CVE-2026-109298: Open WebUI: regex DoS via grep tools exhausts memory

MEDIUMCISA: TRACK*
Published October 12, 2026
CISO Take

Open WebUI 0.11.0 through 0.11.3 compiles user-chosen grep patterns in its knowledge and chat file search tools with a regex engine that expands nested counted repetition without limit and caches the result, so a short pattern can block the event loop and exhaust worker memory. Any authenticated low-privilege user can trigger it by getting the model to call grep_knowledge_files, so the barrier is an ordinary account, not admin rights. Exploitation is low-complexity, but there is no CISA KEV listing, no public exploit, no Nuclei template and no EPSS score, so near-term mass exploitation looks unlikely. Because the cached result persists, a single bad pattern can degrade the whole instance and affect every user sharing the workers. Upgrade to Open WebUI 0.11.4 or later, and until then restrict registrations and monitor worker memory and event-loop latency.

Sources:NVDGitHub AdvisoryATLASVulnCheck

What is the risk?

Medium (CVSS 6.5, AV:N/AC:L/PR:L/UI:N, availability impact high only). Exploitation needs a valid account and the ability to prompt the model into calling the grep tool, which is low effort. No confidentiality or integrity impact. No known active exploitation, KEV listing, public exploit or scanner template, and EPSS is unavailable. Exposure is highest on multi-user or internet-facing instances with open sign-up. The package has 184 other CVEs, which signals a heavy vulnerability history, though only 3 downstream dependents limit the supply-chain blast radius.

How does the attack unfold?

Initial Access
Attacker signs in with a valid low-privilege Open WebUI account.
AML.T0012
Tool Invocation
Attacker prompts the model to call grep_knowledge_files with a short nested-repetition regex.
AML.T0053
Resource Exhaustion
The regex engine expands the pattern at compile time and caches it, blocking the event loop and consuming worker memory.
AML.T0034.002
Impact
Open WebUI becomes unresponsive for all users until workers are restarted, and the attack can be repeated.
AML.T0029

What systems are affected?

PackageEcosystemVulnerable RangePatched
Open WebUIpip—No patch
156.7K3 dependentsPushed 6d ago76% patched~20d to patchFull package profile →

Do you use Open WebUI? You're affected.

How severe is it?

CVSS 3.1
6.5 / 10
EPSS
0.5%
chance of exploitation in 30 days
Higher than 41% of all CVEs
Exploitation Status
Exploit Available
Exploitation: MEDIUM
Sophistication
Trivial
Exploitation Confidence
medium
○CISA SSVC: Public PoC
Composite signal derived from CISA KEV, VulnCheck KEV, CISA SSVC, EPSS, Metasploit, Exploit-DB, trickest/cve, Nuclei templates, and inthewild.io exploitation reports.

What is the attack surface?

AVACPRUISCIA
AVNetwork
ACLow
PRLow
UINone
SUnchanged
CNone
INone
AHigh

What should I do?

1 step
  1. 1) Upgrade open-webui to 0.11.4 or later (fix commit 97e013a6). 2) Until patched, disable or restrict the knowledge and chat file grep tools, for example through tool and function permissions, and limit who can use knowledge bases. 3) Disable open self-registration and review low-privilege accounts. 4) Set container memory limits and restart policies, and run multiple workers behind a health-checked proxy. 5) Detect: alert on event-loop latency spikes, worker memory growth and OOM restarts, and review logs for grep_knowledge_files calls with patterns containing nested quantifiers such as (a{1000}){1000}. 6) Rate-limit authenticated API usage per account.

What does CISA's SSVC say?

DecisionTrack*
Exploitationpoc
AutomatableNo
Technical Impactpartial

Source: CISA Vulnrichment (SSVC v2.0). Decision based on the CISA Coordinator decision tree.

How is it classified?

Which compliance frameworks are affected?

This CVE is relevant to:

EU AI Act
Art. 15-Accuracy, robustness and cybersecurity
ISO 42001
A.6.2.6-AI system operation and monitoring
NIST AI RMF
MANAGE 2.3-Respond to unknown risks and resource failures
OWASP LLM Top 10
LLM06-Excessive AgencyLLM10-Unbounded Consumption

Frequently Asked Questions

What is CVE-2026-109298?

