2, fastify unconditionally sets the Transfer-Encoding: chunked header, which is forbidden on HTTP/2, so Node.js throws while serializing the response headers. The exception is not caught and becomes an uncaughtException, so a single unauthenticated HTTP/2 request to any route that uses trailers crashes the server process and drops all in-flight requests, and it can be repeated on every restart. The issue is fixed in fastify 5.12.5, and users should upgrade to 5.12.5 or later. As a workaround, avoid registering response trailers with reply.trailer() on routes served over HTTP/2 until upgrading.
🎖@cveNotify
cna.openjsf.org
Security Advisories | OpenJS Foundation CVE Numbering Authority
The OpenJS Foundation's CVE Numbering Authority (CNA)
🚨 CVE-2026-14917
A SAML authentication bypass vulnerability affects the Kong SAML plugin when the validate_assertion_signature option is explicitly set to false. This option is enabled by default. When disabled, the plugin may extract the SAML identity from an unsigned assertion and authenticate the user without verifying a valid cryptographic signature.
As a result, an unauthenticated remote attacker may be able to submit a crafted SAML response and impersonate arbitrary users, including administrators
🎖@cveNotify
A SAML authentication bypass vulnerability affects the Kong SAML plugin when the validate_assertion_signature option is explicitly set to false. This option is enabled by default. When disabled, the plugin may extract the SAML identity from an unsigned assertion and authenticate the user without verifying a valid cryptographic signature.
As a result, an unauthenticated remote attacker may be able to submit a crafted SAML response and impersonate arbitrary users, including administrators
🎖@cveNotify
Kong Docs
Kong Gateway changelog - Kong Gateway | Kong Docs
Changelog for supported Kong Gateway Enterprise versions.
🚨 CVE-2026-19640
On affected platforms running Arista EOS, an authenticated user with access to the gNMI (gRPC Network Management Interface) may receive incorrect authorization results, potentially allowing access beyond their currently assigned permissions.
This issue was discovered internally by Arista and the company is not aware of any malicious uses of this issue in customer networks.
🎖@cveNotify
On affected platforms running Arista EOS, an authenticated user with access to the gNMI (gRPC Network Management Interface) may receive incorrect authorization results, potentially allowing access beyond their currently assigned permissions.
This issue was discovered internally by Arista and the company is not aware of any malicious uses of this issue in customer networks.
🎖@cveNotify
Arista Networks
Security Advisory 0170 - Arista
September 9, 2026 Certain gNMI requests may bypass authorization checks, allowing access beyond authorized scope.
🚨 CVE-2026-73436
On affected platforms running Arista EOS with OSPFv2 and OSPFv2 segment routing configured, a specially crafted OSPFv2 packet from an adjacent OSPF neighbor may cause OSPF to restart unexpectedly.
🎖@cveNotify
On affected platforms running Arista EOS with OSPFv2 and OSPFv2 segment routing configured, a specially crafted OSPFv2 packet from an adjacent OSPF neighbor may cause OSPF to restart unexpectedly.
🎖@cveNotify
Arista Networks
Security Advisory 0171 - Arista
Date: September 9, 2026 Crafted packets from an adjacent device can crash the OSPF agent, causing routing disruption. Crafted packets can crash the OSPF agent when Segment Routing is enabled.
🚨 CVE-2026-73440
On affected platforms running Arista EOS with Simple Network Management Protocol (SNMP) configured, SNMPv3 local or remote user credentials may be exposed as a one-way hashed, localized key value within the device's running and sanitized configurations. An authenticated user who gains access to this sensitive information could leverage it to perform unauthorized read operations on SNMP tables or to send fraudulent trap notifications to the Network Management System (NMS).
This issue was discovered internally by Arista, and the company is not aware of any malicious exploitation of this vulnerability in customer networks.
🎖@cveNotify
On affected platforms running Arista EOS with Simple Network Management Protocol (SNMP) configured, SNMPv3 local or remote user credentials may be exposed as a one-way hashed, localized key value within the device's running and sanitized configurations. An authenticated user who gains access to this sensitive information could leverage it to perform unauthorized read operations on SNMP tables or to send fraudulent trap notifications to the Network Management System (NMS).
This issue was discovered internally by Arista, and the company is not aware of any malicious exploitation of this vulnerability in customer networks.
