CVE-2025-1793: CWE-89 Improper Neutralization of Special Elements used in an SQL Command in run-llama run-llama/llama_index
Multiple vector store integrations in run-llama/llama_index version v0.12.21 have SQL injection vulnerabilities. These vulnerabilities allow an attacker to read and write data using SQL, potentially leading to unauthorized access to data of other users depending on the usage of the llama-index library in a web application.
AI Analysis
Technical Summary
CVE-2025-1793 is a critical SQL injection vulnerability identified in multiple vector store integrations within the run-llama/llama_index library, specifically version v0.12.21. The vulnerability stems from improper neutralization of special elements used in SQL commands (CWE-89), allowing attackers to inject malicious SQL code. This flaw enables unauthorized reading and writing of data within the underlying database. Since llama_index is a library commonly used to facilitate data indexing and retrieval in applications, especially those leveraging vector stores for AI or search functionalities, the vulnerability can be exploited if the library is integrated into web applications without proper input sanitization or parameterization. Exploitation requires no authentication or user interaction, and the attack vector is network accessible (AV:N), making it highly exploitable remotely. The CVSS 3.0 score of 9.8 reflects the critical nature of this vulnerability, with high impacts on confidentiality, integrity, and availability. An attacker could potentially access sensitive data belonging to other users, modify or delete data, or disrupt service availability. The lack of available patches at the time of publication increases the risk for applications relying on this library. Given the widespread adoption of AI and vector-based indexing in modern applications, this vulnerability poses a significant threat to any system using the affected versions of run-llama/llama_index in a web-facing context.
Potential Impact
For European organizations, the impact of CVE-2025-1793 could be severe. Many enterprises and service providers in Europe utilize AI-driven search, recommendation, and data indexing solutions that may incorporate the run-llama/llama_index library or similar vector store integrations. Exploitation could lead to unauthorized disclosure of personal data, violating GDPR and other data protection regulations, resulting in legal penalties and reputational damage. Additionally, data integrity compromises could disrupt business operations, cause financial losses, and erode customer trust. Critical sectors such as finance, healthcare, telecommunications, and government services, which often handle sensitive or regulated data, are particularly at risk. The vulnerability’s ease of exploitation without authentication means attackers can remotely compromise systems without insider access, increasing the threat landscape. Furthermore, the absence of known exploits in the wild currently does not preclude rapid weaponization, especially given the critical severity and public disclosure. Organizations relying on web applications that integrate this library must consider the risk of data breaches, service interruptions, and compliance violations.
Mitigation Recommendations
To mitigate CVE-2025-1793, European organizations should take immediate and specific actions beyond generic patching advice: 1) Inventory all applications and services to identify any usage of run-llama/llama_index version v0.12.21 or earlier. 2) If possible, upgrade to a patched version once available; if no patch exists yet, consider temporarily disabling or isolating affected functionalities to reduce exposure. 3) Implement strict input validation and parameterized queries around any database interactions involving vector store integrations to prevent injection. 4) Employ Web Application Firewalls (WAFs) with custom rules to detect and block SQL injection patterns targeting the vulnerable endpoints. 5) Conduct thorough code reviews and penetration testing focused on injection vectors in applications using the library. 6) Monitor logs and network traffic for unusual database queries or access patterns indicative of exploitation attempts. 7) Apply the principle of least privilege to database accounts used by applications, limiting the scope of potential damage. 8) Educate developers and security teams about secure coding practices related to SQL injection and the specific risks of third-party libraries. These targeted steps will help reduce the attack surface and prepare organizations to respond effectively once patches are released.
Affected Countries
Germany, France, United Kingdom, Netherlands, Sweden, Finland, Belgium, Italy
CVE-2025-1793: CWE-89 Improper Neutralization of Special Elements used in an SQL Command in run-llama run-llama/llama_index
Description
Multiple vector store integrations in run-llama/llama_index version v0.12.21 have SQL injection vulnerabilities. These vulnerabilities allow an attacker to read and write data using SQL, potentially leading to unauthorized access to data of other users depending on the usage of the llama-index library in a web application.
