Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB
CVE-2026-82045 Overview
CVE-2026-82045 is a Java Persistence Query Language (JPQL) injection vulnerability in UTMStack versions prior to 11.2.16. The flaw resides in the UtmNetworkScanService.searchPropertyValues() method, which builds a JPQL query using String.format() and executes it through em.createQuery() without parameter binding. Authenticated attackers can inject malicious JPQL through the value parameter of the GET /api/utm-network-scans/searchPropertyValues endpoint. Successful exploitation allows arbitrary entity reads, including credential tables such as jhi_user. The weakness is tracked under [CWE-89].
Critical Impact
Any authenticated UTMStack user can extract arbitrary database entity data, including stored user credentials from the jhi_user table.
Affected Products
- UTMStack versions prior to 11.2.16
- UTMStack backend UtmNetworkScanService component
- The GET /api/utm-network-scans/searchPropertyValues API endpoint
Discovery Timeline
- 2026-10-02 - CVE-2026-82045 published to the National Vulnerability Database
- 2026-10-06 - CVE-2026-82045 last updated in NVD
Technical Details for CVE-2026-82045
Vulnerability Analysis
The UtmNetworkScanService.searchPropertyValues() method constructs a JPQL query by concatenating untrusted input through String.format(). The user-supplied value parameter is interpolated directly into the query string as the %3$s positional placeholder. The resulting string is passed to EntityManager.createQuery() without any parameter binding. Because the attacker controls raw JPQL fragments, they can terminate the intended LIKE clause and append arbitrary subqueries that enumerate other entities. Authentication is required, but no elevated role is needed to reach the vulnerable endpoint.
Root Cause
The root cause is unsafe query assembly. Instead of binding parameters with setParameter(), the pre-patch code embedded the raw value string into the JPQL template, treating user input as trusted query syntax. This matches the classic server-side injection pattern described in [CWE-89], applied to the JPA query language rather than native SQL.
Attack Vector
An authenticated attacker sends a crafted GET request to /api/utm-network-scans/searchPropertyValues with a value parameter containing JPQL metacharacters and clauses. The injected fragment can read other mapped entities, including the jhi_user table used for user accounts and credentials. The vulnerability does not require user interaction and is reachable over the network against any exposed UTMStack console.
// Pre-patch (vulnerable) query construction in UtmNetworkScanService
if (prop == PropertyFilter.PORTS)
sb.append(" AND cast(port as string) LIKE '%%%3$s%%'");
else if (prop == PropertyFilter.ALIVE && !value.isEmpty()) {
sb.append(" AND %1$s = %3$s");
}
else
sb.append(" AND lower(%1$s) LIKE '%%%3$s%%'");
String query = String.format(
sb.toString(),
prop.getPropertyName(),
prop.getFromTable(),
StringUtils.hasText(value) ? value.toLowerCase() : null,
prop.getJoinTable()
);
return em.createQuery(query)
.setFirstResult(/* ... */)
.setMaxResults(/* ... */);
Source: GitHub Commit 4e7a727
Detection Methods for CVE-2026-82045
Indicators of Compromise
- Requests to GET /api/utm-network-scans/searchPropertyValues where the value query parameter contains JPQL keywords such as SELECT, FROM, WHERE, UNION, or references to entities like jhi_user.
- Unusually long or URL-encoded value parameters containing single quotes, parentheses, or comment markers.
- Spikes in searchPropertyValues request volume from a single authenticated session.
- Application logs showing javax.persistence.PersistenceException or JPQL parse errors tied to this endpoint.
Detection Strategies
- Deploy a Web Application Firewall rule to inspect the value parameter on the vulnerable endpoint for JPQL syntax tokens.
- Enable Hibernate or JPA SQL logging in UTMStack to review the queries generated by searchPropertyValues and alert on queries referencing unrelated tables.
- Correlate authenticated session identifiers with abnormal query patterns against user or credential entities.
Monitoring Recommendations
- Forward UTMStack application and reverse-proxy logs to a centralized analytics platform for query inspection and retention.
- Alert on any successful read of jhi_user data originating from the network scan API path.
- Track failed authentication followed by rapid API enumeration from the same source address.
How to Mitigate CVE-2026-82045
Immediate Actions Required
- Upgrade UTMStack to version 11.2.16 or later, which replaces string interpolation with parameterized JPQL.
- Restrict network access to the UTMStack management interface so that only trusted administrators can reach /api/utm-network-scans/*.
- Rotate credentials stored in the jhi_user table if there is evidence the endpoint was reachable by untrusted users prior to patching.
Patch Information
The fix is delivered in UTMStack release v11.2.16 and implemented in commit 4e7a727. The patch rewrites the JPQL template to use the named parameter :value and binds it through Query.setParameter(), with type-specific handling for PORTS and ALIVE filters. Additional context is available in the Vulncheck Security Advisory.
Workarounds
- Block or filter requests to GET /api/utm-network-scans/searchPropertyValues at the reverse proxy until the patch is applied.
- Enforce strict allow-list validation on the value parameter, rejecting any input containing quotes, spaces, or JPQL keywords.
- Reduce the exposure of UTMStack accounts by disabling unused users and requiring multi-factor authentication for console access.
# Example NGINX rule to block JPQL metacharacters on the vulnerable endpoint
location /api/utm-network-scans/searchPropertyValues {
if ($arg_value ~* "('|\"|--|;|\bselect\b|\bunion\b|\bfrom\b|jhi_user)") {
return 403;
}
proxy_pass http://utmstack_backend;
}
Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB

