A security issue in MongoDB Server allows an authenticated user with elevated internal privileges to bypass a disabled feature gate in the applyOps command by specifying an internal replication mode value that was not intended to be client-selectable. This bypass enables execution of container operations that are disabled by default in production configurations, allowing direct storage-engine writes to arbitrary internal storage tables. The authorization check for these operations validates only the operation's namespace, not the actual storage target, enabling writes to unrelated internal metadata or other collections' data.

A use-after-free security issue exists in the server's query execution memory tracking subsystem. An authenticated user with read privileges can trigger a write to freed heap memory through a sequence of standard database commands, leading to server process crash or potential memory corruption. No user interaction is required.

In MongoDB, insufficient validation of shard key values during document insertion allowed authenticated users to store documents with specially crafted, operator-shaped objects as shard key values in sharded collections. When change stream events for such documents were processed with the updateLookup full document mode, the crafted values were embedded into internal post-image lookup queries without proper sanitization, causing them to be interpreted as query operators rather than literal equality values. This could result in change stream consumers receiving incorrect post-image documents or encountering non-resumable fatal errors.

An internal aggregation expression in MongoDB Server was incorrectly registered as accessible to any authenticated user rather than being restricted to internal cluster operations. By crafting a malformed index specification within this expression, an authenticated user with read-only privileges could trigger an assertion failure in the index key generation code path. In certain build configurations, this assertion failure results in termination of the mongod process, causing a denial of service to all connected clients.

A flaw in MongoDB's JSON Schema validation error generation code allows an authenticated user with readWrite privileges to crash the mongod server. When a BSON document containing an array with a malformed numeric field name fails a $jsonSchema items type constraint, the error generation path performs unsafe numeric conversion on the user-controlled field name without proper exception handling, resulting in an uncaught exception that terminates the server process. This is possible because incoming wire protocol BSON validation does not enforce that array element field names are valid, in-range numeric indices.

A security issue was discovered in MongoDB where an authenticated user with readWrite privileges could crash the mongod server process. By specifying a custom WiredTiger storage configuration option with an incompatible value during collection creation, a user could cause a type confusion in the storage engine layer. When documents were subsequently read from the misconfigured collection, the resulting mismatch in expected data format led to corrupted memory interpretation and a server crash. The crafted collection configuration persists across restarts, requiring manual operator intervention to remediate.

A race condition in MongoDB server's text index query parsing can cause a heap use-after-free read when handling upsert retry paths. Under certain concurrent index lifecycle operations, a raw pointer to internal text index metadata may be dereferenced after the underlying structures have been freed, leading to a server crash. An authenticated user with readWrite privileges can trigger this condition through specific concurrent text-search and index management operations, resulting in denial of service for all connected clients. This

A security issue exists in MongoDB's 2dsphere index key generation that can cause a server crash due to a null pointer dereference. When a specially crafted GeoJSON document is inserted into a collection with a 2dsphere index, an inconsistency in geometry parsing can leave an internal object in an invalid, partially initialized state. During subsequent index key generation, access to this improperly initialized object results in a null pointer dereference that terminates the mongod process. An authenticated user with write access can use this to cause a denial of service.

A security issue exists in MongoDB server's JSON Pointer parser used during $jsonSchema query filter processing. When a find command includes a specially crafted $jsonSchema filter field, the parser processes the input without enforcing adequate limits on iteration count or total allocation size, resulting in significant memory amplification. Under concurrent request load, the cumulative memory consumption can exhaust available heap memory, causing the server's out-of-memory handler to terminate the mongod process and deny service to all connected clients.

A security issue exists in MongoDB's LDAP authorization integration where pooled LDAP connections can retain stale authentication identities after user authentication under certain configurations. Subsequent authorization queries may execute under an unintended LDAP identity rather than the expected one. This can result in incorrect role assignments based on the LDAP directory's access control configuration, potentially allowing an authenticated user to acquire elevated privileges that were not intended by the deployment's authorization policy.

The $regexFindAll expression can be used by an authenticated user who can run aggregation pipeline stages to crash a MongoDB server (mongod). Under certain specific conditions the  regex match can start in the middle of a multi-code-unit character, triggering an assertion during query execution.

An information disclosure vulnerability has been identified in Omada Controller.  An API endpoint intended for Controller initialization remains accessible after completion and may disclose account-related information to unauthenticated remote users.  Successful exploitation may allow an attacker to remote query the affected endpoint that may facilitate user enumeration and subsequent attacks targeting administrative accounts.

mfish-nocode-pro v1.0.0 was discovered to contain a SQL injection vulnerability in the tableName parameter at /sys/dbConnect/data. This vulnerability allows attackers to access sensitive database information via a crafted SQL statement.

Movie_Recommend v1.0.0 was discovered to contain a SQL injection vulnerability in the sort parameter at /loadingmore. This vulnerability allows attackers to access sensitive database information via a crafted SQL statement.

UC Browser for Android (package com.UCMobile.intl, version 13.7.8.1314) contains a Universal Cross-Site Scripting vulnerability that allows an attacker to execute arbitrary JavaScript in the context of any origin. An attacker hosts a specially crafted URL on a UC-owned domain (via a reflected XSS) that leverages the browser's internal JavaScript bridge to register a deferred callback, navigate the tab to a victim site, and then execute attacker-controlled code on that site when a login dialog is dismissed.

