CVE-2026-7189: Proliz's OBS Sensitive Information Exposure Vulnerability
TL;DR - Proliz’s OBS before v3.6.0 is affected by a network-exploitable sensitive data exposure issue. - No authentication or user interaction is required per the CNA CVSS vector. - Upgrade to v3.6.0 immediately; public PoC and in-the-wild exploitation are not currently confirmed.
Vulnerability at a Glance
| Field | Value |
|---|---|
| CVE ID | CVE-2026-7189 |
| CVSS | 7.5 High (CNA/MITRE), 8.2 in NVD |
| Attack vector | Network |
| Privileges required | None |
| Patch available | Yes, fixed in v3.6.0 |
CVE-2026-7189 is a high-severity vulnerability in Proliz’s OBS, published on 2026-07-17 and assigned by TR-CERT. The primary technical classification is CWE-201: Insertion of Sensitive Information Into Sent Data. The CVE record also maps to CAPEC-1: Accessing Functionality Not Properly Constrained by ACLs, which suggests the exposure likely stems from insufficient access control around one or more application functions.
The most defensible severity interpretation comes from the CNA/MITRE CVSS v3.1 vector because it includes full scoring details: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N. That means defenders should treat this as remotely reachable, low-complexity, unauthenticated exposure with high confidentiality impact, even though integrity and availability impacts were not assessed as direct outcomes in the record.
What Is This Vulnerability?
At a high level, this is a sensitive information disclosure issue in Proliz’s OBS. The published description states that the product “allows Accessing Functionality Not Properly Constrained by ACLs,” which points to an access control weakness rather than classic memory corruption, injection, or authentication bypass in the narrow sense. In practice, that kind of flaw often means a request can reach a function or response path that returns data to a caller who should not be able to access it.
The CWE mapping to CWE-201 is important because it narrows the likely behavior: the application is sending data that contains sensitive information when it should not. The CNA vector says PR:N and UI:N, so the working assumption for defenders should be that an attacker may be able to trigger disclosure remotely without logging in and without relying on a victim to click a link or perform any action.
What is not publicly disclosed, based on the currently available primary sources, is equally important. The advisory material retrieved does not identify the exact endpoint, parameter, module, request path, or response object involved. It also does not clarify whether the exposed data includes PII, configuration details, internal identifiers, tokens, or business records. Because those specifics are unavailable, defenders should assume the issue may affect any externally reachable OBS function that returns structured data and should review exposure accordingly.
Another practical implication of the ACL-related description is that this may not look like a crash or service fault. Instead, exploitation could appear as normal-looking HTTP or API traffic that receives a successful response containing data an unauthenticated user should never see. That makes logging, access review, and external exposure reduction especially important until patched.
Who Is Affected?
The affected product named in the CNA/MITRE record is Proliz’s OBS. The affected version range is listed as before v3.6.0. The version expression in the record uses a custom versioning type with version: 0 and lessThan: v3.6.0, which operationally means organizations should treat all releases earlier than v3.6.0 as affected, unless Proliz publishes a narrower exception later.
The fixed version boundary is clearly stated by the affected-range metadata: v3.6.0. If your environment runs Proliz’s OBS and the installed version is older than v3.6.0, it should be considered vulnerable. If you do not have complete version inventory, assume exposed instances are affected until verified otherwise.
This matters most for organizations that have Proliz’s OBS accessible from untrusted networks, including public internet exposure, partner access networks, or flat internal segments where many users or systems can reach the application. Because the CVSS vector indicates network exploitation with no privileges required, internet-facing deployments carry the highest immediate risk.
If your team cannot quickly confirm version status, treat this as a patch-validation problem as well as an exposure-management problem. Asset owners should identify where OBS is hosted, which reverse proxies or load balancers publish it, and whether there are cloned, DR, staging, or unmanaged instances still running older builds. In many incidents, overlooked secondary instances remain the weakest link after the primary system is patched.
CVSS Score Breakdown
The CNA/MITRE vector is CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N, which corresponds to a 7.5 High severity score. Each component tells a useful story for defenders. AV:N means the issue is reachable over the network, so the attacker does not need local shell access or adjacent network placement. AC:L means exploitation is not expected to require unusual preconditions, race timing, or complex chaining.
