Loading vulnerability…
CVE-2026-22730 is a critical SQL injection vulnerability affecting Spring AI's MariaDBFilterExpressionConverter in the spring-ai-mariadb-store package. According to the provided content, the flaw is caused by missing input sanitization in the converter, allowing attacker-controlled input to be incorporated into SQL queries. Successful exploitation can bypass metadata-based access controls and result in execution of arbitrary SQL commands against the backend MariaDB database.
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.
This repository is a small standalone Python exploit/scanner project, not part of a larger framework. It contains two files: a README with usage guidance and payload examples, and a single executable script, cve-2026-22730_scanner.py, which is the main entry point. The script is designed to assess CVE-2026-22730, described here as a SQL injection flaw in Spring AI’s MariaDB vector store. Its primary workflow is: accept a target URL and request mode, send a baseline request, send a time-based SQLi payload, compare response timing to determine likely vulnerability, and if the target appears vulnerable, prompt the operator to either retrieve all documents or issue a destructive delete-all action. The code supports GET requests with query parameters, POST requests with JSON bodies, and DELETE requests using either query parameters or JSON bodies. It uses the requests library, a fixed User-Agent string, a 10-second timeout, and a __PAYLOAD__ placeholder for JSON template substitution. Capabilities visible from the code and README include: time-based detection using SLEEP-style SQL injection, optional exploitation to force broad data retrieval via a tautology payload, optional destructive exploitation via DELETE, interactive CLI prompting, command-line argument handling, and report generation in JSON/TXT/HTML formats. This makes it more than a pure detector; it includes active exploitation logic, including destructive functionality. Repository structure is minimal and purpose-built: README.md documents the vulnerability, command-line options, example target URLs, payload categories, and output artifacts; cve-2026-22730_scanner.py implements request construction, vulnerability detection, exploitation selection, and result saving. The code shown is partially truncated, but enough remains to establish that it is a real operational PoC/exploit utility rather than a simple advisory or fake sample.
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.
27 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.
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.