AshAi exposes Ash read actions to language-model tool calls. The read tool accepts an aggregate result type (min, max, sum, avg) that builds an ad-hoc Ash.Query.Aggregate over a named field and returns its raw value. Ash field policies redact forbidden fields on returned records (replacing them with %Ash.ForbiddenField{}), but that redaction does not apply to aggregate values. A tool caller could therefore read a field the calling actor's field policies forbid by requesting it as an aggregate; min/max in particular return an actual field value. This includes fields that are public? true but restricted per-actor by a field policy, such as sensitive PII. The tool's existing check only required the field to be public, which is a separate axis from per-actor field-policy authorization. The fix authorizes the aggregated field against the resource's field policies, so aggregating over a field the actor may not see is refused or scoped to the rows where it is visible. This issue affects ash_ai: from 0.1.0 before 1.0.3.

AshLua exposes Ash read actions to Lua scripts run through an eval action. A read call accepts an operation (list, min, max, first, sum, avg) that builds an ad-hoc Ash.Query.Aggregate over a named field and returns its raw value. Ash field policies redact forbidden fields on returned records (replacing them with %Ash.ForbiddenField{}), but that redaction does not apply to aggregate values. A script could therefore read a field the calling actor's field policies forbid by requesting it as an aggregate instead of as a field. This includes fields that are public? true but restricted per-actor by a field policy, such as sensitive PII. The prior hardening only enforced the exposed-field allow-list (field visibility), which is a separate axis from per-actor field-policy authorization. The fix authorizes the aggregated field against the resource's field policies, so aggregating over a field the actor may not see is refused or scoped to the rows where it is visible. This issue affects ash_lua: from 0.1.0 before 0.2.2.

Apache Airflow FAB provider versions 3.7.3 through 3.8.0 do not validate the issuer or audience of Azure AD `id_token`s during OAuth login. Deployments are affected only when the FAB auth manager is configured with Azure AD as an OAuth provider. Because the signing keys are fetched from Microsoft's **multi-tenant** JWKS endpoint, an `id_token` minted in *any* Azure tenant — including one the attacker creates — passes signature verification, and the username and role assignments are then read from that attacker-controlled token. Anyone able to register an Azure tenant can therefore authenticate to the Airflow UI with no prior access to the deployment. The fix for **CVE-2026-59243** was incomplete, and this advisory closes the remaining gap: that fix made the provider verify the `id_token` signature, but did not add issuer or audience checks. Operators who already applied the CVE-2026-59243 fix are **still affected and must upgrade again** — 3.7.3 is the release that shipped that fix, so every version containing it falls inside this affected range. Upgrade to apache-airflow-providers-fab `3.8.1` or later.

An authenticated stored cross-site scripting (XSS) vulnerability in the Column Management component of ClassCMS 1CMS v5.6 allows attackers to execute arbitrary web scripts or HTML via injecting a crafted payload into the title field.

NVIDIA Triton Inference Server for Linux contains a vulnerability where an attacker could abuse missing authorization. A successful exploit of this vulnerability might lead to information disclosure, data tampering, and denial of service.

A improper access control vulnerability in Fortinet FortiSandbox 5.0.0 through 5.0.5, FortiSandbox 4.4.0 through 4.4.8, FortiSandbox Cloud 5.0.4 through 5.0.5, FortiSandbox PaaS 5.0.4 through 5.0.5 may allow attacker to access sensitive information via crafted HTTP requests.

An improper access control vulnerability in Fortinet FortiManager 7.6.0 through 7.6.4, FortiManager 7.4.0 through 7.4.10, FortiManager 7.2 all versions, FortiManager Cloud 7.6.2 through 7.6.4, FortiManager Cloud 7.4.1 through 7.4.10, FortiManager Cloud 7.2 all versions may allow an administrator to bypass the approval process for workflow sessions via crafted HTTP or HTTPs requests.

A vulnerability in the Unified Extensible Firmware Interface (UEFI) Shell implementation of Cisco UCS Servers and UCS-based appliances could allow an authenticated attacker with valid credentials for a user account with the role of user or admin or an unauthenticated attacker with physical access to an affected device to bypass UEFI Secure Boot validation checks and execute unauthorized software. This vulnerability is due to the availability of memory write commands in the UEFI Shell while UEFI Secure Boot is enabled on a device. An attacker could exploit this vulnerability by selecting the UEFI Shell boot option at boot time and using available shell commands to modify UEFI memory variables. A successful exploit could allow the attacker to manipulate the preboot environment, overwrite UEFI Secure Boot-related memory values, and execute unauthorized software on the affected device.

NVIDIA Triton Inference Server for Linux contains a vulnerability where an attacker could cause excessive iteration. A successful exploit of this vulnerability might lead to denial of service.

A malicious webpage could repeatedly trigger external URL schemes, causing system prompts or external application launches. This could make Firefox for iOS temporarily unusable until the page is closed. This vulnerability was fixed in Firefox for iOS 155.1.