Vulnerabilities

Leantime before 3.9.6 contains an authorization bypass (CWE-862) in its HTMX plugin install endpoint, which lacks permission validation. Any authenticated user with a limited or low-privilege role can call this network-reachable endpoint and install plugins from the marketplace, fully controlling arbitrary properties including the plugin identifier, version, and license key. By abusing this, an attacker can deploy a malicious or attacker-controlled plugin onto the server, compromising the integrity of the installation (CVSS 4.0: 7.1 High, network vector, low privileges, no user interaction). All Leantime deployments running a version earlier than 3.9.6 that host low-privileged user accounts are affected. No public proof-of-concept or in-the-wild exploitation is known, and the issue is not listed in CISA KEV.

Do: Upgrade Leantime to version 3.9.6 or later, which adds the missing permission validation on the plugin install endpoint. Until patched, review which low-privileged users hold accounts on the instance and audit the installed plugin list for unrecognized identifiers, versions, or license keys. Limit network exposure of the Leantime instance to trusted users as an interim mitigation.

Twenty, an open-source CRM, prior to version 2.35.0 fails to enforce field- and row-level permissions in the groupBy-with-records GraphQL resolver (CWE-863, incorrect authorization). An authenticated user who holds object-level read permission (canReadObjectRecords) but has field-level read denied (canReadFieldValue false) can call the groupBy resolver and retrieve restricted field values that the permission system would normally hide. Exploitation requires only a valid low-privilege account and no user interaction, resulting in high confidentiality impact (CVSS 4.0: 7.1 High). Any deployment running an affected version is exposed, since the flaw is reachable through the standard GraphQL API. No public proof-of-concept or in-the-wild exploitation is currently known.

Do: Upgrade Twenty to version 2.35.0 or later, which fixes the groupBy resolver permission enforcement. Until patched, review user roles for accounts holding canReadObjectRecords and restrict GraphQL API access (e.g., via reverse-proxy rules or allowlists) for untrusted users, and inspect access logs for groupBy queries against objects with restricted fields. Note that any authenticated user account, including low-privilege ones, could have been used to read hidden field values, so consider auditing sensitive fields for unauthorized reads.

Harbor through 2.15.2 fails to restrict the q query-parameter filtering on scanner registration records, leaving the AccessCredential column open to server-side fuzzy matching (CWE-200, information exposure). A user holding the project administrator role can issue crafted filter queries against the scanner registration API and use the returned row counts as an oracle, recovering the scanner adapter's stored secret one character at a time. With that secret, a semi-trusted project-level administrator can authenticate to the scanner adapter service, gaining access to scanner capabilities and data beyond their intended privileges. Any Harbor deployment that assigns the project administrator role to users who should not see scanner credentials is affected, particularly where a scanner adapter with access credentials is configured. No public proof-of-concept is known, the flaw is not in CISA KEV, and no exploitation has been confirmed.

Do: Upgrade Harbor to a release newer than 2.15.2 that restricts q-parameter filtering on the AccessCredential field as soon as one is available. In the interim, limit the project administrator role to fully trusted users, audit API logs for unusual q-filter queries against scanner registration list endpoints, and rotate scanner adapter credentials after patching.

CVE-2026-92765 is an organization-ownership validation flaw (CWE-639, broken object-level authorization) in the WebScanVulnList endpoint of ArcherySec, an open-source vulnerability assessment and management platform. An authenticated user can supply arbitrary scan identifiers belonging to a different organization, and the endpoint returns the results without checking tenant ownership. This lets the attacker read complete web vulnerability records from other tenants, including titles, severities, statuses, and analyst notes, though no write or availability impact is involved. Any multi-tenant deployment of ArcherySec up to and including version 2.0.6 is affected, with the greatest risk where low-privileged accounts exist across tenant boundaries or where the instance is internet-facing. No public proof-of-concept, KEV listing, or confirmed in-the-wild exploitation is known; the issue was assigned by the VulnCheck CNA.

Do: Upgrade ArcherySec to the latest release (any version newer than 2.0.6) once a patched build is available, since the flaw affects everything through 2.0.6. In the meantime, restrict access to WebScanVulnList at the reverse proxy or API-gateway layer, minimize the number of authenticated accounts (especially low-privileged ones in multi-tenant setups), and review access logs for requests enumerating scan IDs outside the user's own organization. If the instance is internet-exposed, limit it to trusted networks or VPN access.

Rundeck through 6.2.1 fails to enforce authorization on the importConfig and importNodesSources parameters of the project archive import endpoint, a missing-authorization flaw (CWE-862). An authenticated attacker who holds only the 'import' action on a project — without broader admin or configuration privileges — can invoke the import endpoint with those parameters enabled. This allows the attacker to replace project configuration files, including security-relevant settings such as node executors and SSH key paths that control how jobs execute, giving high integrity and confidentiality impact per the CVSS score. Any Rundeck deployment at version 6.2.1 or earlier in which less-privileged users are granted the import action is affected. No public proof-of-concept, CISA KEV listing, or confirmed in-the-wild exploitation is currently known.

Do: Upgrade Rundeck to a release newer than 6.2.1 as soon as a patched version is available. Until then, restrict the project 'import' action to trusted administrators, audit recent project archive import activity, and review project configuration files (node executors, SSH key paths) for unauthorized modifications.

Pelican Panel versions before 1.0.0-beta35 enforce startup write permissions only through disabled form controls in the web UI, with no corresponding server-side authorization checks (CWE-862). A user holding only the startup.read permission — typically a delegated subuser on a game server — can craft Livewire state updates that invoke afterStateUpdated callbacks, causing the panel to accept changes it should have rejected. By modifying the startup command, Docker image, and container variables this way, the attacker achieves arbitrary command execution inside that server's container. Any operator running an affected Pelican Panel version who has granted limited accounts startup read access is exposed, a common pattern among game-server hosting providers and communities. No public proof-of-concept is known, the flaw is not in CISA's KEV catalog, and no in-the-wild exploitation has been reported.