🎖@cveNotify
Arista Networks
Security Advisory 0178 - Arista
September 9, 2026 SNMPv3 credentials may appear in clear text in logs or running config, enabling unauthorized SNMP access.
🚨 CVE-2026-73453
An unauthenticated P4Runtime (Programming Protocol-Independent Packet Processors Runtime) client can achieve arbitrary code execution under certain conditions on affected platforms running Arista EOS configured with P4Runtime. P4Runtime is disabled by default in Arista EOS. By crafting a malicious packet during the initiation of a P4Runtime session, an attacker can obtain complete administrative control over the compromised switch.
This issue was discovered internally by Arista, and the company is not aware of any malicious exploitation of this vulnerability in customer networks.
🎖@cveNotify
An unauthenticated P4Runtime (Programming Protocol-Independent Packet Processors Runtime) client can achieve arbitrary code execution under certain conditions on affected platforms running Arista EOS configured with P4Runtime. P4Runtime is disabled by default in Arista EOS. By crafting a malicious packet during the initiation of a P4Runtime session, an attacker can obtain complete administrative control over the compromised switch.
This issue was discovered internally by Arista, and the company is not aware of any malicious exploitation of this vulnerability in customer networks.
🎖@cveNotify
Arista Networks
Security Advisory 0174 - Arista
September 9, 2026 Unauthenticated attacker can send crafted requests to execute arbitrary code, leading to full device compromise.
🚨 CVE-2026-73468
A specially crafted packet can cause the premature expiry of multicast forwarding state on affected interfaces, potentially resulting in temporary multicast traffic loss during the affected period.
🎖@cveNotify
A specially crafted packet can cause the premature expiry of multicast forwarding state on affected interfaces, potentially resulting in temporary multicast traffic loss during the affected period.
🎖@cveNotify
Arista Networks
Security Advisory 0175 - Arista
September 9, 2026 A specially crafted packet can cause the premature expiry of multicast forwarding state on affected interfaces.
🚨 CVE-2026-73469
When specific platforms are using Arista EOS with a loose Unicast Reverse Path Forwarding (uRPF) configuration, certain traffic may not be subjected to the intended verification drop. Consequently, traffic that should be dropped based on these routes could still be processed and forwarded by the device.
This issue was discovered internally by Arista and the company is not aware of any malicious uses of this issue in customer networks.
🎖@cveNotify
When specific platforms are using Arista EOS with a loose Unicast Reverse Path Forwarding (uRPF) configuration, certain traffic may not be subjected to the intended verification drop. Consequently, traffic that should be dropped based on these routes could still be processed and forwarded by the device.
This issue was discovered internally by Arista and the company is not aware of any malicious uses of this issue in customer networks.
🎖@cveNotify
Arista Networks
Security Advisory 0176 - Arista
September 9, 2026 Misconfigured uRPF allows spoofed-source traffic to bypass the check and traverse the device.
🚨 CVE-2026-76151
Out-of-bounds read (buffer over-read) in the HTTP Cache-Control response header parsing in the QtNetwork module in Qt Group Qt 6.0.0 through 6.8.8, and 6.9.0 through 6.11.1, allows remote attackers to cause a denial of service (application crash) via an excessively large Cache-Control header value returned by an untrusted or compromised HTTP server to an application using QNetworkAccessManager. Only the client side of the connection is affected and 32-bit builds are not affected; the out-of-bounds access is read-only, with no information disclosure and no code execution.
🎖@cveNotify
Out-of-bounds read (buffer over-read) in the HTTP Cache-Control response header parsing in the QtNetwork module in Qt Group Qt 6.0.0 through 6.8.8, and 6.9.0 through 6.11.1, allows remote attackers to cause a denial of service (application crash) via an excessively large Cache-Control header value returned by an untrusted or compromised HTTP server to an application using QNetworkAccessManager. Only the client side of the connection is affected and 32-bit builds are not affected; the out-of-bounds access is read-only, with no information disclosure and no code execution.
🎖@cveNotify
🚨 CVE-2026-77190
On affected platforms running Arista EOS, an unauthenticated attacker who is network-adjacent to the switch and able to connect to a device with PIM Sparse Mode and MLAG configured, can send malformed messages that cause the Pimsm agent to terminate unexpectedly. The Pimsm agent is automatically restarted, but repeated attacks can cause the agent to restart continuously, resulting in a sustained denial of service.
