CVE-2024-42474 — MEDIUM (CVSS 6.5) AI Security Vulnerability

CISO Take

Streamlit deployments on Windows with static file sharing enabled expose the running user's NTLM password hash to any low-privileged attacker. That hash can be cracked offline or weaponized in pass-the-hash attacks for lateral movement. Patch to v1.37.0 immediately and audit whether static file sharing is active across your data science environment.

Risk Assessment

CVSS 6.5 understates enterprise risk. Streamlit is pervasive in data science teams on Windows workstations and servers, frequently running under domain accounts with elevated privileges for database or GPU access. An exposed NTLM hash enables offline brute-force or direct pass-the-hash attacks, potentially granting access to broader enterprise resources. Risk escalates significantly in domain-joined environments where a single compromised account can cascade. The low attack complexity and no user interaction requirement mean exploitation is straightforward once a vulnerable instance is identified.

Affected Systems

Package	Ecosystem	Vulnerable Range	Patched
streamlit	pip	—	No patch
44.4K OpenSSF 7.2 2.8K dependents Pushed 7d ago 8% patched ~0d to patch Full package profile →

Do you use streamlit? You're affected.

Severity & Risk

CVSS 3.1

6.5 / 10

EPSS

1.7%

chance of exploitation in 30 days

Higher than 82% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

Exploit Available

Exploitation: MEDIUM

Sophistication

Trivial

Exploitation Confidence

medium

○ Public PoC indexed (trickest/cve)

Composite signal derived from CISA KEV, CISA SSVC, EPSS, trickest/cve, and Nuclei templates.

Attack Surface

AV Network

AC Low

PR Low

UI None

S Unchanged

C High

I None

A None

Recommended Action

6 steps

Patch: Upgrade Streamlit to v1.37.0 or later immediately—this is the only complete fix.
Disable: If patching is delayed, set enableStaticServing = false in .streamlit/config.toml or via environment variable STREAMLIT_SERVER_ENABLE_STATIC_SERVING=false.
Audit: Inventory all Streamlit deployments on Windows; grep config files and startup scripts for enableStaticServing.
Network controls: Block outbound SMB (TCP 445/139) from Streamlit hosts to prevent NTLM relay to attacker-controlled servers.
Monitor: Alert on path traversal patterns (../, ..\, URL-encoded variants) in static file request logs.
Rotate credentials: If exposure is suspected on any host, rotate credentials for accounts that ran Streamlit and check for lateral movement indicators.

CISA SSVC Assessment

Decision Track

Exploitation none

Automatable No

Technical Impact partial

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

Classification

Data Leakage Auth Bypass Framework API AML.T0025 - Exfiltration via Cyber Means AML.T0049 - Exploit Public-Facing Application AML.T0055 - Unsecured Credentials AML.T0106 - Exploitation for Credential Access

Compliance Impact

This CVE is relevant to:

EU AI Act

Art. 15 - Accuracy, robustness and cybersecurity

ISO 42001

A.6.2.6 - Information security for AI systems

NIST AI RMF

MANAGE-2.2 - Mechanisms are in place and applied to sustain the value of AI and to manage associated risks

OWASP LLM Top 10

LLM06:2025 - Excessive Agency

Frequently Asked Questions

What is CVE-2024-42474?

Streamlit deployments on Windows with static file sharing enabled expose the running user's NTLM password hash to any low-privileged attacker. That hash can be cracked offline or weaponized in pass-the-hash attacks for lateral movement. Patch to v1.37.0 immediately and audit whether static file sharing is active across your data science environment.

Is CVE-2024-42474 actively exploited?

Proof-of-concept exploit code is publicly available for CVE-2024-42474, increasing the risk of exploitation.

How to fix CVE-2024-42474?

1. Patch: Upgrade Streamlit to v1.37.0 or later immediately—this is the only complete fix. 2. Disable: If patching is delayed, set enableStaticServing = false in .streamlit/config.toml or via environment variable STREAMLIT_SERVER_ENABLE_STATIC_SERVING=false. 3. Audit: Inventory all Streamlit deployments on Windows; grep config files and startup scripts for enableStaticServing. 4. Network controls: Block outbound SMB (TCP 445/139) from Streamlit hosts to prevent NTLM relay to attacker-controlled servers. 5. Monitor: Alert on path traversal patterns (../, ..\, URL-encoded variants) in static file request logs. 6. Rotate credentials: If exposure is suspected on any host, rotate credentials for accounts that ran Streamlit and check for lateral movement indicators.

What systems are affected by CVE-2024-42474?

This vulnerability affects the following AI/ML architecture patterns: ML demo applications, model serving, data science internal tooling, ML prototyping environments.

What is the CVSS score for CVE-2024-42474?

CVE-2024-42474 has a CVSS v3.1 base score of 6.5 (MEDIUM). The EPSS exploitation probability is 1.65%.

Technical Details

NVD Description

Streamlit is a data oriented application development framework for python. Snowflake Streamlit open source addressed a security vulnerability via the static file sharing feature. Users of hosted Streamlit app(s) on Windows were vulnerable to a path traversal vulnerability when the static file sharing feature is enabled. An attacker could utilize the vulnerability to leak the password hash of the Windows user running Streamlit. The vulnerability was patched on Jul 25, 2024, as part of Streamlit open source version 1.37.0. The vulnerability only affects Windows.

Exploitation Scenario

An adversary with any authenticated access to a Streamlit app (legitimate user, insider, or anyone who can reach the endpoint) crafts a path traversal request targeting the static file serving endpoint. On Windows, this can reference a UNC path (e.g., \\attacker-controlled-server\share\file), forcing the Windows host to initiate NTLM authentication to the attacker's server. Using a tool like Responder, the attacker captures the NTLMv2 hash. They either crack it offline using a GPU cluster if the password is weak, or relay it directly in a pass-the-hash attack against other Windows systems in the environment. In an AI/ML context, the Streamlit host almost certainly has access to training data stores, model registries, MLflow tracking servers, or database credentials—making it an ideal lateral movement pivot into the broader AI infrastructure.