The Federated Learning and Interoperability Platform (FLIP) developed by londonaicentre is an open-source platform designed to facilitate federated training and evaluation of medical imaging AI models across multiple healthcare institutions. Versions 0.1.1 and earlier of FLIP contain a security vulnerability identified as CVE-2026-33879, classified under CWE-307 (Improper Restriction of Excessive Authentication Attempts). Specifically, the FLIP login page does not implement any rate limiting or CAPTCHA mechanisms, allowing attackers to perform unlimited authentication attempts. This absence of controls enables brute-force and credential-stuffing attacks, where attackers systematically try numerous password combinations or reused credentials to gain unauthorized access. Since FLIP users are external to the organizations hosting the platform, the risk of credential reuse attacks is elevated, as attackers may leverage leaked credentials from other breaches. The vulnerability has a CVSS 4.0 base score of 2.7, indicating low severity primarily due to the lack of direct impact on confidentiality, integrity, or availability without successful authentication. No patches or fixes have been published at the time of disclosure, and no known exploits are in the wild. The vulnerability's exploitation could lead to unauthorized access to federated learning environments, potentially exposing sensitive medical AI model data or enabling manipulation of training processes. Given the critical nature of medical data and AI model integrity, this vulnerability poses a notable risk to healthcare institutions collaborating via FLIP. Organizations using FLIP should prioritize implementing protective measures to mitigate brute-force risks while awaiting official patches.

The primary impact of CVE-2026-33879 is the increased risk of unauthorized access to FLIP instances through brute-force or credential-stuffing attacks. Successful exploitation could allow attackers to compromise user accounts, potentially exposing sensitive medical imaging AI training data or enabling manipulation of federated learning processes. This could undermine the confidentiality and integrity of AI models used for critical healthcare diagnostics. Although the CVSS score is low, the specialized nature of FLIP and its use in healthcare elevates the potential consequences. Credential reuse risk is significant because FLIP users are external and may use passwords compromised elsewhere. Additionally, unauthorized access could facilitate lateral movement within healthcare networks or data exfiltration. The lack of rate limiting also increases the likelihood of automated attacks succeeding over time. While no direct availability impact is indicated, the breach of trust and data integrity in medical AI systems could have downstream effects on patient care quality and regulatory compliance. Overall, the vulnerability poses a moderate operational and reputational risk to organizations deploying FLIP in sensitive healthcare environments.

To mitigate CVE-2026-33879, organizations should implement immediate compensating controls given the absence of an official patch. Key recommendations include: 1) Deploy rate limiting on the FLIP login endpoint to restrict the number of authentication attempts per IP address or user account within a defined time window. 2) Integrate CAPTCHA challenges on the login page to prevent automated brute-force attacks. 3) Enforce strong password policies and encourage users to avoid credential reuse by leveraging password managers and multi-factor authentication (MFA) where possible. 4) Monitor authentication logs for unusual patterns such as repeated failed login attempts or logins from suspicious IP addresses, and trigger alerts or account lockouts accordingly. 5) Segment FLIP infrastructure within the network to limit lateral movement if an account is compromised. 6) Educate users about phishing and credential hygiene to reduce the risk of credential theft. 7) Stay updated with londonaicentre announcements for patches or security updates and apply them promptly. 8) Consider deploying Web Application Firewalls (WAFs) with rules to detect and block brute-force attack signatures. These measures collectively reduce the attack surface and improve resilience against exploitation until an official fix is available.