🎖@cveNotify
On affected platforms running Arista EOS, an unauthenticated attacker who is network-adjacent to the switch and able to connect to a device with PIM Sparse Mode and MLAG configured, can send malformed messages that cause the Pimsm agent to terminate unexpectedly. The Pimsm agent is automatically restarted, but repeated attacks can cause the agent to restart continuously, resulting in a sustained denial of service.
🎖@cveNotify
Arista Networks
Security Advisory 0177 - Arista
September 9, 2026 Malformed packets can cause the S4Sync agent to crash, leading to DoS for MLAG synchronization.
🚨 CVE-2026-86443
Cleartext storage of sensitive information in the DuoxMe application for Android, in versions prior to 4.3.4, allows an attacker with local access to the device to retrieve the credentials stored by the application and impersonate the user account.
🎖@cveNotify
Cleartext storage of sensitive information in the DuoxMe application for Android, in versions prior to 4.3.4, allows an attacker with local access to the device to retrieve the credentials stored by the application and impersonate the user account.
🎖@cveNotify
Fermax
Fermax Professional | FERMAX
🚨 CVE-2026-86474
The lack of TLS certificate validation when downloading firmware updates in VEO and VEO-XS Wi-Fi monitors, in versions prior to 01.48.001, allows an attacker to perform man-in-the-middle attacks on the update channel.
🎖@cveNotify
The lack of TLS certificate validation when downloading firmware updates in VEO and VEO-XS Wi-Fi monitors, in versions prior to 01.48.001, allows an attacker to perform man-in-the-middle attacks on the update channel.
🎖@cveNotify
Fermax
Fermax Professional | FERMAX
🚨 CVE-2026-92456
yshop-crm through 2.1.3 fails to enforce authorization on the saveRedisSet and getRedisSet endpoints in CrmCustomerController, allowing any authenticated back-office user to read and modify installation-wide lead-allocation and customer auto-recycling policy. Attackers can invoke these endpoints to manipulate shared Redis keys controlling customer auto-recycling behavior, causing mass customer data deletion, disabling lead recycling, or blocking customer creation across the deployment.
🎖@cveNotify
yshop-crm through 2.1.3 fails to enforce authorization on the saveRedisSet and getRedisSet endpoints in CrmCustomerController, allowing any authenticated back-office user to read and modify installation-wide lead-allocation and customer auto-recycling policy. Attackers can invoke these endpoints to manipulate shared Redis keys controlling customer auto-recycling behavior, causing mass customer data deletion, disabling lead recycling, or blocking customer creation across the deployment.
🎖@cveNotify
GitHub
cve-request-poc/yshop-crm/C04_crm_customer_save_redis_set.py at master · LinYuanyi1/cve-request-poc
poc repo. Contribute to LinYuanyi1/cve-request-poc development by creating an account on GitHub.
🚨 CVE-2026-92461
yshop-crm through 2.1.3 contains a missing authorization vulnerability in the GET /admin-api/crm/flow/flow-users endpoint that allows any logged-in back-office user to access approval workflow data. Attackers can retrieve approval chain topology, step ordering, approver identifiers, and personal information including login names, nicknames, departments, email addresses, mobile numbers and last login IP addresses.
🎖@cveNotify
yshop-crm through 2.1.3 contains a missing authorization vulnerability in the GET /admin-api/crm/flow/flow-users endpoint that allows any logged-in back-office user to access approval workflow data. Attackers can retrieve approval chain topology, step ordering, approver identifiers, and personal information including login names, nicknames, departments, email addresses, mobile numbers and last login IP addresses.
🎖@cveNotify
GitHub
cve-request-poc/yshop-crm/C08_crm_flow_users.py at master · LinYuanyi1/cve-request-poc
poc repo. Contribute to LinYuanyi1/cve-request-poc development by creating an account on GitHub.
🚨 CVE-2026-88817
An authenticated, non-guest user of Curiosity Workspace could enroll themselves as an administrator and member of an existing access group without an invitation or approval.
It did not grant application-wide administrator privileges, and the vulnerability could not be used to obtain root access to the application or its underlying host.
