Vulnerabilities · Application Security
Researcher shows how attackers can run code on major database servers
A security researcher has demonstrated ways to run code on PostgreSQL, MySQL and Oracle servers using low-privilege accounts. The findings highlight risks even when users have limited access.
Ross Grant2 min read

The short version
A security researcher has found ways to run code on popular database servers like PostgreSQL, MySQL and Oracle, even with very limited user access. This means attackers could take control of a server without needing high-level permissions. The findings have not yet been confirmed by the database vendors. Organisations using these databases should be alert and consider limiting user access until more is known.
What happened?
A security researcher has demonstrated ways to run code on PostgreSQL, MySQL and Oracle servers using low-privilege accounts. The proof-of-concept exploits target memory-corruption bugs in the database engines. Only The Hacker News has reported this so far, and there is no vendor confirmation yet.
How the flaw works
A security researcher has shown that even low-privilege database accounts, users who cannot read a single application table, can be used to run code on the database server itself. The researcher demonstrated this using proof-of-concept exploits against PostgreSQL, MySQL and ClickHouse. For Oracle, a separate exploit chain was shown. The technical details were published on 18 November 2026, according to The Hacker News (opens in a new tab).
The core of the issue is that attackers can reach memory-corruption bugs in the database engine code by sending carefully crafted SQL commands. Memory corruption means the attacker can change how the server behaves, often letting them run their own code, this is known as remote code execution (RCE). RCE is one of the most serious types of vulnerability because it can give attackers control of the server.
Who is affected
The researcher’s demonstrations targeted PostgreSQL, MySQL, ClickHouse and Oracle. These are some of the most widely used database systems in the world, powering everything from websites to enterprise applications. The proof-of-concept attacks worked even when the attacker had only a low-privilege account. That means organisations that think their databases are safe because users have limited permissions may still be at risk.
What you should do
There is no mention yet of official patches or vendor advisories. Organisations using PostgreSQL, MySQL, ClickHouse or Oracle should review user permissions and limit access as much as possible. Monitoring for unusual SQL queries or signs of exploitation is also important. Until the vendors confirm and address these flaws, database administrators should be cautious about who can access their systems and what commands they can run.
So far, only The Hacker News has reported these findings. There is no confirmation from the database vendors or any government agency. The full technical details and the impact on different versions of the databases are not yet clear.
Why it matters
Many organisations rely on these databases to store sensitive data. If attackers can run code on a database server using a low-privilege account, they could bypass normal security controls. This could lead to data theft or further attacks.
Key points
- A researcher showed remote code execution is possible on major databases using low-privilege accounts.
- The affected systems include PostgreSQL, MySQL, ClickHouse and Oracle.
- Attackers exploit memory-corruption bugs in the database engine code.
- No official patches or vendor statements have been released yet.
- Organisations should review permissions and monitor for suspicious activity.
Sources
The Hacker News
Reference

