Aix-DB SQLi flaw exposes databases via unauthenticated endpoint
CVE-2026-8335 allows unauthenticated attackers to execute arbitrary SELECT queries against Aix-DB instances via a missing auth check on the /llm/process_llm_out endpoint.
Sourcecert.pl — open the original
What to check
- Applies to
- The 1 CVE named below. Whether you run an affected version is not assessed: check the versions the source lists.
- Exploitation
- No exploitation in the wild and no public exploit is recorded for the CVEs named.
CERT Polska disclosed CVE-2026-8335, a missing authentication vulnerability in Aix-DB software that permits unauthenticated remote attackers to execute arbitrary SELECT SQL queries and exfiltrate database contents. The flaw exists in the /llm/process_llm_out endpoint, which fails to implement the token validation enforced across other application endpoints.
All Aix-DB releases through version 1.2.4 are confirmed vulnerable. As of publication, the vendor has not issued a patch, and the vulnerability status of subsequent releases remains unknown. The disclosure follows a coordinated process initiated by a responsible report from researcher Eryk Winiarz.
Mentioned in this report
Detection guidance
Unauthenticated Access to Aix-DB /llm/process_llm_out Endpoint
A behavioural hunt for the technique in general, not a signature for the CVEs this report names.
Detects HTTP requests to the vulnerable Aix-DB endpoint that lacks token validation, indicating potential exploitation of CVE-2026-8335. Auto-generated starting point — validate and tune in your environment before deploying. IOC matches can false-positive on shared infrastructure and decay as adversary infrastructure rotates.
title: Unauthenticated Access to Aix-DB Vulnerable Endpoint
id: 8048497b-6220-56a9-b874-b59f6ce975c3
status: experimental
description: 'Detects HTTP requests to the /llm/process_llm_out endpoint in Aix-DB,
which is
affected by CVE-2026-8335, a missing authentication vulnerability that allows
unauthenticated attackers to execute arbitrary SELECT SQL queries. Legitimate
traffic to this endpoint should always carry a validated auth token; requests
without one, or from unexpected sources, may indicate exploitation attempts.
'
references:
- CVE-2026-8335
author: Vorant
tags:
- attack.t1190
logsource:
category: webserver
detection:
selection:
cs-uri-stem|contains: /llm/process_llm_out
filter:
cs-uri-query|contains: token=
condition: selection and not filter
falsepositives:
- Legitimate application traffic that includes a valid token parameter but is logged
differently by the web server
- Internal health checks or monitoring probes hitting the endpoint with a valid session
cookie instead of a query token
level: high
Potential SQL Injection Payload in Aix-DB LLM Endpoint Request
A behavioural hunt for the technique in general, not a signature for the CVEs this report names.
Detects SQL injection indicators (UNION SELECT, information_schema, SLEEP, etc.) in requests to the Aix-DB /llm/process_llm_out endpoint used to exploit CVE-2026-8335. Auto-generated starting point — validate and tune in your environment before deploying. IOC matches can false-positive on shared infrastructure and decay as adversary infrastructure rotates.
title: SQL Injection Pattern in Aix-DB Endpoint Request
id: bd243e99-239b-5bee-9a3d-6bf68193d66c
status: experimental
description: 'Detects common SQL injection syntax (UNION SELECT, information_schema,
SLEEP, benchmark, or stacked SELECT statements) sent to the vulnerable
Aix-DB /llm/process_llm_out endpoint. This endpoint accepts arbitrary
SELECT SQL queries without authentication (CVE-2026-8335), so the presence
of SQL keywords in the request body or query string is a strong indicator
of exploitation and data exfiltration attempts.
'
references:
- CVE-2026-8335
author: Vorant
tags:
- attack.t1190
- attack.t1213
logsource:
category: webserver
detection:
selection_endpoint:
cs-uri-stem|contains: /llm/process_llm_out
selection_sqli:
cs-uri-query|contains:
- UNION SELECT
- union select
- information_schema
- SLEEP(
- sleep(
- benchmark(
- --
- ''' OR ''1''=''1'
condition: selection_endpoint and selection_sqli
falsepositives:
- Automated vulnerability scanners or penetration testing tools intentionally probing
the endpoint
- Security research or bug bounty validation traffic against the same endpoint
level: high
Repeated SELECT-Only Queries via Aix-DB Public Endpoint Indicating Bulk Data Exfiltration
A behavioural hunt for the technique in general, not a signature for the CVEs this report names.
Detects a pattern of successive SELECT-style queries submitted to the unauthenticated Aix-DB endpoint, consistent with bulk exfiltration of database contents after exploiting CVE-2026-8335. Auto-generated starting point — validate and tune in your environment before deploying. IOC matches can false-positive on shared infrastructure and decay as adversary infrastructure rotates.
title: Bulk SELECT Query Pattern Against Aix-DB Public Endpoint
id: 50c75f82-7a89-5db3-ab0d-1695576d5b4b
status: experimental
description: 'Detects HTTP requests to the vulnerable /llm/process_llm_out endpoint
that
carry SELECT statements targeting sensitive database objects (e.g. users,
credentials, accounts tables). Because the endpoint lacks authentication,
attackers can freely query and exfiltrate database contents; this rule
flags requests referencing sensitive table/column names commonly targeted
for data theft, per CVE-2026-8335. Analysts should review request volume
and source over time to confirm bulk exfiltration.
'
references:
- CVE-2026-8335
author: Vorant
tags:
- attack.t1213
- attack.t1190
logsource:
category: webserver
detection:
selection_endpoint:
cs-uri-stem|contains: /llm/process_llm_out
selection_select:
cs-uri-query|contains: SELECT
selection_sensitive:
cs-uri-query|contains:
- password
- credential
- user
- account
- token
- secret
condition: selection_endpoint and selection_select and selection_sensitive
falsepositives:
- Legitimate administrative queries against user/account tables performed through
an approved, authenticated integration
- Internal reporting tools that happen to query similarly named tables via a whitelisted
service account
level: medium
Behavioural rules are generated from public reporting — validate in your environment before deploying.
Source reporting: https://cert.pl/en/posts/2026/06/CVE-2026-8335/
Free public brief. The detections with Splunk SPL already written (plus Elastic and KQL where they apply) are in the app with a free account, no card. Slack alerts on the vendors you run come with the paid plans, and a new account gets three days of them.
What this brief leaves out
This page is the free tier: the write-up, the severity, the names. In the app the same report carries its detections (Splunk SPL already written, plus Elastic and Microsoft KQL where they apply) and live profiles of the CVEs it names. The paid plans add the security research and news reporting the free tier leaves out.
It sits in a corpus of 16,445 reports from 149 sources, 666 of them written in the last seven days, and it grows through the day.
A new account starts with three days of all of it, no card, then keeps the government and CERT reporting free. Slack alerts are a paid-plan feature: name the vendors you run and Slack hears about the reports that touch them, not the rest of the week's. Nothing is charged unless you subscribe.
Create a free account What it costs