Do: Upgrade Pelican Panel to 1.0.0-beta35 or later, which adds server-side enforcement of startup write permissions. Until upgraded, audit all subuser accounts and remove or restrict startup read grants for any user who should not be able to run code in the container, and treat existing startup.read holders as potential code-execution risk. Review container logs for unexpected changes to startup commands, Docker images, or variables, and limit panel network exposure where possible.

WebVirtCloud, a web panel for managing KVM virtual machines, fails to validate the permission flags attached to UserInstance grants, checking only that a grant exists via its get_instance gate. As a result, any user holding even a view-only grant can invoke privileged instance actions remotely, including powering off virtual machines, resetting root passwords, installing SSH keys, and managing ISO images. Resetting a VM's root password and injecting an SSH key effectively hands the attacker full control of the guest system, and the attack requires only valid low-privilege credentials with no user interaction. Any WebVirtCloud operator that has granted instance access to untrusted or shared users is affected. No exploitation in the wild or public proof-of-concept is currently known, and the issue is not listed in CISA's KEV catalog.

Do: Update WebVirtCloud to the latest available release once a patched version is published, since the data does not yet identify a fixed version. In the interim, audit all UserInstance grants and revoke or restrict view-only grants for untrusted users, and monitor instance action logs for unauthorized power-offs, root password resets, SSH key additions, or ISO changes. Apply the CVSS low-privilege caveat: only operators who share panel accounts with third parties are meaningfully exposed.

Shlink, a self-hosted URL shortener, through version 5.1.6 fails to enforce API key role restrictions when issuing Mercure real-time subscription tokens, so keys scoped to specific authors or domains receive tokens valid for all topics (CWE-863, incorrect authorization). An attacker who already holds an author-only or domain-only API key can call the mercure-info endpoint and subscribe to topics outside their authorization boundary. This exposes live visit data — including referrer, user agent, and geolocation — plus full short URL objects for URLs the key should not be able to see, a high-confidentiality issue with no integrity or availability impact. Only deployments running Shlink 5.1.6 or earlier that issue restricted API keys to third parties or limited users are materially affected. No public proof-of-concept is known, the issue is not in CISA KEV, and no exploitation has been reported.

Do: Upgrade to a Shlink release newer than 5.1.6 that enforces API key roles in Mercure subscription tokens as soon as one is available. Until then, rotate or restrict author-only and domain-only API keys, limit external access to the Mercure hub and mercure-info endpoint (e.g., at the reverse proxy), and review hub/subscription logs for topic access outside authorized boundaries.

SecObserve, an open-source vulnerability management platform, fails to strip the basic_auth_password field from API configuration responses in its ApiConfigurationSerializer. Any product member with view-only permissions can therefore call standard REST endpoints and retrieve the decrypted basic-auth password of configured scanner or integration service accounts. An attacker holding a low-privileged account gains working credentials for the connected scanner or integration services, which can enable further access to those upstream systems. All SecObserve deployments running versions before 1.59.1 are affected. No public proof-of-concept or exploitation has been reported, and the issue is not in CISA's KEV catalog.

Do: Upgrade SecObserve to version 1.59.1 or later. Rotate the basic-auth credentials of all scanner and integration service accounts configured in SecObserve, since any view-only member may have already been able to read them via the REST API. Review product member roles and confirm low-privileged users only have the access they need.

PatrowlManager, an open-source threat-intelligence and security-operations management platform, fails through version 1.8.4 to enforce object-level authorization on its events and alerts API endpoints, which lack ownership filtering (CWE-862). Any attacker with a valid low-privilege account and network access to the instance can call these endpoints to read the platform's full event history, delete arbitrary events, and modify alert records belonging to other users. The flaw requires no user interaction and is scored 7.1 (high) under CVSS 4.0. All deployments running PatrowlManager 1.8.4 or earlier are affected, especially multi-user or internet-exposed instances. No public proof-of-concept is known, the flaw is not in CISA KEV, and no exploitation has been reported.

Do: Check the Patrowl GitHub project for a patched release beyond 1.8.4 and upgrade as soon as one is available. Until then, restrict access to the PatrowlManager web UI and API to trusted networks via firewall or VPN, limit the number of low-privilege accounts on shared instances, and review logs for unexplained event deletions or alert modifications by non-owner accounts.

CVE-2026-92752 is a broken object-level authorization flaw (CWE-639) in metasfresh's REST API: the DocumentAttachmentsRestController and CommentsRestController endpoints only verify that the caller is authenticated and never check whether the caller's role is permitted to access the referenced record. An authenticated user can enumerate sequential document identifiers and, for each, list/read, replace, or delete attachments and comments belonging to records outside their permitted scope. This can expose sensitive business documents (e.g., invoices, contracts, internal notes) held on other users' or other roles' records, and allows tampering with or destroying that content. Any metasfresh deployment using the web UI REST API with multiple roles or restricted records is affected. Exploitation has not been observed: the issue is not in CISA's KEV and no public proof-of-concept is known.

Do: Upgrade to the fixed metasfresh release once the vendor advisory identifies it, and in the interim restrict or place authentication/authorization checks in front of the DocumentAttachments and Comments REST endpoints (e.g., reverse-proxy rules or WAF). Audit attachment and comment records for unexpected reads, replacements, or deletions by low-privileged accounts, and review role definitions to confirm record-level access is enforced elsewhere. Note that the attack requires valid credentials, so also review and rotate accounts with low-privileged roles if misuse is suspected.