🎖@cveNotify
An authenticated, non-guest user of Curiosity Workspace could enroll themselves as an administrator and member of an existing access group without an invitation or approval.
It did not grant application-wide administrator privileges, and the vulnerability could not be used to obtain root access to the application or its underlying host.
🎖@cveNotify
docs.curiosity.ai
Curiosity - CVE-2026-88817
Security information for Curiosity Workspace: controls, advisories, and how to report a vulnerability.
🚨 CVE-2026-92359
A security flaw has been discovered in ag-ui-protocol ag-ui 0.3.0. The affected element is the function create_strands_app of the file integrations/aws-strands/python/src/ag_ui_strands/utils.py of the component CORSMiddleware. The manipulation results in permissive cross-domain policy with untrusted domains. The attack may be launched remotely. The attack requires a high level of complexity. The exploitability is described as difficult. Upgrading to version AGUI.Abstractions@0.0.6 is sufficient to fix this issue. The patch is identified as 9b143b9668fa52c2054ede9d34a45ac4b4401089. It is suggested to upgrade the affected component.
🎖@cveNotify
A security flaw has been discovered in ag-ui-protocol ag-ui 0.3.0. The affected element is the function create_strands_app of the file integrations/aws-strands/python/src/ag_ui_strands/utils.py of the component CORSMiddleware. The manipulation results in permissive cross-domain policy with untrusted domains. The attack may be launched remotely. The attack requires a high level of complexity. The exploitability is described as difficult. Upgrading to version AGUI.Abstractions@0.0.6 is sufficient to fix this issue. The patch is identified as 9b143b9668fa52c2054ede9d34a45ac4b4401089. It is suggested to upgrade the affected component.
🎖@cveNotify
GitHub
GitHub - ag-ui-protocol/ag-ui: AG-UI: the Agent-User Interaction Protocol. Bring Agents into Frontend Applications.
AG-UI: the Agent-User Interaction Protocol. Bring Agents into Frontend Applications. - ag-ui-protocol/ag-ui
🚨 CVE-2026-19662
An attacker may be able to cause a `named` resolver to abort. The attack requires inducing the victim resolver to send multiple queries for a DNSSEC-signed zone hosted by an authoritative server under the control of the attacker. If the auth responds with a particular sequence of crafted answers, and those answers arrive in a particular order with particular timing, the `named` resolver will encounter a use-after-free bug, and abort.
This issue affects BIND 9 versions 9.11.0 through 9.18.50, 9.20.0 through 9.20.27, 9.11.3-S1 through 9.18.50-S1, and 9.20.9-S1 through 9.20.27-S1.
🎖@cveNotify
An attacker may be able to cause a `named` resolver to abort. The attack requires inducing the victim resolver to send multiple queries for a DNSSEC-signed zone hosted by an authoritative server under the control of the attacker. If the auth responds with a particular sequence of crafted answers, and those answers arrive in a particular order with particular timing, the `named` resolver will encounter a use-after-free bug, and abort.
This issue affects BIND 9 versions 9.11.0 through 9.18.50, 9.20.0 through 9.20.27, 9.11.3-S1 through 9.18.50-S1, and 9.20.9-S1 through 9.20.27-S1.
🎖@cveNotify
🚨 CVE-2026-19667
If an attacker-controlled authoritative server can produce a negative answer that is exactly 65536 bytes, then a flaw in `named` results in a negative cache entry of 0 bytes. When this entry is subsequently read, `named` aborts.
This issue affects BIND 9 versions 9.11.0 through 9.18.50, 9.20.0 through 9.20.27, 9.21.0 through 9.21.25, 9.11.3-S1 through 9.18.50-S1, and 9.20.9-S1 through 9.20.27-S1.
🎖@cveNotify
If an attacker-controlled authoritative server can produce a negative answer that is exactly 65536 bytes, then a flaw in `named` results in a negative cache entry of 0 bytes. When this entry is subsequently read, `named` aborts.
This issue affects BIND 9 versions 9.11.0 through 9.18.50, 9.20.0 through 9.20.27, 9.21.0 through 9.21.25, 9.11.3-S1 through 9.18.50-S1, and 9.20.9-S1 through 9.20.27-S1.
🎖@cveNotify