Submit #773906: vanna.ai vanna 2.0.2 SQL Injection

Vanna utilizes large language models to understand user intents and generate SQL, enabling users to interact with databases using natural language. However, Vanna executes all SQL statements generated by the LLM without filtration. Malicious users may control the model’s output through prompt injection, leading to arbitrary SQL execution. While this may be an intentional design choice of Vanna, due to the powerful features of Oracle databases, allowing arbitrary SQL execution could enable attackers to run arbitrary commands on the target server, resulting in full server compromise.

• Source: ⚠️ https://github.com/Ka7arotto/cve/blob/main/vanna-text2sql/vanna-sql-rce.md
• User: Goku (UID 80486)
• Submission: 03/06/2026 12:48 (1 month ago)
• Moderation: 03/20/2026 15:28 (14 days later)
• Status: Accepted
• VulDB entry: 352078 [vanna-ai vanna up to 2.0.2 base.py ask sql injection]
• Points: 20