CMAK (Cluster Manager for Apache Kafka, formerly Yahoo Kafka Manager) through version 3.0.0.6 does not install a cross-site request forgery (CSRF) filter, leaving its state-changing endpoints unprotected. An attacker can lure an authenticated CMAK operator to a page containing hidden forms that auto-submit to destructive endpoints such as topic deletion and cluster configuration changes. The requests are authenticated using the operator's HTTP Basic credentials or the play-basic-authentication cookie, which lacks SameSite protection and therefore is sent along with the forged cross-origin request. Any organization running CMAK 3.0.0.6 or earlier is affected, with operators' browser sessions as the attack vector. No public proof-of-concept, CISA KEV listing, or confirmed in-the-wild exploitation is known.

Do: No fixed release is identified in the available data, so monitor the CMAK project for a patched version beyond 3.0.0.6 and upgrade when one is published. In the interim, restrict CMAK access to trusted networks or a VPN, place it behind a reverse proxy that enforces Origin/Referer validation on state-changing requests, and add SameSite=Lax or Strict attributes to the play-basic-authentication cookie. Audit recent topic deletions and cluster configuration changes for signs of forged requests.

CVE-2026-92750 is a missing access control (CWE-862) flaw in Harness, which through version 3.3.0 fails to validate authorization on the infrastructure provider read endpoint. Any authenticated user can send GET requests to /api/v1/infraproviders with arbitrary space identifiers belonging to spaces they are not a member of, and the server returns the provider configuration. This leaks sensitive provider metadata such as Docker endpoints, TLS certificate paths, and cloud project identifiers, which could aid further attacks against the underlying infrastructure. Any Harness deployment at version 3.3.0 or earlier with more than one user or space is affected. No public proof-of-concept, KEV listing, or confirmed exploitation has been reported.

Do: Until a patched release beyond 3.3.0 is available, restrict which authenticated accounts can reach the Harness API, audit access logs for GET /api/v1/infraproviders calls referencing space identifiers outside the caller's membership, and apply network-level access controls to the API. Treat Docker endpoints, TLS certificate paths, and cloud project identifiers returned by this endpoint as potentially disclosed, and rotate or tighten related credentials and network exposure if compromise is suspected. Monitor vendor advisories for a fixed version and upgrade promptly once released.

SafeLine, an open-source web application firewall, derives the secret used to sign management-console session cookies from a weakly seeded math/rand generator (CWE-338), making the key reconstructable offline. An unauthenticated remote attacker who can narrow the window of when the instance was installed can regenerate the session-signing secret and forge valid administrator session cookies. Successful exploitation grants full admin access to the management console, letting the attacker change WAF rules, inspect or manipulate traffic, and effectively control the protected sites. All SafeLine deployments at version 9.4.1 or earlier are affected, with the greatest risk on instances whose console is internet-facing. No public proof-of-concept or in-the-wild exploitation has been reported, and the flaw is not on CISA's KEV.

Do: Upgrade SafeLine to a fixed release beyond 9.4.1 as soon as one is available and, where possible, regenerate the console's secret by reinstalling/upgrading. Until patched, restrict the management console to trusted networks (VPN or IP allowlist) and verify it is not exposed to the internet, since the attack requires no credentials or user interaction. Review console logs for unexplained administrator sessions or rule changes that could indicate forged-cookie access.

BC Security Empire, an open-source command-and-control (C2) post-exploitation framework, fails to validate the multipart filename parameter on its upload endpoints in versions before 6.7.1. An authenticated operator can supply path traversal sequences (e.g., '../') in the filename to escape the intended upload directory and write files to arbitrary paths on the C2 server. Because the write is attacker-directed, an operator can plant malicious files in sensitive locations such as startup or plugin directories to achieve code execution on the C2 server with its privileges. Any deployment running Empire prior to 6.7.1 is affected; the user base consists mainly of red teams and penetration testers, and the framework is also known to be abused by threat actors. No public proof-of-concept is known, the flaw is not listed in CISA's KEV catalog, and there is no confirmed in-the-wild exploitation.

Do: Upgrade Empire to 6.7.1 or later immediately. Restrict operator accounts to trusted personnel with least privilege, since any authenticated operator can exploit this, and review the C2 server filesystem for unexpected files in sensitive locations (startup directories, plugin/module folders). Limit network exposure of the C2 management/API interfaces to only those who need it.

Cisco Identity Services Engine (ISE) and the Cisco ISE Passive Identity Connector (ISE-PIC) contain an incorrect use of privileged APIs flaw (CWE-648) affecting the web-based management interface. An unauthenticated, remote attacker with network access to that interface can send requests that invoke privileged APIs without authenticating, bypassing the interface's access controls. Successful exploitation grants the attacker unauthorized access to the affected device, presumably with the administrative capabilities available through the management interface, such as control over network access policy and visibility into identity data. Any organization running an affected Cisco ISE or ISE-PIC release is potentially affected, with risk highest where the management interface is reachable from untrusted networks. The flaw was added to CISA's Known Exploited Vulnerabilities catalog on 2026-09-16, indicating exploitation in the wild, though no public proof-of-concept is known and CVSS scoring is pending.

Do: Upgrade ISE and ISE-PIC to the fixed releases specified in Cisco's security advisory (fixed versions are not provided in the available data); because the flaw is on CISA's KEV list, federal agencies must patch or apply mitigations per BOD 26-04 timelines. Until patched, restrict access to the web-based management interface to trusted administrative networks only, verify no unintended exposure via firewalls/ACLs, and monitor for unauthenticated access attempts against the interface.

Cisco Identity Services Engine (ISE) contains a SQL injection flaw in its APIs caused by insufficient validation of parameters that are concatenated directly into backend database queries. An attacker who sends a specially crafted request containing SQL statements to an affected API endpoint can trigger the flaw, but valid administrative credentials are required to do so. A successful exploit allows the attacker to read arbitrary content from the SQL database and to conduct server-side request forgery (SSRF) attacks. All organizations running affected releases of Cisco ISE are potentially affected, though exploitation requires a privileged attacker account. As of now there is no evidence of in-the-wild exploitation and no public proof-of-concept is known, and the flaw is not listed in CISA's KEV catalog.

