Executive Summary
A researcher demonstrated database remote-code-execution exploits affecting PostgreSQL, MySQL, ClickHouse, and Oracle, including scenarios involving low-privilege accounts. Teams operating AI agents or LLM applications with database connectivity should review database permissions and test whether agent-accessible accounts could trigger vulnerable behavior.
What Happened
Low-privilege database accounts that cannot read a single application table can still be used to run code on the database server itself, a researcher has shown, by reaching memory-corruption bugs in the engine code behind ordinary SQL. The exploits, published yesterday, are the researcher's proofs of concept against PostgreSQL, MySQL and ClickHouse, plus a separate Oracle chain that needs no
Affected Products
- PostgreSQL
- MySQL
- ClickHouse
- Oracle
Vulnerability / Attack Technique
Memory-corruption exploitation through SQL
Technical Mechanism
The reported exploits use ordinary SQL to reach memory-corruption bugs in database engine code. The summary states that low-privilege accounts unable to read an application table could still be used to run code on the database server; it does not provide further technical details about the Oracle chain.
AI Security Category
AI-assisted attack
Part of the AI-Assisted Attacks topic.
Relevant Frameworks
RealGround Analysis
This is not an AI-system weakness, but it can affect AI agents and LLM applications that issue SQL using database credentials. Agent workflows should be evaluated for whether their database access could provide a path to exploit vulnerable engine code or compromise the host.
Recommended Controls
- Apply vendor security updates for PostgreSQL, MySQL, ClickHouse, and Oracle when available.
- Use narrowly scoped database roles for AI agents and prohibit unnecessary execution of arbitrary SQL.
- Isolate database servers from agent runtimes and other sensitive systems using network segmentation.
- Monitor database activity for unusual SQL patterns, memory-corruption indicators, and unexpected operating-system process creation.
- Include database-engine exploitation scenarios in AI agent red-team and business-logic assessments.
Related RealGround Service
AI Agent Business Logic Audit · or start with the free AI Agent Security Scanner.
Sources
thehackernews.com: https://thehackernews.com/2026/11/researcher-demonstrates-database-rce.html
Summary and analysis written by RealGround from the source above; last updated 2026-11-19.
