🚨 CVE-2026-57131
PraisonAI is a multi-agent teams system. Prior to 4.6.58, praisonai.jobs.server.create_app mounts praisonai.jobs.router.create_router under /api/v1/runs without authentication or per-job authorization. Network clients can submit attacker-controlled prompts and agent configuration, list and read jobs, stream results, and cancel or delete other jobs, exposing service credentials and connected tool capabilities to unauthorized agent execution. This vulnerability is fixed in 4.6.58.

🎖@cveNotify

🚨 CVE-2026-57132
PraisonAI is a multi-agent teams system. Prior to 4.6.62, setting PRAISONAI_CALL_AUTH to disabled makes verify_token accept requests to /api/v1/agents/{id}/invoke without CALL_SERVER_TOKEN authentication. Deployments that use the application's advertised opt-out can expose registered agents and their connected tools or private context to unauthenticated invocation. The vulnerability is fixed in 4.6.62.

🎖@cveNotify

🚨 CVE-2026-55416
Pimcore is an Open Source Data & Experience Management Platform. Prior to 11.5.19, 12.3.10, and 2026.1.6, an authenticated user with reports_config permission can place attacker-controlled SQL fragments in the sql, from, where, and groupby fields of a Custom Reports configuration processed by bundles/CustomReportsBundle/src/Tool/Adapter/Sql.php. The buildQueryString() method concatenates these values into a database query, while a blacklist omits dangerous constructs such as additional data-manipulation statements, comments, subqueries, and multiple statements. The getData() method also previously interpolated offset and limit values into a LIMIT clause without integer casting. Executing the configured report reaches fetchAllAssociative() with the constructed query and can disclose, modify, or delete arbitrary database data. This issue is fixed in versions 11.5.19, 12.3.10, and 2026.1.6.

🎖@cveNotify

🚨 CVE-2026-91021
Trilium Notes, version v0.103.0 and earlier, contains a stored cross-site scripting (XSS) vulnerability in the share renderer for webView notes due to improper HTML escaping of user-controlled #webViewSrc values. This vulnerability allows attackers with note-authoring privileges to inject arbitrary JavaScript that executes for any user who opens the shared note, including administrators.

🎖@cveNotify

🚨 CVE-2026-54246
Skipper is an HTTP router and reverse proxy for service composition. Prior to 0.27.13, the routesrv component serves cluster-wide control-plane data without application-layer authentication through /routes, /routes/{zone}, /swarm/redis/shards, and /swarm/valkey/shards. The handlers registered in routesrv/routesrv.go, eskipBytes.ServeHTTP in routesrv/eskipbytes.go, RedisHandler.ServeHTTP in routesrv/redishandler.go, and ValkeyHandler.ServeHTTP in routesrv/valkeyhandler.go only restrict HTTP methods and do not authenticate callers. A compromised or otherwise attacker-controlled pod that can reach routesrv on the Kubernetes cluster network can bypass its namespace-scoped RBAC and read Ingress and RouteGroup backend URLs, filter-chain and OAuth or OIDC path details, and Redis or Valkey shard addresses across namespaces. NetworkPolicy can restrict reachability but does not remove the missing-authentication condition, and the established impact is confidentiality exposure rather than direct integrity or availability loss. This issue is fixed in version 0.27.13.

🎖@cveNotify

🚨 CVE-2026-54247
Skipper is an HTTP router and reverse proxy for service composition. Prior to 0.26.22, Handler in dataclients/kubernetes/admission/admission.go passes the body of requests to the Kubernetes admission endpoint at :9443/admission directly to io.ReadAll(r.Body) without a size limit. An attacker with in-cluster network access and a valid Kubernetes client certificate can send a very large body that causes unbounded memory allocation and an out-of-memory termination of the Skipper process. The disruption is limited to Ingress and RouteGroup admission rather than pod creation or unrelated admission controllers, and Kubernetes normally restarts the process. This issue is fixed in version 0.26.22.

🎖@cveNotify

🚨 CVE-2026-65838
Skipper is an HTTP router and reverse proxy for service composition. Prior to 0.27.35, the opaAuthorizeRequestWithBody filter in filters/openpolicyagent/openpolicyagent.go can allow an oversized declared Content-Length request to bypass a deny-on-presence Rego policy because ExtractHttpBodyOptionally leaves OPA with an empty parsed_body while forwarding the complete request body upstream. This incomplete remediation of CVE-2026-50197 affects deployments that authorize request-body content and exceed -open-policy-agent-max-request-body-size, which defaults to 1 MB. Policy logic that does not reject input.attributes.request.http.truncated_body can therefore fail open and permit a forbidden payload to reach the protected service, while small bodies and the previously fixed chunked-body case are evaluated normally. This issue is fixed in version 0.27.35.