Do: Check the Cisco PSIRT advisory for CVE-2026-76425 and upgrade ISE to a fixed release as soon as one is identified. Because exploitation requires valid administrative credentials, restrict access to the ISE admin and API interfaces to trusted management networks, enforce strong authentication and least privilege for admin accounts, and audit/rotate administrative credentials. Monitor API logs for unexpected queries and outbound requests that could indicate attempted SSRF.

Cisco Identity Services Engine (ISE) contains a path traversal flaw (CWE-23) in its REST API caused by insufficient validation of file operations. An attacker who already holds valid administrative credentials can send an API request uploading a file with a crafted path, allowing the file to be written to arbitrary locations on the device. Successful exploitation lets the attacker upload files anywhere on the appliance and execute arbitrary commands with root privileges, giving full compromise of the ISE deployment. Because high privileges are required, the flaw is rated CVSS 3.1 7.2 (High) with AV:N/AC:L/PR:H. It is not currently listed in CISA's KEV catalog and no public proof-of-concept or in-the-wild exploitation is known.

Do: Upgrade ISE to the fixed releases listed in the Cisco PSIRT advisory for CVE-2026-76424. Until patched, restrict access to the ISE REST API to trusted management networks, audit privileged/administrative accounts for compromise, and review API logs for file-upload requests containing path traversal characters (e.g., ../ sequences).

Cisco Secure Firewall Management Center (FMC) Software contains a flaw in the Adaptive Security Device Manager (ASDM) single sign-on (SSO) handler, caused by improper management of the ASDM SSO token (CWE-1259). An unauthenticated, remote attacker can exploit it by performing session token forgery techniques against the ASDM SSO login mechanism. A successful attack lets the attacker log in as the ASDM administrator without any credentials, and by repeatedly forging tokens they can keep legitimate administrators locked out of ASDM indefinitely, consistent with the high-availability impact reflected in the 8.2 CVSS score. Any organization running Cisco Secure FMC with ASDM SSO enabled is exposed, since no user interaction or prior access is required. As of now there is no known exploitation, no public proof-of-concept, and the flaw is not listed in CISA KEV.

Do: Upgrade Cisco ASDM and Cisco Secure FMC to the fixed releases identified in Cisco's advisory for CVE-2026-76413 (no fixed version numbers were provided in this data). Until patched, restrict ASDM and FMC management interfaces to trusted administrative networks or VPN access via access control lists, and monitor for unexpected ASDM administrator logins or repeated administrator lockouts, which could indicate token-forgery attempts.

CVE-2026-76412 is a privilege-escalation flaw (CWE-264) in the remote diagnostics debugger of Cisco Secure FMC (Firewall Management Center) Software, caused by an error when checking the privilege level of a user invoking remote diagnostics. An attacker who holds valid credentials on the device—no matter how low-privileged—can authenticate via the web-based management interface or the REST API and use the remote diagnostics debugger to enable the service and grant elevated privileges. A successful exploit elevates the attacker to root on the FMC, the central management platform for Cisco Secure Firewalls, giving full control over firewall management (scope-changing, with high confidentiality, integrity, and availability impact). Any organization running Cisco Secure FMC with user accounts is potentially affected, though attack complexity is high because exploitation requires valid credentials and a multistage process. No public proof-of-concept is known, exploitation has not been observed in the wild, and the flaw is not on the CISA KEV list.

Do: Upgrade FMC to the fixed release specified in Cisco's security advisory for CVE-2026-76412 (exact version numbers are not provided in this dataset). Until patched, restrict access to the FMC web-based management interface and REST API to trusted management networks or a dedicated management interface, audit and remove unnecessary local user accounts, and monitor device logs for enabling or use of the remote diagnostics debugger service.

CVE-2026-76409 covers multiple improper limitation of a pathname (path traversal, CWE-22) vulnerabilities in Cisco Nexus Dashboard, discovered during Cisco's internal security review and fixed via a software hardening release. A remote attacker who already holds low-privileged credentials can submit crafted pathnames that bypass directory restrictions on the appliance. Per the CVSS 8.8 (high) score, successful exploitation can yield high-impact consequences for confidentiality, integrity, and availability, potentially including arbitrary file read/write on the management platform. Organizations running Cisco Nexus Dashboard as the management control plane for their data center fabrics are affected. No public proof-of-concept is known, the issue is not in CISA's KEV catalog, and there is no evidence of exploitation in the wild.

Do: Upgrade Cisco Nexus Dashboard to the latest hardening release referenced in Cisco's advisory, and check the advisory for the exact fixed version applicable to your deployment track. Because exploitation requires valid low-privileged credentials, enforce strong authentication and restrict management interfaces to trusted administrative networks; monitor Cisco PSIRT for updated details, as the flaw was internally found and public information is limited.

Marten, a .NET transactional document database and event store built on PostgreSQL, fails to escape or parameterize runtime strings interpolated into single-quoted SQL literals in versions 7.0.0 through before 9.13.0. The confirmed injection point is the dictionary indexer key used by LINQ Where filters (DictionaryItemMember.cs), with additional sinks in SelectParser.cs, DatabaseScopedTenantPartitions.cs, DeleteAllForTenant.cs (reachable via IEventStore.DeleteProjectionProgressAsync), the Newtonsoft-serializer-only DictionaryContainsKeyFilter.cs (System.Text.Json is not affected), and a defense-in-depth per-tenant partition literal in EventLoader.cs. An attacker who controls a value containing a single quote — for example a dictionary key used in a query filter or a tenant identifier — can break out of the generated literal to bypass query filters or multi-tenant authorization and perform blind data exfiltration. Where the deployment permits semicolon-batched Npgsql statements, the injection can extend to data modification. Any application built with Marten 7.0.0–9.12.x on PostgreSQL, particularly multi-tenant apps that pass attacker-influenced keys or tenant IDs into LINQ queries, is affected; no public PoC exists, it is not in CISA KEV, and exploitation is not known to date.

