GPAC versions prior to 26.07.0 have an uncontrolled recursion vulnerability in the dasher_gather_deps() function within src/filters/dasher.c. This flaw fails to detect cyclic stream dependency chains when processing MPEG-2 TS files with the MP4Box tool using the -dash option. An attacker can exploit this by supplying a crafted MPEG-2 TS file that causes stack exhaustion and crashes the process, resulting in a denial of service.
The REV.exe module distributed with Farmasimplex contains embedded OAuth client credentials for integration with the IZS platform. An attacker who gains access to this executable could extract these credentials and potentially reuse them to access integration resources or manipulate associated information, depending on the privileges and authentication controls in place. This vulnerability is identified as CVE-2026-123051 and has a medium severity rating.
A code injection vulnerability exists in react-native-webview versions 13.8.0 through 17.1.0. This flaw allows attackers to execute arbitrary JavaScript by injecting unescaped backticks or ${} sequences in the injectedJavaScriptObject string values. Exploitation can lead to unauthorized access to the page DOM, non-HttpOnly cookies, and the ReactNativeWebView message bridge.
GPAC versions prior to 26.07.0 have a heap-based buffer overflow vulnerability in the gf_mpd_resolve_url() function located in src/media_tools/mpd.c. This flaw allows attackers to craft DASH manifests with segment templates that expand beyond allocated heap buffers due to unchecked strcat() calls. Exploiting this can cause MP4Box to crash, resulting in a denial of service condition.
GPAC versions prior to 26.07.0 have an uninitialized pointer access vulnerability in the GetESD() function within src/isomedia/track.c. This flaw can be triggered by a crafted MP4 file containing an rvcc box with predefined_rvc_config set to 0 and lacking a meta item index. The vulnerability leads to an invalid memory read due to a strcmp() call on an uninitialized mime_type pointer, causing the MP4Box -add operation to crash.
A vulnerability exists in react-native-webview versions 13.7.0 through 17.1.0 on Windows where the WebView2 component's new window request handler allows launching URIs with unrestricted schemes. This occurs because the handler passes requested URIs to LaunchUriAsync without verifying the URI scheme or requiring user initiation. Consequently, malicious web content can open new windows or links using schemes like ms-settings: to invoke Windows protocol handlers without triggering the usual external-protocol confirmation prompt.
GPAC versions before 26.02.0 have an out-of-bounds read vulnerability in the m2tsdmx_send_packet() function within src/filters/dmx_m2ts.c. This flaw is caused by double-skipping the VC-1 start-code offset, allowing a crafted MPEG-2 TS VC-1 stream to trigger an underflow of the remaining length. This results in memcpy() reading beyond the intended buffer, leading to a crash of MP4Box during import.
GPAC version 26.02.0 has a NULL pointer dereference vulnerability in the dasher_configure_pid() function within src/filters/dasher.c. This flaw occurs because the code dereferences the pointer ds2->set without verifying if it is NULL. An attacker can exploit this by providing a specially crafted MPEG-2 TS file processed with MP4Box -dash, leading to a crash of MP4Box and causing a denial of service.
Camunda versions 7.24.0 before 7.24.14 and 7.23.0 before 7.23.17 contain an XML External Entity (XXE) injection vulnerability in the camunda-external-task-client component. This vulnerability arises because the defaultDocumentBuilderFactory() method does not disable external entity resolution when parsing XML serialized object variables. An attacker who can write a process variable may exploit this to cause arbitrary file reads and out-of-band HTTP requests from worker JVMs, potentially exposing sensitive credentials and private keys.
GPAC versions before 26.02.0 have a use-after-free vulnerability in the function gf_filter_pid_set_args_internal(). This flaw occurs when a stale pointer remains after gf_props_reset_single() frees a parsed PID property value. An attacker can exploit this by supplying a crafted filelist input to MP4Box with the -add option, causing the application to access freed memory and crash.
The CeliachiaRL integration component used in Farmasimplex shares OAuth credentials across multiple customer installations. This shared credential usage allows an attacker who obtains credentials from one installation to potentially access or modify data in other installations. The vulnerability could lead to unauthorized access to sensitive information or unauthorized modification of integration data.
Multiple Farmasimplex healthcare integration modules fail to properly validate TLS certificates. This vulnerability allows attackers who can intercept network traffic to impersonate trusted healthcare services, enabling man-in-the-middle attacks that may disclose or modify sensitive data.
phpMyFAQ versions 4.0.12 through 4.0.19, all 4.1.x versions before 4.1.12, and 4.2.0 pre-releases before 4.2.0-RC.1 contain a stored cross-site scripting (XSS) vulnerability. This flaw exists in the startpage.twig template where top-ten FAQ titles are rendered using Twig's raw filter. Authenticated users with FAQ edit permissions can inject malicious scripts by saving questions containing crafted double quotes and event handlers, causing the script to execute for every visitor of the start page.
Scalar scalar-ninja versions before 0.3.3 have a cross-site scripting (XSS) vulnerability in the get_scalar_api_reference() function. This flaw allows attackers to inject malicious scripts via crafted OpenAPI documents containing a </script><script> sequence, which are embedded unescaped inside an inline script element. Users viewing the affected API documentation may have scripts executed in their browser context.
Farmasimplex retrieves a CIP News file over HTTP and executes embedded SQL statements with administrative privileges against the local FarmaCIP database. This allows an attacker who can tamper with network communications or compromise vendor infrastructure to inject arbitrary SQL commands, potentially leading to unauthorized database access, modification, deletion, and possibly operating system command execution.
53AIHub versions before 0.5.5 have a missing authentication vulnerability that allows unauthenticated attackers to read files by accessing specific API endpoints without authentication. Attackers can enumerate sequential numeric IDs to access raw knowledge-base files and stored file versions belonging to any enterprise tenant.
Eclipse Zenoh's zenoh-pico implementation versions 0.7.2-rc up to but excluding 1.10.2 contain a vulnerability where the Zenoh identifier (ZID) length is not validated during timestamp deserialization. This can lead to out-of-bounds writes and reads, causing memory corruption and denial of service. Remote attackers can exploit this by sending crafted timestamp-bearing messages. The issue is fixed in version 1.10.2.
SQLBot versions 1.10.0 through 1.10.3 contain a vulnerability where the PostgreSQL dangerous-function denylist is not properly applied due to a mismatch in identifier strings. This allows authenticated users to bypass the check_sql_read() function and execute dangerous PostgreSQL functions such as pg_read_file, pg_write_file, lo_import, and lo_export, potentially enabling reading or writing files on the database host.
Emitter versions through 3.1 have a privilege escalation vulnerability due to insecure key encryption. Channel keys use XOR Salsa20 encryption without integrity protection, allowing any channel key holder to forge a master key by manipulating the permissions byte. This forged master key can then be used to generate keys with arbitrary permissions across all channels of the contract.
FusionPBX versions through 5.6.5 have a path traversal vulnerability in the fax_emails.php component that allows unauthenticated remote attackers to write files on the server. By sending a crafted MIME attachment with directory traversal sequences in the filename and spoofing an authorized From address, attackers can create or overwrite files with .pdf or .tif extensions in locations writable by the web server.