CVE-2026-25130
CriticalSummary
Framework Cybersecurity AI (CAI) w wersjach do 0.5.10 zawiera wiele podatności na wstrzykiwanie argumentów w swoich funkcjach. Użytkownik może wprowadzać dane, które są bezpośrednio przekazywane do poleceń powłoki, co pozwala atakującym na wykonywanie dowolnych poleceń na systemie gospodarza.
Risk Assessment
Podatność ta umożliwia zdalne wykonanie kodu (RCE), co stwarza poważne zagrożenie dla bezpieczeństwa systemu i danych organizacji. Atakujący może wykorzystać tę lukę do przejęcia kontroli nad systemem.
Recommendation
Zaleca się aktualizację frameworka CAI do wersji zawierającej poprawkę, aby usunąć te podatności. Należy również przeprowadzić audyt bezpieczeństwa, aby zidentyfikować i zminimalizować ryzyko związane z wstrzykiwaniem argumentów.
Related vulnerabilities
- CVE-2026-108707Critical
Wukong_HRM through commit 186115e contains an authentication bypass vulnerability in ParamAspect that allows unauthenticated attackers to call every HRM API endpoint by omitting the AUTH-TOKEN header. Attackers gain HR administrator access to read payslips, salary history and employee personal data, download attachments, and modify or delete company-wide HR records.
- CVE-2026-81797Critical
The WordPress theme Buzz Stone | Magazine & Viral Blog in versions up to 1.0.2 contains a PHP Object Injection vulnerability. An unauthenticated attacker can inject a malicious PHP object, potentially leading to arbitrary code execution or other severe impacts.
- CVE-2026-78535Critical
The WordPress theme Photolia in versions up to 1.0.3 contains a PHP Object Injection vulnerability. An unauthenticated attacker can inject a malicious PHP object, potentially leading to arbitrary code execution or other severe impacts.
- CVE-2026-78533Critical
The WordPress theme Qwery in versions up to 3.6.1 contains a PHP Object Injection vulnerability. An unauthenticated attacker can inject a malicious PHP object, potentially leading to arbitrary code execution or other severe impacts.
- CVE-2026-78531Critical
The WordPress theme Jacqueline in versions up to 2.22 contains a PHP Object Injection vulnerability. An unauthenticated attacker can inject a malicious PHP object, potentially leading to arbitrary code execution or other severe impacts.
- CVE-2026-78529Critical
The WordPress theme Alliance in versions up to 3.11 contains a PHP Object Injection vulnerability. An unauthenticated attacker can inject a malicious PHP object, potentially leading to arbitrary code execution or other severe impacts.
- CVE-2026-66569Critical
The WordPress theme Kicker in versions up to 2.2.1 contains a PHP Object Injection vulnerability. An unauthenticated attacker can inject a malicious PHP object, potentially leading to arbitrary code execution or other severe impacts.
- CVE-2026-66568Critical
The WordPress theme Original in versions up to 1.9.0 contains a PHP Object Injection vulnerability. An unauthenticated attacker can inject a malicious PHP object, potentially leading to arbitrary code execution or other severe impacts.
- CVE-2026-66567Critical
The WordPress theme Anesta in versions up to 1.5.3 contains a PHP Object Injection vulnerability. An unauthenticated attacker can inject a malicious PHP object, potentially leading to arbitrary code execution or other severe impacts.
- CVE-2026-66565Critical
The WordPress theme FC United in versions up to 1.1.1 contains a PHP Object Injection vulnerability. An unauthenticated attacker can inject a malicious PHP object, potentially leading to arbitrary code execution or other severe impacts.
Original NVD description (English source)
Cybersecurity AI (CAI) is a framework for AI Security. In versions up to and including 0.5.10, the CAI (Cybersecurity AI) framework contains multiple argument injection vulnerabilities in its function tools. User-controlled input is passed directly to shell commands via `subprocess.Popen()` with `shell=True`, allowing attackers to execute arbitrary commands on the host system. The `find_file()` tool executes without requiring user approval because find is considered a "safe" pre-approved command. This means an attacker can achieve Remote Code Execution (RCE) by injecting malicious arguments (like -exec) into the args parameter, completely bypassing any human-in-the-loop safety mechanisms. Commit e22a1220f764e2d7cf9da6d6144926f53ca01cde contains a fix.
Vulnerability data from NVD (NIST) · CISA KEV · EPSS