The combination of PR:N and UI:N is what should drive urgency. An attacker does not need to authenticate, and a legitimate user does not need to assist by clicking or opening anything. In operational terms, that frequently translates to lower attacker cost and broader opportunistic scanning potential if the vulnerable service is externally reachable.
The S:U component means the vulnerability impact remains within the same security scope rather than crossing trust boundaries in the CVSS model. C:H is the core impact metric here: confidentiality impact is high. Even if there is no proven direct modification or disruption, exposure of sensitive information can still lead to downstream compromise, including account takeover, internal reconnaissance, or abuse of leaked business data.
The vector assigns I:N and A:N, so no direct integrity or availability impact is currently claimed in the public record. That does not mean the business impact is low. Data disclosure vulnerabilities can still enable further attacks, especially if the exposed information includes authentication tokens, identifiers, internal architecture details, or customer records. NVD lists a score of 8.2, but absent a more detailed alternate vector in the retrieved material, the CNA vector remains the most technically specific primary-source basis for assessment.
Exploitation Status
At the time of writing, there is no confirmed public proof of concept in the retrieved primary references. No GitHub repository, exploit write-up, or vendor-published technical reproduction steps were identified in the available source set. That means defenders should avoid assuming the issue is harmless simply because exploit code has not yet surfaced publicly.
There is also no primary-source confirmation of active exploitation in the wild. The CVE is not currently listed in CISA’s Known Exploited Vulnerabilities catalog, which means there is no CISA-confirmed evidence of in-the-wild exploitation at this time based on KEV status. That is a useful signal, but not a guarantee of safety.
Sparse public detail can cut both ways. On one hand, lack of endpoint and exploit specifics may slow opportunistic exploitation. On the other hand, a simple ACL weakness in a web application can often be rediscovered quickly by attackers once version targets are known. Because this issue is remotely reachable and does not require authentication per the CVSS vector, defenders should act as though weaponization is plausible even if it is not yet documented.
For change control and risk communication, the right statement is: patch availability is known, affected versions are known, public PoC is not currently known, and in-the-wild exploitation is not currently confirmed. That is enough to justify accelerated remediation for exposed systems.
How to Detect It
Detection is challenging because the public advisory material does not disclose the exact vulnerable route or response pattern. That means there is no authoritative single URI, parameter name, or response signature to match on today. In the absence of those details, defenders should focus on identifying unauthenticated requests to OBS endpoints that return unusually rich data, especially where 200 responses include structured records, identifiers, or fields that should normally require login.
Start with basic exposure and traffic review. Identify all hosts serving Proliz’s OBS, then review reverse proxy, web server, WAF, and application logs for requests from unauthenticated sources that returned successful responses and nontrivial payload sizes. If the application separates authenticated and unauthenticated routes, prioritize any endpoint that returned data before session establishment or without expected authorization checks.
Technical Notes
Example grep patterns for web or reverse-proxy logs, assuming common combined formats:
grep -Ei 'GET|POST' /var/log/nginx/access.log | grep -Ei ' 200 | 206 ' | grep -Ei 'obs|api|export|report|user|detail|list'
Look for suspicious patterns such as repeated unauthenticated requests from the same IP, especially if they enumerate resources or fetch large responses:
awk '$9 ~ /200|206/ {print $1, $7, $9, $10}' /var/log/nginx/access.log | sort | uniq -c | sort -nr | head -50
If logs include authenticated user context, hunt for empty or anonymous principals receiving successful responses:
SELECT timestamp, src_ip, request_path, status, bytes_sent
FROM web_requests
WHERE app = 'proliz_obs'
AND status IN (200,206)
AND (user_id IS NULL OR user_id = '' OR auth_state = 'anonymous')
ORDER BY timestamp DESC;
A generic Splunk search for potentially relevant activity:
index=web sourcetype IN (nginx,apache,iis)
("obs" OR "proliz")
status IN (200,206)
| eval unauth=if(isnull(user) OR user="" OR like(auth_state,"anonymous"),1,0)
| search unauth=1
| stats count sum(bytes) as total_bytes values(uri_path) as paths by src_ip
| sort - total_bytes
On the network side, large HTTP responses to unauthenticated requests against OBS from internet IPs are worth triaging. Because the flaw is described as sensitive information sent in data, a useful heuristic is to alert on repeated successful requests with unusually high response sizes or content types such as JSON, CSV, or export/download responses from anonymous sessions.
