CVE-2024-45758 is a critical security vulnerability affecting H2O.ai H2O versions through 3.46.0.4. The vulnerability arises from unsafe deserialization of user-supplied data within the ImportSQLTable API endpoint, which accepts JSON documents containing a connection_url property. Attackers can exploit this by sending a specially crafted JSON payload with a malicious JDBC URL, including payloads that leverage queryInterceptors or other JDBC features to trigger deserialization attacks. This allows arbitrary code execution on the server, unauthorized file reads, and other malicious activities. The vulnerability does not require authentication or user interaction, making it remotely exploitable over the network. The root cause is improper validation and deserialization of attacker-controlled input, classified under CWE-502 (Deserialization of Untrusted Data). The CVSS v3.1 base score is 9.1 (Critical), reflecting network attack vector, low attack complexity, no privileges required, no user interaction, and high impact on confidentiality and integrity. No patches or official fixes are currently listed, and no public exploits have been reported yet. The vulnerability threatens the confidentiality and integrity of data processed by H2O.ai H2O installations, potentially leading to full system compromise.

The impact of CVE-2024-45758 is severe for organizations using H2O.ai H2O, particularly those processing sensitive or proprietary data. Successful exploitation can lead to arbitrary command execution on the host system, allowing attackers to gain full control over the affected server. This can result in data exfiltration, destruction, or manipulation, undermining data confidentiality and integrity. Additionally, attackers could read sensitive files or deploy further malware, potentially pivoting within internal networks. Since the vulnerability requires no authentication and can be triggered remotely, it significantly increases the attack surface. Organizations in sectors such as finance, healthcare, government, and technology that rely on H2O.ai for machine learning or data analytics are at heightened risk. The lack of known exploits in the wild currently provides a window for proactive defense, but the critical severity demands immediate attention to prevent potential future attacks.

To mitigate CVE-2024-45758, organizations should immediately restrict network access to the ImportSQLTable API endpoint, ideally limiting it to trusted internal networks or VPNs. Implement strict input validation and sanitization on all JSON payloads, especially the connection_url parameter, to prevent injection of malicious JDBC URLs. Employ application-layer firewalls or web application firewalls (WAFs) to detect and block suspicious payloads targeting deserialization vulnerabilities. Monitor logs for unusual activity related to the ImportSQLTable endpoint and JDBC connection attempts. Until an official patch is released by H2O.ai, consider disabling or restricting the ImportSQLTable functionality if not essential. Additionally, conduct thorough code reviews and security testing focusing on deserialization and JDBC connection handling. Prepare incident response plans to quickly address any signs of exploitation. Stay updated with vendor advisories for patches or mitigations.