Summary
ChurchCRM prior to version 7.3.2 exposes several deletion endpoints that can be triggered via cross-site GET requests without additional confirmation or CSRF protection.
If a logged-in user with the appropriate permissions visits an attacker-controlled page, crafted navigation can silently delete fundraiser, property type, or note records, including cascaded associations.
Affected context
This vulnerability affects ChurchCRM deployments where authenticated users with deletion privileges are actively using a browser that can be navigated or embedded by untrusted sites.
The issue occurs in endpoints such as FundRaiserDelete.php, PropertyTypeDelete.php, and NoteDelete.php that accept state-changing operations via top-level HTTP GET requests without robust cross-site request forgery protections or explicit user confirmation.
Exploit preconditions
- Attack surface: Web browser of an authenticated ChurchCRM user.
- Required access: The attacker must be able to entice a privileged, logged-in ChurchCRM user to visit or be redirected from an attacker-controlled web page.
- Trigger point: A top-level navigation (e.g., link click, automatic redirect, meta refresh, or certain frame-based techniques) from the attacker page to deletion URLs such as
FundRaiserDelete.php,PropertyTypeDelete.php, orNoteDelete.php. - Why exploitation is possible: The deletion actions are performed via GET requests that rely solely on existing session cookies for authorization and do not require an anti-CSRF token, re-authentication, or explicit confirmation from the user, allowing the browser to perform unintended state-changing actions in the background.
Why this happens
The core issue aligns with CWE-352: Cross-Site Request Forgery, where state-changing actions are bound only to a user's authenticated session and not to an explicit, user-initiated request from the application's own origin.
ChurchCRM's deletion endpoints accept sensitive operations over GET without CSRF tokens or same-site interaction checks, and without an additional confirmation step.
This pattern also intersects with CWE-650: Trusting HTTP GET Request Data for Sensitive Actions, because destructive operations are bound to easily triggered URLs that can be invoked through normal browser navigation from untrusted origins.
Potential impact
- Silent deletion of fundraiser records initiated by targeted users.
- Removal of property types and associated property assignments.
- Deletion of notes linked to people or other entities in the system.
- Loss of historical and operational data that may be difficult or impossible to reconstruct.
- Potential for targeted data disruption against specific congregations, groups, or individuals.
General mitigation guidance
Effective mitigation generally focuses on ensuring that destructive or state-changing actions cannot be triggered purely through cross-site navigation:
- Introduce robust CSRF protections for all state-changing endpoints, such as per-request tokens tied to the user session and origin.
- Avoid using HTTP GET for operations that create, modify, or delete data; prefer POST/DELETE with appropriate CSRF validation.
- Require explicit user interaction for destructive actions, such as confirmation dialogs backed by action-specific tokens or short-lived re-authentication prompts.
- Apply the principle of least privilege to authorization checks, ensuring only narrowly scoped roles can perform bulk or cascading deletions.
- Consider implementing additional safeguards such as soft-delete, audit logging, or delayed/queued deletion to reduce the impact of accidental or malicious removal.
How teams detect issues like this
Organizations can look for this pattern through a mix of design review and automated checks:
- Static analysis and code scanning rules that flag state-changing operations on GET routes, especially those containing terms like "Delete", "Remove", or "Destroy".
- Security-focused code review checklists that require CSRF tokens and POST/DELETE semantics for endpoints that modify database state.
- Runtime or proxy-level policies that monitor for cross-site navigation to sensitive endpoints, such as referrer-based anomaly detection or Content Security Policy (CSP) reports.
- Automated testing in CI/CD that attempts CSRF-style requests (without tokens or from off-origin contexts) against known sensitive routes to detect missing protections early.