🎖@cveNotify

🚨 CVE-2026-76081
ZITADEL is an open source identity management platform. Prior to version 4.16.0, a bug in how ZITADEL updates permissions when multiple project roles are deleted at the same time can cause some user permissions to be missed. This issue specifically affects User Grants on Granted Projects (projects shared between different organizations), potentially allowing users to keep access rights that were supposed to be completely removed. This issue has been fully resolved in version 4.16.0. There are no configuration workarounds. Upgrading to a patched version is the only way to trigger the automatic cleanup migration. Those who cannot upgrade immediately should manually review user permissions specifically for Granted Projects where multiple roles were recently deleted.

🎖@cveNotify

🚨 CVE-2026-91819
Affected versions of MISP rely on CakePHP request-method override processing in a way that can disable CSRF and form-security validation.

CakePHP honors a _method field or X-HTTP-Method-Override header by rewriting the effective request method. For override values outside the normal write verbs POST, PUT, PATCH, and DELETE, CakePHP also clears the parsed request body. MISP’s security component then determines whether to perform _validatePost() and _validateCsrf() based on whether request data remains. With a value such as:


_method=GET

the body becomes empty before those checks run, so both protections are skipped. A cross-site form containing only that override can therefore reach actions whose parameters are taken from the URL rather than the request body

Version affected: ≤2.5.45

🎖@cveNotify

🚨 CVE-2026-76159
Incorrect Permission Assignment for Critical Resource in the configuration loader of Duplicati for Windows versions before v2.4.0.0 allows a local low-privileged attacker to escalate privileges to NT AUTHORITY\SYSTEM via an attacker-controlled preload.json file.

🎖@cveNotify

🚨 CVE-2026-91825
Affected versions of MISP fail to authorize a submitted sharing group in a specific event-edit path.


The vulnerable logic checked whether the acting user could use a sharing_group_id only when the request explicitly supplied distribution = 4. If the attacker instead omitted distribution but supplied a different sharing_group_id, that authorization branch was skipped. Later, MISP’s field-recovery logic restored the existing event distribution from storage. For events already configured with sharing-group distribution, the unauthorized sharing-group ID could therefore be saved.


The fix adds authorization checks in both the controller and Event::_edit() whenever a non-empty sharing_group_id is supplied without distribution. The model now calls SharingGroup::checkIfAuthorised() before persisting the change.



Version affected: ≤2.5.45

🎖@cveNotify

🚨 CVE-2026-91846
Affected versions of MISP allow a collection element to be created from a bare UUID without consistently checking whether the acting user is allowed to access the referenced object.


The commit explains that collection elements themselves only store UUIDs, while the collection view later resolves those UUIDs into their underlying objects. Before this fix, the generic add() path could therefore persist a UUID for an Event or Galaxy Cluster that the caller could not normally read. The patch explicitly notes that this made collections a way to reference another organisation’s private data and had caused disclosure of organisation-only events in the beta collection view.


The fix centralizes authorization in __assertCanUseElements(). Event UUIDs are validated through Event::fetchSimpleEvent() under the current user’s ACL, while Galaxy Cluster UUIDs are checked through GalaxyCluster::fetchGalaxyClusters(). The check is applied both to the CRUD add() path and to addElementToCollection().

Version affected: ≤2.5.45

🎖@cveNotify

🚨 CVE-2026-91851
Affected versions of MISP incorrectly filter dashboard templates that are restricted to a specific permission flag.


DashboardsController::listTemplates() allowed a template when either:



 - its restrict_to_permission_flag matched one of the current user’s permission flags, or

 - restrict_to_permission_flag equaled integer 0






However, restrict_to_permission_flag is a varchar. MySQL therefore performed numeric coercion when comparing the column against integer 0. Strings such as perm_site_admin convert numerically to zero, making expressions such as perm_site_admin = 0 evaluate true and causing the “unrestricted” branch to match permission-restricted templates as well.



Version affected: ≤2.5.45

🎖@cveNotify

🚨 CVE-2026-91857
Affected versions of MISP expose several state-changing controller actions without restricting them to POST.


The affected actions are:



 - EventReportsController::purgeUnusedPictures()

 - NoticelistsController::enableNoticelist()

 - ServersController::removeOrphanedCorrelations()

 - WorkflowsController::rebuildRedis()






The patch adds allowMethod(['post']) to each action, preventing them from being triggered through ordinary GET requests.


