Summary
Smanga 3.2.7 contains an authentication bypass that lets an unauthenticated attacker reset the password of any user, including administrators, by tampering with POST parameters.
Because permission checks in check-power.php are insufficient, attacker-controlled input can be used to impersonate arbitrary accounts and fully take them over.
Affected context
This vulnerability arises in Smanga 3.2.7 within the account management and privilege-checking logic implemented in check-power.php.
It is most relevant in deployments where the application is exposed to untrusted networks, and password reset or privilege-check actions are handled via POST parameters.
The issue is triggered when the server trusts client-supplied identifiers and state without reliably confirming the caller's authenticated identity and authorization level.
Exploit preconditions
- Attack surface: Public-facing Smanga web interface
- Required access: Network access to the Smanga instance (no prior authentication required)
- Trigger point: HTTP POST requests to functionality governed by
check-power.php(e.g., password reset or privilege-related actions) - Why exploitation is possible: The server-side permission validation incorrectly trusts POST parameters and does not consistently enforce that the caller is authenticated as the affected user or as an appropriately privileged administrator.
Why this happens
The core issue is a combination of broken authentication (CWE-287) and insecure direct object references / missing authorization (CWE-639).
The check-power.php logic appears to make authorization decisions based on attacker-controlled POST parameters, such as user identifiers or action flags, instead of deriving identity from a validated session.
This insecure pattern allows callers to invoke sensitive operations on arbitrary accounts without proving ownership or sufficient privileges, effectively bypassing both authentication and access control.
Potential impact
- Full takeover of any user account, including administrator accounts
- Unauthorized password resets leading to persistent control of targeted accounts
- Access to all data and actions available to compromised users
- Potential change of application-wide settings through compromised admin accounts
- Use of compromised accounts as a pivot for further attacks against integrated systems
General mitigation guidance
Effective mitigation focuses on enforcing strict server-side authentication and authorization checks independent of client-supplied parameters.
Common defensive patterns include:
- Deriving the acting user's identity exclusively from authenticated session state or secure tokens, never from POST parameters.
- Implementing robust authorization checks for every sensitive action, verifying that the authenticated user is permitted to perform the requested operation on the specified account.
- Validating that password reset and account recovery flows require strong proof of identity (e.g., verified email/OTP) rather than simple parameter manipulation.
- Centralizing permission checks in well-reviewed middleware or access-control layers instead of duplicating ad-hoc logic across scripts.
- Applying the principle of least privilege so that even administrative endpoints require explicit, verified roles and cannot be accessed anonymously.
How teams detect issues like this
Organizations can detect similar issues through a combination of code-level and behavior-level analysis:
- Static analysis: Look for authorization decisions based on request parameters (e.g.,
$_POST['user_id']) without corresponding checks against authenticated session variables. - Code review: Flag any password reset or privilege-modifying logic that does not verify current authentication state and user role on the server side.
- Security testing: Use automated and manual tests to attempt sensitive actions (like changing passwords) from unauthenticated or low-privileged contexts and verify these are consistently rejected.
- Logging and monitoring: Monitor for unusual patterns of password reset activity, especially from unauthenticated sessions or from a single IP targeting many accounts.