AI-Powered Analysis
Technical Analysis
CVE-2025-1793 is a critical SQL injection vulnerability identified in multiple vector store integrations within the run-llama/llama_index library, specifically version v0.12.21. The vulnerability stems from improper neutralization of special elements used in SQL commands (CWE-89), allowing attackers to inject malicious SQL code. This flaw enables unauthorized reading and writing of data within the underlying database. Since llama_index is a library commonly used to facilitate data indexing and retrieval in applications, especially those leveraging vector stores for AI or search functionalities, the vulnerability can be exploited if the library is integrated into web applications without proper input sanitization or parameterization. Exploitation requires no authentication or user interaction, and the attack vector is network accessible (AV:N), making it highly exploitable remotely. The CVSS 3.0 score of 9.8 reflects the critical nature of this vulnerability, with high impacts on confidentiality, integrity, and availability. An attacker could potentially access sensitive data belonging to other users, modify or delete data, or disrupt service availability. The lack of available patches at the time of publication increases the risk for applications relying on this library. Given the widespread adoption of AI and vector-based indexing in modern applications, this vulnerability poses a significant threat to any system using the affected versions of run-llama/llama_index in a web-facing context.
Potential Impact
For European organizations, the impact of CVE-2025-1793 could be severe. Many enterprises and service providers in Europe utilize AI-driven search, recommendation, and data indexing solutions that may incorporate the run-llama/llama_index library or similar vector store integrations. Exploitation could lead to unauthorized disclosure of personal data, violating GDPR and other data protection regulations, resulting in legal penalties and reputational damage. Additionally, data integrity compromises could disrupt business operations, cause financial losses, and erode customer trust. Critical sectors such as finance, healthcare, telecommunications, and government services, which often handle sensitive or regulated data, are particularly at risk. The vulnerability’s ease of exploitation without authentication means attackers can remotely compromise systems without insider access, increasing the threat landscape. Furthermore, the absence of known exploits in the wild currently does not preclude rapid weaponization, especially given the critical severity and public disclosure. Organizations relying on web applications that integrate this library must consider the risk of data breaches, service interruptions, and compliance violations.
Mitigation Recommendations
To mitigate CVE-2025-1793, European organizations should take immediate and specific actions beyond generic patching advice: 1) Inventory all applications and services to identify any usage of run-llama/llama_index version v0.12.21 or earlier. 2) If possible, upgrade to a patched version once available; if no patch exists yet, consider temporarily disabling or isolating affected functionalities to reduce exposure. 3) Implement strict input validation and parameterized queries around any database interactions involving vector store integrations to prevent injection. 4) Employ Web Application Firewalls (WAFs) with custom rules to detect and block SQL injection patterns targeting the vulnerable endpoints. 5) Conduct thorough code reviews and penetration testing focused on injection vectors in applications using the library. 6) Monitor logs and network traffic for unusual database queries or access patterns indicative of exploitation attempts. 7) Apply the principle of least privilege to database accounts used by applications, limiting the scope of potential damage. 8) Educate developers and security teams about secure coding practices related to SQL injection and the specific risks of third-party libraries. These targeted steps will help reduce the attack surface and prepare organizations to respond effectively once patches are released.
Affected Countries
For access to advanced analysis and higher rate limits, contact root@offseq.com
Technical Details
- Data Version
- 5.1
- Assigner Short Name
- @huntr_ai
- Date Reserved
- 2025-02-28T18:09:06.434Z
- Cvss Version
- 3.0
- State
- PUBLISHED
Threat ID: 68418437182aa0cae2dcccbd
Added to database: 6/5/2025, 11:49:11 AM
Last enriched: 7/7/2025, 3:55:20 AM
Last updated: 8/1/2025, 4:21:57 PM
Views: 18
Related Threats
CVE-2025-8972: SQL Injection in itsourcecode Online Tour and Travel Management System
MediumCVE-2025-51986: n/a
HighCVE-2025-52335: n/a
HighCVE-2025-8971: SQL Injection in itsourcecode Online Tour and Travel Management System
MediumCVE-2025-8970: SQL Injection in itsourcecode Online Tour and Travel Management System
MediumActions
Updates to AI analysis are available only with a Pro account. Contact root@offseq.com for access.
Need enhanced features?
Contact root@offseq.com for Pro access with improved analysis and higher rate limits.