Mitigation and Patching
The primary remediation is straightforward from the CVE record: upgrade Proliz’s OBS to v3.6.0 or later. The affected range is before v3.6.0, so v3.6.0 is the minimum fixed version currently identified in primary sources. If your estate includes multiple OBS instances, verify all of them, including standby, test, and geographically distributed deployments.
Because the public advisory material does not provide a deeper workaround narrative, defenders should combine patching with exposure reduction. If immediate upgrade is not possible, restrict network access to OBS as tightly as possible using reverse proxy ACLs, VPN-only access, source IP allowlists, and segmentation. Do not rely on obscurity or non-default URLs. If the product is internet-facing, that exposure should be treated as higher priority until patched.
You should also review whether any sensitive information may already have been disclosed. That means checking logs for anomalous anonymous access, reviewing exported or downloaded response paths, and rotating credentials or tokens if the application could have exposed them. Since the record does not specify the data classes involved, assume sensitive operational data may have been visible and scope impact accordingly.
Technical Notes
Because Proliz has not published a source-backed universal package command in the materials provided, the exact upgrade command depends on how OBS was deployed in your environment. Use your standard software delivery path, but verify that the installed application version becomes v3.6.0 or later after maintenance.
Generic Linux package verification example:
obs_binary --version
# or
rpm -qa | grep -i proliz
# or
dpkg -l | grep -i proliz
If OBS is containerized, verify the image tag and redeploy only with the fixed release:
docker ps --format '{{.Image}} {{.Names}}' | grep -i obs
docker pull <your-registry>/proliz-obs:v3.6.0
docker stop <container_name> && docker rm <container_name>
docker run -d --name <container_name> <your-registry>/proliz-obs:v3.6.0
If immediate patching is blocked, enforce a temporary reverse-proxy restriction to trusted source ranges only:
location / {
allow 10.0.0.0/8;
allow 192.168.0.0/16;
deny all;
proxy_pass http://proliz_obs_backend;
}
And verify externally that the service is no longer broadly reachable:
curl -i https://your-obs-host.example.com/
In the absence of vendor-published endpoint-level workarounds, the safest temporary control is to remove or strictly gate untrusted network access until v3.6.0 is deployed and validated.
References
The primary source for the vulnerability record is the MITRE/CVE Services entry for CVE-2026-7189, which provides the affected version boundary, fixed version cutoff, CVSS vector, CWE mapping, and advisory linkage. The NVD entry is also relevant, especially because it currently reflects a higher score, though the CNA vector remains the most technically specific scoring basis in the available materials.
The advisory reference associated with the CVE is TR-26-0571 from TR-CERT. However, based on the retrieved material used here, that page did not expose additional technical details beyond what is already present in the CVE record. CISA KEV is useful as an exploitation-status check and currently does not list this CVE.
- MITRE CVE record API: https://cveawg.mitre.org/api/cve/CVE-2026-7189
- NVD entry: https://nvd.nist.gov/vuln/detail/CVE-2026-7189
- TR-CERT advisory TR-26-0571: https://siberguvenlik.gov.tr/guvenlik-bildirimleri/detay/tr-26-0571
- CISA Known Exploited Vulnerabilities catalog: https://www.cisa.gov/known-exploited-vulnerabilities-catalog
For defenders, the key takeaways remain stable even with sparse public detail: Proliz’s OBS before v3.6.0 is affected, v3.6.0 is the fix boundary, exploitation is not currently confirmed, and internet-exposed deployments should be prioritized for upgrade and log review immediately.
Additionally, for enhancing your digital privacy while managing vulnerabilities, consider using a VPN. Check out our comparisons of the best VPN for digital privacy 2026: Proton vs. Mullvad to find the right fit for your needs.
This article may contain affiliate links. We earn a commission on qualifying purchases at no extra cost to you.