CVE-2025-62155
HighSummary
Nowe API to brama do dużych modeli językowych (LLM) oraz system zarządzania zasobami sztucznej inteligencji (AI). Przed wersją 0.9.6 istniała podatność SSRF, która mogła być obejściem istniejącej poprawki zabezpieczeń, umożliwiającym dostęp do intranetu.
Risk Assessment
Organizacja może być narażona na nieautoryzowany dostęp do wewnętrznych zasobów, co może prowadzić do wycieku danych lub innych poważnych incydentów bezpieczeństwa.
Recommendation
Zaleca się aktualizację do wersji 0.9.6, aby usunąć tę podatność oraz zapewnić odpowiednie zabezpieczenia przed atakami SSRF.
Related vulnerabilities
- CVE-2025-12999Critical
A vulnerability in Open VSX allows an unauthenticated attacker to poison the per-extension metadata cache by supplying a crafted X-Forwarded-Host header, replacing download, signature, and public-key URLs with attacker-controlled ones. The cache keys do not include the host, so poisoned entries are served to all clients for the entry's lifetime (1 hour by default, cluster-wide with Redis).
- CVE-2026-94101Critical
A security vulnerability has been detected in Netcore NBR200V2 1.3.241127.071246. The affected function is vlan_load_form_uci in /usr/bin/routerd, where manipulation of the wan_num argument leads to buffer overflow. The attack can be initiated remotely and the exploit is publicly disclosed.
- CVE-2026-94100Critical
A weakness has been identified in Netcore NBR200V2 1.3.241127.071246. The affected function is wan_config_set_vlan in /usr/bin/routerd, where manipulation of the vlan_wanX.ports argument leads to buffer overflow. The attack can be performed remotely and the exploit is publicly available.
- CVE-2026-94099Critical
A security flaw has been discovered in Netcore NBR200V2 1.3.241127.071246. The issue affects the restore.cgi file in the Backup Restore component, where manipulation of the QUERY_STRING argument results in command injection. The attack can be carried out remotely and the exploit is publicly released.
- CVE-2026-94098Critical
A vulnerability was identified in Netcore NBR200V2 1.3.241127.071246. The vulnerability affects the /www/cgi-bin/upgrade file in the Firmware Upgrade CGI Endpoint component, where manipulation of the QUERY_STRING argument leads to command injection. The attack can be executed remotely and the exploit is publicly available.
- CVE-2026-94097Critical
A vulnerability was determined in Netcore NBR200V2 1.3.241127.071246. The vulnerability affects the /www/cgi-bin/network_tools file in the CGI Diagnostic Endpoint component, where manipulation of the param/key/val arguments causes command injection. Remote exploitation is possible and the exploit has been publicly disclosed.
- CVE-2026-94096Critical
A vulnerability was found in Netcore NBR200V2 1.3.241127.071246. The vulnerability affects the /usr/bin/network_tools file in the LAN IP Configuration Handler component, where manipulation of the ipv4 argument results in command injection. The attack may be launched remotely and the exploit has been made public.
- CVE-2026-94095Critical
A vulnerability has been found in Netcore NBR200V2 1.3.241127.071246. The vulnerability affects the /usr/bin/network_tools file in the Traceroute Diagnostic Feature component, where manipulation of the url argument leads to command injection. The attack may be initiated remotely and the exploit has been disclosed publicly.
- CVE-2026-94089Critical
A vulnerability was found in D-Link DIR-868L 2.01b05 in the strcpy function of the file /webfa_authentication.cgi in the Authentication Handler component. Manipulating the id/password arguments leads to a stack-based buffer overflow. The attack can be executed remotely, and the exploit has been publicly disclosed.
- CVE-2026-88857Critical
In the OrdaSoft Joomla Gallery extension for Joomla versions below 6.2.7, an authenticated privileged remote code execution vulnerability was found. The saveWatermark() function copies an uploaded file to a web-accessible directory using the client-supplied filename exactly as sent, with no extension check, content check, or filename sanitization. An authenticated core.manage user could upload a .php file disguised with an image Content-Type header and execute it directly by requesting the resulting path.
Original NVD description (English source)
New API is a large language mode (LLM) gateway and artificial intelligence (AI) asset management system. Prior to version 0.9.6, a recently patched SSRF vulnerability contains a bypass method that can bypass the existing security fix and still allow SSRF to occur. Because the existing fix only applies security restrictions to the first URL request, a 302 redirect can bypass existing security measures and successfully access the intranet. This issue has been patched in version 0.9.6.