Do: Upgrade to Marten 9.13.0 or later. Until patched, validate/escape single quotes in values used as dictionary indexer keys in Where filters and in tenant identifiers, avoid enabling semicolon-batched Npgsql statements, and audit multi-tenant query paths (select parsing, tenant partition management, DeleteAllForTenant/DeleteProjectionProgressAsync); note the ContainsKey sink applies only when the Newtonsoft serializer is used.

CVE-2026-63506 is an authorization bypass (CWE-639) in the self-hosted authentication logic of Tina, a headless CMS: the isAuthorized function in @tinacms/auth and next-tinacms-azure accepts a request-controlled clientID and validates the bearer token against that attacker-chosen TinaCloud app instead of the victim site's configured app. An attacker holding any valid TinaCloud account can submit their own app ID and token to a victim endpoint, causing TinaCloudBackendAuthProvider or affected media authorization callbacks to accept the attacker as verified across the tenant boundary. Successful exploitation permits listing, reading, uploading, or deleting media and, where TinaCloudBackendAuthProvider is used, GraphQL create, read, update, and delete operations on the victim's content, with no victim account or interaction required. Only self-hosted Tina deployments using @tinacms/auth prior to 1.1.4 or next-tinacms-azure prior to 15.0.1 are affected. No public proof-of-concept or confirmed in-the-wild exploitation is known, and the issue is not listed in CISA KEV.

Do: Upgrade to @tinacms/auth 1.1.4 or later and next-tinacms-azure 15.0.1 or later, verifying the resolved versions in your lockfile (package-lock.json / yarn.lock). As an interim mitigation, restrict network access to Tina API and media routes and reject client-supplied clientID values in authorization checks. Review media and content audit logs for unexpected tokens or GraphQL mutations from unfamiliar TinaCloud apps that could indicate cross-tenant access.

Kedro-Datasets' opt-in experimental PyTorch connector (kedro_datasets_experimental.pytorch.PyTorchDataset) loads .pt model files via torch.load without enforcing weights_only=True, and user-supplied load_args are silently dropped, so safe loading cannot be enabled through configuration (CWE-502). It is triggered when a Kedro pipeline loads a pickle-backed .pt model from an attacker-influenced source, such as a shared registry, a downloaded checkpoint, or a partitioned external data source, on PyTorch versions earlier than 2.6. An attacker gains arbitrary code execution on the host running the pipeline. Only users of the opt-in experimental component who load untrusted files are affected; pipelines loading only trusted files are not impacted. No public proof-of-concept is known, it is not in CISA KEV, and no exploitation has been reported; the flaw is fixed in version 9.5.0.

Do: Upgrade kedro-datasets to 9.5.0 or later, or upgrade PyTorch to 2.6+ so torch.load defaults to weights_only=True. Audit pipelines for use of kedro_datasets_experimental.pytorch.PyTorchDataset and ensure any .pt files it loads come only from trusted sources (shared registries, downloaded checkpoints, and partitioned external stores are the risky paths). Note that supplying weights_only=True via load_args is not a reliable mitigation because load_args are silently dropped in affected versions.

Cisco Identity Services Engine (ISE) contains a denial-of-service vulnerability in its RADIUS feature caused by improper handling of certain RADIUS requests (CWE-119). An unauthenticated, remote attacker can trigger it by sending a crafted RADIUS request directly to an affected ISE node, with no user interaction or credentials required. A successful exploit makes the ISE node unavailable; in single-node deployments, endpoints that have not already authenticated cannot access the network until the node recovers on its own, and even in larger clusters RADIUS-based network access can be disrupted. All organizations running Cisco ISE with the RADIUS/network-access-control feature enabled are potentially affected, particularly single-node deployments. As of now there is no evidence of exploitation in the wild, no CISA KEV listing, and no public proof-of-concept.

Do: Check the Cisco PSIRT advisory for this CVE and upgrade ISE nodes to the fixed releases it specifies. As an interim mitigation, restrict which devices can send RADIUS traffic (UDP 1812/1813) to the ISE policy service nodes using ACLs or firewall rules so only trusted network access devices can reach the service. Operators of single-node deployments should prioritize patching and ensure ISE is monitored so an unavailable node is detected and allowed to recover.

A SQL injection vulnerability (CWE-89) in the web-based management interface of Cisco Secure FMC Software is caused by insufficient validation of user-supplied input. To trigger it, an attacker must already hold a valid account on the device with the Security Approver, Access Admin, or Network Admin role and send a crafted HTTP request to the management interface. A successful exploit allows the attacker to read any data from the database, obtain the session credentials of an authenticated Administrator, and then take actions with administrative privileges on the affected device, effectively escalating a limited role to full admin control. Organizations running Cisco Secure FMC to manage their Cisco Secure Firewall estates are potentially affected. As of this writing, the flaw is not in the CISA KEV catalog and no public proof-of-concept or confirmed in-the-wild exploitation is known.

Do: Upgrade Cisco Secure FMC to the fixed release identified in Cisco's security advisory (fixed version numbers are not included in the available data). Until patched, restrict access to the FMC web management interface to trusted management networks and audit accounts holding the Security Approver, Access Admin, or Network Admin roles. Because successful attacks can harvest administrator session credentials, consider rotating admin passwords and invalidating active sessions on unpatched devices.

CVE-2026-20343 is a missing-authentication flaw (CWE-306) in a critical API of Cisco Secure Firewall Management Center (FMC) Software, the central management console for Cisco Secure Firewall (Firepower) devices. Because the API lacks authentication, an unauthenticated, remote attacker who can reach it can invoke it repeatedly to download files that should be restricted and consume unbounded disk space. A successful exploit can expose sensitive files and fill the FMC's disk until the device becomes unresponsive, causing a denial-of-service condition; the published CVSS 3.1 vector (7.5, AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H) rates availability impact as High, even though the description also includes disclosure of sensitive files. Any organization running an affected FMC release is exposed, particularly where the FMC management interface is reachable from untrusted networks. There is currently no evidence of exploitation: the flaw is not in CISA's KEV catalog and no public proof-of-concept is known.