For purgeUnusedPictures(), the corresponding UI previously used $.get(). The fix converts that request to POST and supplies X-CSRF-Token, while the controller enables header-only CSRF validation for that AJAX action.


Because GET requests can be induced cross-origin through links, images, redirects, or navigation, accepting GET for these state-changing operations can let an attacker trigger them using the authenticated victim's session.

Version affected: ≤2.5.45

🎖@cveNotify

🚨 CVE-2026-91859
Affected versions of MISP can record incorrect access-log data for requests that terminate in an exception.


Because CakeErrorController extends AppController, exception rendering runs the application startup path a second time. As a result, __accessMonitor() calls AccessLog::logRequest() twice for one HTTP request. The second deferred writer measures the error-controller execution instead of the original request and can overwrite the row created by the first pass. The corrupted fields include request duration, SQL query count, memory usage, and potentially the recorded query log.


The bug was masked because the same model instance retained the ID of the first saved row, causing the later save to issue an UPDATE rather than insert an obvious duplicate row.

Version affected: ≤2.5.45

🎖@cveNotify

🚨 CVE-2026-92002
Affected versions of MISP use Redis to throttle repeated authentication-failure log entries. The intent is to avoid excessive duplicate logs while still recording failed authentication activity.


However, User->setupRedis() returns false when Redis cannot be reached. The vulnerable _shouldLog() logic only returned true when a Redis instance existed and no throttle key was present. Therefore, when Redis was unavailable, the function did not allow the log write at all, effectively silencing authentication-failure logging for the duration of the outage.

Version affected: ≤2.5.45

🎖@cveNotify

🚨 CVE-2026-92003
Affected versions of MISP do not consistently apply the existing authentication-failure logging throttle.


Two API authentication failure branches wrote directly to the Log model:

 - API requests with no authentication key;
 - requests supplying an API key with an incorrect length




Unlike other authentication failures, these paths bypassed _shouldLog(), so every request could create another durable auth_fail entry.

Version affected: ≤2.5.45

🎖@cveNotify

🚨 CVE-2026-48737
pyLoad is a free and open-source download manager written in Python. Prior to 0.5.0b3.dev101, is_global_address in src/pyload/core/utils/web/check.py relies on Python's global-address classification without examining IPv4 destinations embedded in 6to4 or NAT64 IPv6 addresses. A low-privileged user can submit an IPv6 literal through parse_urls to the pre-resolution is_global_host guard. Because host_to_ip is pinned to AF_INET, that guard does not evaluate a hostname's AAAA result. Separately, curl resolves hostnames before the pycurl PREREQFUNC in src/pyload/core/network/http/http_request.py applies the same vulnerable is_global_address check to the actual connection address, so a transition-form AAAA result can be permitted even when it terminates at an embedded loopback, private, CGNAT, or link-local IPv4 address. Exploitation requires the pyLoad host to route the applicable transition mechanism, including 6to4 on affected Python 3.9 through 3.11 deployments or NAT64 on a network with a NAT64 gateway. Successful exploitation can enable internal-network reconnaissance, timing-based confirmation, limited service disruption, or cloud metadata disclosure where the wrapped address is routable. This issue is fixed in version 0.5.0b3.dev101.

🎖@cveNotify

🚨 CVE-2026-48987
pyLoad is a free and open-source download manager written in Python. Prior to 0.5.0b3.dev101, EventManager in src/pyload/core/managers/event_manager.py appends a Client object to the clients list for each unique uuid submitted to the authenticated getEvents API endpoint, but get_events does not invoke the available clean method to remove inactive clients. An authenticated user can repeatedly submit unique UUID values, causing retained client objects and process memory to grow without bound even after requests stop. The resulting memory exhaustion can trigger an operating-system out-of-memory termination of pyLoad or host-wide instability and denial of service. This issue is fixed in version 0.5.0b3.dev101.

🎖@cveNotify

🚨 CVE-2023-54397
Tornado before 6.3.3 contains an HTTP request smuggling vulnerability due to improper parsing of Content-Length headers accepting non-standard characters. Attackers can send crafted HTTP requests with these characters to bypass proxy validation and smuggle requests when deployed behind certain proxies.

🎖@cveNotify

🚨 CVE-2024-14029
Tornado before 6.4.1 ignores duplicate Transfer-Encoding: chunked headers, treating requests as having no message body and parsing the chunked body as a subsequent request. Attackers can exploit this inconsistency when Tornado is deployed behind proxies to perform HTTP request smuggling, enabling access control bypass, cache poisoning, or connection desynchronization.

🎖@cveNotify