Open WebUI 0.11.0 through 0.11.3 compiles user-chosen grep patterns in its knowledge and chat file search tools with a regex engine that expands nested counted repetition without limit and caches the result, so a short pattern can block the event loop and exhaust worker memory. Any authenticated low-privilege user can trigger it by getting the model to call grep_knowledge_files, so the barrier is an ordinary account, not admin rights. Exploitation is low-complexity, but there is no CISA KEV listing, no public exploit, no Nuclei template and no EPSS score, so near-term mass exploitation looks unlikely. Because the cached result persists, a single bad pattern can degrade the whole instance and affect every user sharing the workers. Upgrade to Open WebUI 0.11.4 or later, and until then restrict registrations and monitor worker memory and event-loop latency.

Is CVE-2026-109298 actively exploited?

No confirmed active exploitation of CVE-2026-109298 has been reported, but organizations should still patch proactively.

How to fix CVE-2026-109298?

1) Upgrade open-webui to 0.11.4 or later (fix commit 97e013a6). 2) Until patched, disable or restrict the knowledge and chat file grep tools, for example through tool and function permissions, and limit who can use knowledge bases. 3) Disable open self-registration and review low-privilege accounts. 4) Set container memory limits and restart policies, and run multiple workers behind a health-checked proxy. 5) Detect: alert on event-loop latency spikes, worker memory growth and OOM restarts, and review logs for grep_knowledge_files calls with patterns containing nested quantifiers such as (a{1000}){1000}. 6) Rate-limit authenticated API usage per account.

What systems are affected by CVE-2026-109298?

This vulnerability affects the following AI/ML architecture patterns: LLM chat web interfaces, RAG pipelines, Agent tool-calling frameworks, Self-hosted model front-ends.

What is the CVSS score for CVE-2026-109298?

CVE-2026-109298 has a CVSS v3.1 base score of 6.5 (MEDIUM). The EPSS exploitation probability is 0.47%.

What is the AI security impact?

Affected AI Architectures

LLM chat web interfacesRAG pipelinesAgent tool-calling frameworksSelf-hosted model front-ends

MITRE ATLAS Techniques

AML.T0012Valid Accounts
AML.T0029Denial of AI Service
AML.T0034.002Agentic Resource Consumption
AML.T0053AI Agent Tool Invocation

Compliance Controls Affected

EU AI Act: Art. 15
ISO 42001: A.6.2.6
NIST AI RMF: MANAGE 2.3
OWASP LLM Top 10: LLM06, LLM10

What are the technical details?

Original Advisory

Open WebUI 0.11.0 before 0.11.4 compiles user-chosen grep patterns for its knowledge and chat file search tools with a regex engine that expands nested counted repetition unbounded at compile time and caches the result. Authenticated users can have the model call grep_knowledge_files with short nested-repetition patterns, blocking the event loop and exhausting worker memory.

Exploitation Scenario

An attacker with a basic account, from open sign-up or a compromised low-privilege user, opens a chat with knowledge search enabled. They instruct the model to run grep_knowledge_files with a short pattern using nested counted repetition. The regex engine expands the pattern at compile time and caches the result, blocking the event loop and driving worker memory up until the process stalls or is OOM-killed. The attacker repeats the request after each restart to keep the AI platform down for all users.

Weaknesses (CWE)

CWE-400 — Uncontrolled Resource Consumption: The product does not properly control the allocation and maintenance of a limited resource.

  • [Architecture and Design] Design throttling mechanisms into the system architecture. The best protection is to limit the amount of resources that an unauthorized user can cause to be expended. A strong authentication and access control model will help prevent such attacks from occurring in the first place. The login application should be protected against DoS attacks as much as possible. Limiting the database access, perhaps by caching result sets, can help minimize the resources expended. To further limit the potential for a DoS attack, consider tracking the rate of requests received from users and blocking requests that exceed a defined rate threshold.
  • [Architecture and Design] Mitigation of resource exhaustion attacks requires that the target system either: The first of these solutions is an issue in itself though, since it may allow attackers to prevent the use of the system by a particular valid user. If the attacker impersonates the valid user, they may be able to prevent the user from accessing the server in question. The second solution is simply difficult to effectively institute -- and even when properly done, it does not provide a full solution. It simply makes the attack require more resources on the part of the attacker. recognizes the attack and denies that user further access for a given amount of time, or uniformly throttles all requests in order to make it more difficult to consume resources more quickly than they can again be freed.

Source: MITRE CWE corpus.

CVSS Vector

CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H

Timeline

Published
October 12, 2026
Last Modified
October 23, 2026
First Seen
October 12, 2026

Related Vulnerabilities