Do: Upgrade FMC to the fixed release identified in Cisco's security advisory (fixed version numbers are not provided in the source data). Until patched, restrict access to the FMC management interface to trusted management networks or VPN only, verify whether the management port is internet-exposed, and monitor API access logs and disk utilization for signs of repeated unauthenticated calls or abnormal growth.

Cisco Secure FMC (the management center for Cisco Secure Firewall/Firepower deployments) contains a flaw in a specific file download API where user input is not sanitized, allowing arbitrary file downloads from the appliance. An authenticated, remote attacker can trigger it with a single crafted HTTPS request, provided they hold valid credentials for an account with at least the read-only Security Analyst role, so it is primarily an insider or compromised-account risk. A successful exploit exposes any file on the system, which on an FMC can include configuration data, certificates, and other secrets, with no impact on integrity or availability (CVSS 7.7, high confidentiality impact, scope changed). Any organization running Cisco Secure FMC Software is affected; the source data does not specify version ranges. As of now there is no known exploitation, no public proof-of-concept, and the flaw is not in CISA's KEV catalog.

Do: Upgrade FMC to a fixed release as specified in Cisco's security advisory, and in the meantime restrict access to the FMC web interface to trusted management networks. Audit which accounts hold the Security Analyst role or higher, since any such credential is sufficient to exploit, and monitor HTTPS traffic to file download API endpoints for requests containing unexpected file paths.

Cisco Secure Firewall Management Center (FMC) Software contains a flaw (CWE-502) caused by unsecured deserialization of user-controlled data in the web management interface. An attacker who holds valid credentials for an account with at least the read-only Security Analyst role can authenticate and send a crafted HTTP payload, which is saved by the device and can then be executed on the underlying operating system. Successful exploitation yields arbitrary command execution at the root privilege level, giving full control of the FMC appliance and, by extension, the firewall policies it manages. Any organization running Cisco Secure FMC Software is affected, with the practical risk limited to attackers who can reach the management interface and obtain a low-privileged account. As of now, the flaw is not listed in CISA's KEV catalog and no public proof-of-concept is known.

Do: Upgrade FMC to the fixed release identified in Cisco's security advisory. Restrict access to the FMC web management interface to trusted management networks, audit accounts holding the Security Analyst or higher roles, and review saved/scheduled configurations and system logs for unexpected payloads or command execution as indicators of prior exploitation.

CVE-2026-20332 covers improper access control issues (CWE-284) in Cisco Secure Adaptive Security Appliance (ASA), Secure Firewall Threat Defense (FTD), and Secure Firewall Management Center (FMC) software, discovered during Cisco's internal security review and addressed in a dedicated software hardening release. A remote attacker who already holds a low-privileged account or session (CVSS PR:L over the network, no user interaction) can trigger the flaw. Because the attack scope is changed and confidentiality, integrity, and availability impacts are all rated high, successful exploitation crosses a security boundary, giving the attacker high-impact control over the device or access to data it protects. Any organization running affected ASA, FTD, or FMC releases is exposed, though exploitation requires valid low-privileged credentials. No public proof-of-concept or known exploitation exists; the flaw was internally discovered by Cisco and is not yet in CISA's Known Exploited Vulnerabilities catalog.

Do: Upgrade ASA, FTD, and FMC devices to the Cisco software hardening release cited in the Cisco PSIRT advisory (specific fixed version numbers should be confirmed there). Until patched, restrict management-plane and VPN access to trusted users and networks, and audit low-privileged accounts and their permissions for access-boundary gaps. Monitor Cisco PSIRT for updates, as exploitation requires a valid low-privileged credential and no public exploit is currently known.

CVE-2026-20295 is a memory-exhaustion flaw in sftunnel, the protocol Cisco Secure FMC (Firepower Management Center) and Secure FTD (Firepower Threat Defense) devices use to communicate with each other. Because memory resources are improperly managed during sftunnel TLS connection setup, an unauthenticated remote attacker can send crafted sftunnel TLS frames to exhaust the device's available memory. A successful attack does not disclose data or alter configurations, but can crash or stall the affected device, producing a denial-of-service condition; on FTD that can mean loss of firewall/IPS enforcement or management connectivity, and the scope-changed CVSS score (S:C) reflects that exploitation can impact other devices in the FMC-FTD deployment. Any organization running Cisco Secure FMC or Secure FTD software is potentially affected, particularly where sftunnel-capable interfaces are reachable by untrusted networks. As of this analysis there is no known public proof-of-concept, the flaw is not in CISA's Known Exploited Vulnerabilities catalog, and no in-the-wild exploitation has been confirmed.

Do: Monitor the Cisco PSIRT advisory for fixed releases and upgrade FMC and FTD deployments as soon as fixed software is available. Until then, restrict access to the sftunnel service (TCP 8305 by default) so it is reachable only from trusted management/peer networks and not from the internet, and watch for unexplained memory pressure, service restarts, or DoS symptoms on FMC and FTD devices.

Cisco Identity Services Engine (ISE) contains a SQL injection flaw (CWE-89) caused by improper validation of user-supplied input. A remote attacker who already holds at least low-privileged administrative credentials can send a crafted request to an affected device to trigger the injection. A successful exploit allows the attacker to read or modify data in the underlying ISE database, with high integrity impact reflected in the CVSS 7.1 score (C:L/I:H/A:N). Any organization running Cisco ISE is potentially affected, although exploitation requires valid administrative credentials, so internet-wide opportunistic attacks are unlikely. The flaw is not in CISA KEV, no public proof-of-concept is known, and there are currently no confirmed reports of exploitation in the wild.

