CVE-2024-41113: Streamlit RCE via eval()

CISO Take

Critical unauthenticated RCE in streamlit-geospatial — any public deployment is fully compromised with a single HTTP request crafted in seconds. Patch immediately to commit c4f81d96 or disable the Timelapse page and restrict network access. Organizations using this for Earth Engine workflows risk full server takeover including cloud credential exfiltration.

What is the risk?

Severity ceiling: CVSS 9.8, no authentication, no user interaction, network-exploitable with low complexity. The eval() pattern on user-controlled vis_params input is a textbook code injection — no AI/ML knowledge required to exploit, any script-kiddie with Shodan access can own this. Streamlit apps are routinely deployed for internal data science collaboration or exposed publicly for research, dramatically widening the attack surface. Not in KEV yet, but expect active exploitation given the GitHub Security Lab advisory and trivial exploit path.

What systems are affected?

Package	Ecosystem	Vulnerable Range	Patched
Streamlit	pip	—	No patch
45.0K OpenSSF 7.2 2.9K dependents Pushed 3d ago 7% patched ~0d to patch Full package profile →

Do you use Streamlit? You're affected.

How severe is it?

CVSS 3.1

9.8 / 10

EPSS

1.4%

chance of exploitation in 30 days

Higher than 69% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

Exploit Available

Exploitation: MEDIUM

Sophistication

Trivial

Exploitation Confidence

medium

○ CISA SSVC: Public PoC

○ Public PoC indexed (trickest/cve)

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?

AV Network

AC Low

PR None

UI None

S Unchanged

C High

I High

A High

What should I do?

6 steps

Update immediately to commit c4f81d9616d40c60584e36abb15300853a66e489 or disable the Timelapse page.
If patching is delayed, apply network ACLs to restrict access to trusted IPs only — do not leave publicly exposed.
Audit all other pages in the app for similar eval() or exec() patterns on user input.
Rotate any API keys accessible from the server (GEE tokens, cloud credentials, service accounts).
Review server logs for anomalous vis_params values indicating prior exploitation attempts.
Implement a WAF rule blocking Python built-in patterns (__import__, exec, subprocess) in input parameters as a compensating control.

What does CISA's SSVC say?

Decision Attend

Exploitation poc

Automatable Yes

Technical Impact total

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

How is it classified?

Code Execution Data Extraction Framework AML.T0025 - Exfiltration via Cyber Means AML.T0037 - Data from Local System AML.T0049 - Exploit Public-Facing Application AML.T0050 - Command and Scripting Interpreter AML.T0072 - Reverse Shell

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 security measures

NIST AI RMF

MANAGE 2.4 - Risks and residual risks are treated and managed

OWASP LLM Top 10

LLM05:2025 - Improper Output Handling

Frequently Asked Questions

What is CVE-2024-41113?

Critical unauthenticated RCE in streamlit-geospatial — any public deployment is fully compromised with a single HTTP request crafted in seconds. Patch immediately to commit c4f81d96 or disable the Timelapse page and restrict network access. Organizations using this for Earth Engine workflows risk full server takeover including cloud credential exfiltration.

Is CVE-2024-41113 actively exploited?

Proof-of-concept exploit code is publicly available for CVE-2024-41113, increasing the risk of exploitation.

How to fix CVE-2024-41113?

1. Update immediately to commit c4f81d9616d40c60584e36abb15300853a66e489 or disable the Timelapse page. 2. If patching is delayed, apply network ACLs to restrict access to trusted IPs only — do not leave publicly exposed. 3. Audit all other pages in the app for similar eval() or exec() patterns on user input. 4. Rotate any API keys accessible from the server (GEE tokens, cloud credentials, service accounts). 5. Review server logs for anomalous vis_params values indicating prior exploitation attempts. 6. Implement a WAF rule blocking Python built-in patterns (__import__, exec, subprocess) in input parameters as a compensating control.

What systems are affected by CVE-2024-41113?

This vulnerability affects the following AI/ML architecture patterns: geospatial AI/ML deployments, Streamlit-based ML web interfaces, data science web applications, academic and research ML environments, Earth Engine integration pipelines.

What is the CVSS score for CVE-2024-41113?

CVE-2024-41113 has a CVSS v3.1 base score of 9.8 (CRITICAL). The EPSS exploitation probability is 1.40%.

What is the AI security impact?

Affected AI Architectures

geospatial AI/ML deploymentsStreamlit-based ML web interfacesdata science web applicationsacademic and research ML environmentsEarth Engine integration pipelines

MITRE ATLAS Techniques

AML.T0025 Exfiltration via Cyber Means

AML.T0037 Data from Local System

AML.T0049 Exploit Public-Facing Application

AML.T0050 Command and Scripting Interpreter

AML.T0072 Reverse Shell

Compliance Controls Affected

EU AI Act: Article 15

ISO 42001: A.6.2.6

NIST AI RMF: MANAGE 2.4

OWASP LLM Top 10: LLM05:2025

What are the technical details?

Original Advisory

streamlit-geospatial is a streamlit multipage app for geospatial applications. Prior to commit c4f81d9616d40c60584e36abb15300853a66e489, the `vis_params` variable on line 383 or line 390 in `pages/1_📷_Timelapse.py` takes user input, which is later used in the `eval()` function on line 395, leading to remote code execution. Commit c4f81d9616d40c60584e36abb15300853a66e489 fixes this issue.

Exploitation Scenario

Adversary discovers a publicly exposed streamlit-geospatial instance via Shodan search for 'streamlit port:8501' or Google dork. They craft an HTTP POST to the Timelapse page setting vis_params to a Python expression such as __import__('subprocess').check_output(['env']) to enumerate environment variables. Within seconds they recover Google Earth Engine tokens, cloud IAM credentials, and database connection strings. They then establish a reverse shell via __import__('socket') for persistent C2 access, pivot to cloud storage to exfiltrate training datasets and model weights, and potentially move laterally into the broader cloud environment. The entire kill chain requires zero ML expertise.

Weaknesses (CWE)

CWE-20 Improper Input Validation

CWE-20 — Improper Input Validation: The product receives input or data, but it does not validate or incorrectly validates that the input has the properties that are required to process the data safely and correctly.

[Architecture and Design] Consider using language-theoretic security (LangSec) techniques that characterize inputs using a formal language and build "recognizers" for that language. This effectively requires parsing to be a distinct layer that effectively enforces a boundary between raw input and internal data representations, instead of allowing parser code to be scattered throughout the program, where it could be subject to errors or inconsistencies that create weaknesses. [REF-1109] [REF-1110] [REF-1111]
[Architecture and Design] Use an input validation framework such as Struts or the OWASP ESAPI Validation API. Note that using a framework does not automatically address all input validation problems; be mindful of weaknesses that could arise from misusing the framework itself (CWE-1173).

Source: MITRE CWE corpus.