A critical SQL injection vulnerability in SciPhi-AI R2R allows unauthenticated remote attackers to execute arbitrary database commands, posing a severe risk of full system compromise.

The application utilizes improper string formatting to construct SQL statements within the vector index creation endpoint, failing to sanitize the index name parameter. This allows unauthenticated attackers to perform stacked SQL injection, enabling the execution of arbitrary Data Definition Language (DDL) and Data Manipulation Language (DML) statements as a PostgreSQL superuser.

Successful exploitation of this flaw grants an attacker complete control over the underlying database, including the ability to read, modify, or delete sensitive information. Given the CVSS score of 9.8, this vulnerability is categorized as critical, as it facilitates unauthorized access and potential total system takeover, which could lead to significant data breaches and prolonged operational downtime.

Immediate Action: Update SciPhi-AI R2R to the latest version immediately to patch the vulnerable index creation logic.

Proactive Monitoring: Review database access logs for unusual query patterns, specifically looking for semicolon-separated commands or attempts to access system tables such as pg_authid or pg_catalog.

Compensating Controls: Deploy a Web Application Firewall (WAF) with rules configured to detect and block SQL injection patterns, particularly those targeting the /v3/indices endpoint.

Public Exploit Available: Yes, a published proof-of-concept exists as detailed in the official GitHub issue tracker for the project.

The severity of this vulnerability cannot be overstated, as it provides an unauthenticated path to database superuser privileges. Administrators must prioritize updating the R2R software to the latest release and ensure that database service accounts are configured with the principle of least privilege, rather than running as PostgreSQL superusers. Failure to remediate this issue immediately leaves the environment exposed to full data exfiltration and administrative control by external attackers.