Do: Check the Cisco PSIRT advisory for CVE-2026-20300 and upgrade ISE to the fixed release it specifies. Until patched, audit and rotate low-privileged administrative accounts, enforce least privilege and MFA on the ISE admin interface, and keep that interface off the internet. Review ISE logs for unexpected or malformed requests to administrative endpoints that could indicate exploitation attempts.

Cisco ISE (Identity Services Engine) contains a SQL injection flaw (CWE-943) in its SXP REST API, caused by insufficient validation of user-supplied input in REST API calls. To trigger it, an attacker must send crafted input to the affected device while holding valid administrative credentials, with the SXP service enabled and at least one SXP connection configured. A successful exploit could let the attacker read or modify data in the underlying ISE database, and in single-node deployments could crash the node, denying network access to endpoints that have not yet authenticated. Any organization running Cisco ISE with SXP/TrustSec in this configuration is affected, though the admin-credential requirement makes insider or compromised-credential scenarios the primary risk. No public proof-of-concept or confirmed in-the-wild exploitation is known, and the flaw is not in CISA's KEV catalog.

Do: Upgrade ISE nodes to the fixed release identified in Cisco's security advisory (not specified in the source data). As interim mitigation, restrict access to the ISE admin/REST API to trusted management networks, disable the SXP service on nodes that do not use it, and audit admin accounts for credential compromise. Monitor Cisco PSIRT for updated fixed-version guidance.

CVE-2026-20250 is a denial-of-service flaw in the Datagram TLS (DTLS) message handling of Cisco Secure Firewall ASA Software and Threat Defense (FTD) Software running on Cisco Secure Firewall 3100 and 4200 Series appliances. The bug stems from improper resource management (CWE-772) when processing certain DTLS messages, causing resources to not be released correctly. An unauthenticated, remote attacker can trigger it by sending a crafted stream of DTLS traffic to the device; a successful exploit causes the firewall to reload, taking down traffic inspection and any VPN services it terminates. Only 3100 and 4200 Series devices running ASA or FTD software are affected. There is currently no known exploitation, no CISA KEV listing, and no public proof-of-concept.

Do: Inventory for Cisco Secure Firewall 3100/4200 Series appliances running ASA or FTD software, especially those terminating remote-access VPN with DTLS, and upgrade to the fixed release listed in Cisco's security advisory. As interim mitigation, restrict untrusted DTLS traffic from reaching the device (e.g., filter DTLS/UDP 443 at the perimeter or disable DTLS in VPN group policies if feasible). Monitor appliances for unexpected reloads, which may indicate exploitation attempts.

CVE-2026-20249 is a denial-of-service vulnerability in the certificate authentication feature of Internet Key Exchange version 2 (IKEv2) in Cisco Secure Firewall Adaptive Security Appliance (ASA) Software and Cisco Secure Firewall Threat Defense (FTD) Software, caused by a logic error (CWE-704) during the certificate authentication phase of IKEv2 connection setup. An unauthenticated, remote attacker can trigger it simply by attempting to establish an IKEv2 VPN connection with a crafted certificate, with no valid credentials or user interaction required. A successful exploit crashes the IKEv2 process, causing the device to reload unexpectedly and producing a denial-of-service condition that disrupts VPN services on the appliance. Any organization running an ASA or FTD VPN gateway with IKEv2 enabled, particularly one exposed to the internet, is potentially affected. As of the advisory data there is no public proof-of-concept, the flaw is not in CISA's KEV catalog, and no in-the-wild exploitation is known.

Do: Upgrade ASA/FTD to a fixed release identified in Cisco's advisory and verified with the Cisco Software Checker (no fixed version numbers are present in the source data). Until patching, limit exposure of UDP 500/4500 to known VPN peers, disable IKEv2 remote access or certificate-based authentication where it is not required, and confirm high-availability/failover is configured to absorb a reload. Monitor devices for unexpected IKEv2 process crashes or reloads and track Cisco PSIRT for updates on exploitation status.

Cisco Identity Services Engine (ISE) contains a SQL injection vulnerability (CWE-89) caused by improper validation of user-supplied input. An unauthenticated, remote attacker can trigger it by sending a crafted request to an affected device, requiring no credentials or user interaction. A successful exploit allows the attacker to modify data in the underlying ISE database; Cisco's CVSS scoring (C:N/I:H/A:N) frames this as an integrity-only impact, with no confidentiality or availability loss. Any organization running Cisco ISE is affected, though the device is typically deployed on internal networks rather than exposed to the internet. As of now there is no known exploitation, no public proof-of-concept, and the flaw is not listed in CISA's Known Exploited Vulnerabilities catalog.

Do: Check Cisco's security advisory for CVE-2026-20247 and upgrade ISE to the fixed release it specifies for your deployment. Until patching, restrict network access to ISE web-based interfaces (admin and portal) to trusted management networks and monitor for unexpected database changes. Confirm your ISE node inventory, since distributed deployments mean every node in the cluster must be remediated.

Cisco Secure Firewall ASA and FTD software contain a memory leak (CWE-401) in the handling of EIGRP update messages, caused by improper resource management. An unauthenticated attacker positioned adjacent to the device (on a segment where EIGRP is configured) can send crafted EIGRP updates at a high rate, exhausting memory until the firewall reloads unexpectedly. A successful attack yields only a denial of service (CVSS availability impact, no confidentiality or integrity loss), but repeated exploitation could keep the firewall crashing, disrupting all traffic it inspects. Only deployments of Cisco ASA or FTD software with EIGRP enabled are affected, and the attacker must be able to reach an EIGRP-enabled interface. Exploitation status: no reports of in-the-wild abuse, no public proof-of-concept, and the flaw is not in CISA KEV.

Do: Upgrade ASA and FTD to the fixed releases listed in the Cisco security advisory. As interim mitigation, enable EIGRP neighbor authentication and/or interface access lists restricting which hosts can send EIGRP packets to the firewall, and disable EIGRP where it is not required. Check whether EIGRP is configured on any interface and monitor device memory utilization for unexplained growth.

