CVE-2024-10188: LiteLLM unauthenticated DoS crashes

CISO Take

Any litellm deployment exposed to the network — including internal AI gateways — can be crashed by an unauthenticated attacker with a single crafted request, taking down all LLM routing for dependent applications. Patch to 1.53.1.dev1 immediately; if you cannot patch, place litellm behind an authenticated reverse proxy or WAF as a stopgap. Audit whether litellm endpoints are internet-reachable — many teams expose them naively during POC phases.

What is the risk?

Effective risk is HIGH for any organization using litellm as an LLM gateway or proxy. CVSS 7.5 is accurate: no authentication, no user interaction, network-reachable, and low attack complexity make this trivially exploitable. EPSS (0.00129) is currently low suggesting no observed mass exploitation, but the attack is simple enough that any motivated actor can reproduce it from the public huntr disclosure. The blast radius is limited to availability — no confidentiality or integrity impact — but in AI-dependent workflows, a downed LLM proxy is a full service outage.

What systems are affected?

Package	Ecosystem	Vulnerable Range	Patched
LiteLLM	pip	< 1.53.1.dev1	`1.53.1.dev1`
51.0K OpenSSF 6.1 6 dependents Pushed 2d ago 38% patched ~38d to patch Full package profile →

Do you use LiteLLM? You're affected.

How severe is it?

CVSS 3.1

7.5 / 10

EPSS

0.5%

chance of exploitation in 30 days

Higher than 40% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

No known exploitation

Sophistication

Trivial

What is the attack surface?

AV Network

AC Low

PR None

UI None

S Unchanged

C None

I None

A High

What should I do?

6 steps

PATCH

Upgrade litellm to >= 1.53.1.dev1 immediately (commit 21156ff5).
WORKAROUND (if patch not possible): Place litellm behind an authenticated reverse proxy (nginx + basic auth / mTLS) to eliminate unauthenticated access.
NETWORK CONTROLS

Ensure litellm is not internet-facing; restrict access to known internal CIDR ranges via firewall rules.
RATE LIMITING

Apply request rate limits at the proxy/WAF layer to reduce DoS surface even post-patch.
MONITORING

Alert on litellm process restarts or sudden spikes in 5xx errors from your LLM gateway — these are indicators of exploitation attempts.
DETECTION

Review logs for abnormally large or malformed request bodies to litellm endpoints.

How is it classified?

DoS Framework API Inference AML.T0029 - Denial of AI Service AML.T0034 - Cost Harvesting AML.T0049 - Exploit Public-Facing Application

Which compliance frameworks are affected?

This CVE is relevant to:

EU AI Act

Article 15 - Accuracy, robustness, and cybersecurity

ISO 42001

A.6.2.6 - AI system availability and resilience

NIST AI RMF

MANAGE-2.2 - Risks from third-party AI components

OWASP LLM Top 10

LLM04 - Model Denial of Service

Frequently Asked Questions

What is CVE-2024-10188?

Any litellm deployment exposed to the network — including internal AI gateways — can be crashed by an unauthenticated attacker with a single crafted request, taking down all LLM routing for dependent applications. Patch to 1.53.1.dev1 immediately; if you cannot patch, place litellm behind an authenticated reverse proxy or WAF as a stopgap. Audit whether litellm endpoints are internet-reachable — many teams expose them naively during POC phases.

Is CVE-2024-10188 actively exploited?

No confirmed active exploitation of CVE-2024-10188 has been reported, but organizations should still patch proactively.

How to fix CVE-2024-10188?

1. PATCH: Upgrade litellm to >= 1.53.1.dev1 immediately (commit 21156ff5). 2. WORKAROUND (if patch not possible): Place litellm behind an authenticated reverse proxy (nginx + basic auth / mTLS) to eliminate unauthenticated access. 3. NETWORK CONTROLS: Ensure litellm is not internet-facing; restrict access to known internal CIDR ranges via firewall rules. 4. RATE LIMITING: Apply request rate limits at the proxy/WAF layer to reduce DoS surface even post-patch. 5. MONITORING: Alert on litellm process restarts or sudden spikes in 5xx errors from your LLM gateway — these are indicators of exploitation attempts. 6. DETECTION: Review logs for abnormally large or malformed request bodies to litellm endpoints.

What systems are affected by CVE-2024-10188?

This vulnerability affects the following AI/ML architecture patterns: LLM proxy and gateway, agent frameworks, RAG pipelines, model serving, AI application backends.

What is the CVSS score for CVE-2024-10188?

CVE-2024-10188 has a CVSS v3.1 base score of 7.5 (HIGH). The EPSS exploitation probability is 0.53%.

What is the AI security impact?

Affected AI Architectures

LLM proxy and gatewayagent frameworksRAG pipelinesmodel servingAI application backends

MITRE ATLAS Techniques

AML.T0029 Denial of AI Service

AML.T0034 Cost Harvesting

AML.T0049 Exploit Public-Facing Application

Compliance Controls Affected

EU AI Act: Article 15

ISO 42001: A.6.2.6

NIST AI RMF: MANAGE-2.2

OWASP LLM Top 10: LLM04

What are the technical details?

Original Advisory

A vulnerability in BerriAI/litellm, as of commit 26c03c9, allows unauthenticated users to cause a Denial of Service (DoS) by exploiting the use of ast.literal_eval to parse user input. This function is not safe and is prone to DoS attacks, which can crash the litellm Python server.

Exploitation Scenario

An adversary identifies a litellm endpoint via passive DNS, GitHub leaks, or internal network scanning. They craft an HTTP request to any litellm API endpoint with a payload specifically designed to cause ast.literal_eval to enter a resource-exhaustive evaluation loop — for example, a deeply nested structure or an expression that triggers excessive memory allocation. No credentials are required. The Python server process crashes or becomes unresponsive, taking offline all LLM-dependent services (AI agents, RAG queries, copilot features) routing through that litellm instance. In a CI/CD or automated AI pipeline context, this could silently stall batch enrichment jobs or break production inference without immediate human awareness.

Weaknesses (CWE)

CWE-400 Uncontrolled Resource Consumption Primary CWE-400 Uncontrolled Resource Consumption

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.