Cisco Secure Firewall ASA and FTD Software improperly rate-limit syslog message 419002 (the message generated when the device detects a TCP SYN flood), a flaw tracked as CWE-835 in the system rate-limiting process. An unauthenticated, remote attacker triggers it simply by sending a flood of TCP SYN packets to the affected device. The result is high CPU utilization, performance degradation, and an availability-only denial-of-service condition (CVSS 8.6, Scope changed, Availability: High; confidentiality and integrity are not impacted). Any organization running Cisco ASA or FTD Software is potentially affected, particularly devices with internet-facing interfaces, although exact affected version ranges were not enumerated in the source data. As of this analysis there are no reports of in-the-wild exploitation, no public proof-of-concept, and the flaw is not on the CISA KEV list.

Do: Consult the Cisco PSIRT advisory for CVE-2026-20154 and upgrade ASA/FTD to the fixed software releases it lists for your train. As interim mitigations, rate-limit or filter inbound TCP SYN traffic upstream (e.g., control-plane policing/upstream filtering), reduce unnecessary internet exposure of ASA/FTD interfaces, and monitor CPU usage ('show cpu usage') and syslog 419002 volume for signs of abuse. Inventory all internet-facing ASA/FTD deployments now, since no patch details are in the current data.

CVE-2026-20135 is a double-free (improper buffer management, CWE-415) flaw in the TLS 1.3 implementation of Cisco Secure Firewall Threat Defense (FTD) Software. An unauthenticated, remote attacker can trigger it by sending a crafted TLS 1.3 packet to any TLS 1.3-enabled listening socket, covering both data traffic and user-management traffic, and exploitation can succeed before or after connection authentication. A successful attack crashes the LINA process, forcing the device to reload and causing a denial of service; confidentiality and integrity are not impacted. Any organization running FTD with TLS 1.3 enabled on exposed listeners is affected. No exploitation in the wild or public proof-of-concept is currently known, and the flaw is not in CISA's KEV catalog.

Do: Consult the Cisco security advisory (assigned by [email protected]) for fixed FTD releases and upgrade as soon as fixed software is available for your release train. As interim mitigation, disable TLS 1.3 on listening sockets where possible or restrict access to TLS 1.3-enabled interfaces (especially user-management traffic) to trusted networks only. Check devices for unexpected reloads or LINA process crashes, which may indicate exploitation attempts.

CVE-2026-92808 is a server-side request forgery vulnerability in the UnifiedLogin service of Altium Enterprise Server. An unauthenticated network attacker can trigger the server to issue HTTP requests to destinations of the attacker's choosing, including internal-only services; one such internal service exposes server configuration and stored credential material to any request that appears to originate locally, a check the forged requests satisfy. Using the harvested credentials, the attacker can obtain an administrative session and achieve full compromise of the server and all of its services. Only on-premises/self-hosted deployments are affected; the vulnerable endpoint is disabled in Altium 365 cloud mode. There is no CISA KEV listing and no known public proof-of-concept or reported exploitation to date, though the maximum CVSS 4.0 score of 10.0 reflects trivially exploitable, unauthenticated remote conditions.

Do: Apply the vendor's patched release for Altium Enterprise Server as soon as one is available (no fixed version is stated in the current data) and keep the server off the public internet, reachable only via VPN or IP allowlisting. As interim mitigation, prevent the server process from reaching the internal configuration/credential service, add authentication or network controls around that internal endpoint, and review logs for unauthenticated requests to UnifiedLogin followed by unexpected outbound connections. If compromise is suspected, rotate all credentials stored on the server and invalidate active administrative sessions.

HP has disclosed code-injection flaws (CWE-94) in HP Advance server software that, under certain conditions, can lead to elevation of privilege, remote code execution, or arbitrary file writes on the server hosting the product. The CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:N/UI:N) indicates the flaws are triggerable remotely over a network with low attack complexity, no privileges required, and no user interaction. A successful attacker could gain highly confidential read access on the server and achieve high confidentiality and integrity impact on subsequent (connected) systems, enabling privilege escalation, code execution, or malicious file placement. Any organization running HP Advance, an enterprise print/output-management platform typically hosted on an on-premises server, is affected. No public proof-of-concept is known and the issue is not listed in CISA's KEV, so exploitation has not been confirmed.

Do: Upgrade HP Advance to the fixed release identified in the HP security bulletin, since the advisory data available here does not specify affected or fixed version numbers. Restrict network access to the Advance server (internal or VPN only, not directly internet-exposed) and check server logs for unauthenticated requests or unexpected file writes. Given the high confidentiality and subsequent-system impact ratings, also review accounts and adjacent systems reachable from the Advance host after patching.

Apache NiFi Registry 0.4.0 through 2.11.0 is vulnerable to path manipulation (CWE-22) in the default file persistence provider, which builds storage paths for extension bundles from the group, artifact, and version coordinates in an uploaded NAR manifest without rejecting parent-directory names, while its path-containment check compares an unnormalized resolved path. An authenticated user authorized to write and delete bundles in a bucket can upload a NAR with a crafted manifest, causing file system operations outside the file persistence directory. The attacker gains the ability to write to or delete files elsewhere on the host, producing high integrity and availability impact, though the CVSS score indicates no direct confidentiality impact. All NiFi Registry deployments in the 0.4.0-2.11.0 range are affected where any account holds bundle write/delete permissions, since exploitation requires valid low-privileged credentials. No public proof of concept, CISA KEV listing, or confirmed in-the-wild exploitation is currently known.

Do: Upgrade all NiFi Registry instances to 2.12.0, which rejects parent-directory coordinates and requires a normalized path to remain a strict child of the storage root. Until upgraded, restrict write and delete permissions for extension bundles to fully trusted accounts and review buckets for NARs whose group, artifact, or version coordinates contain parent-directory sequences. Check the host filesystem around the file persistence directory for unexpected files or missing data that could indicate prior abuse.