Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
ReconObtain valid ComfyUI credentials
-
DeliveryAuthenticate to ComfyUI service
-
ExploitCraft HTTP request with path traversal payload
-
InstallSubmit request to get_model_preview endpoint
-
C2Endpoint resolves traversal sequence to target file
-
ExecuteServer returns file contents to attacker
-
ImpactAttacker exfiltrates sensitive data
Vulnerability Assessment
| Weakness |
CWE
OWASP Top 10:2025
A01:2025 - Broken Access Control
ASVS 5.0
V5 - File Handling Path canonicalization, traversal prevention
MITRE ATT&CK
2 techniques across 2 tactics: Discovery, Collection The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up. |
| Exploitation Conditions | Exploitation requires valid authentication to ComfyUI (PR:L indicates authenticated user or lower-privileged account access). The attack is limited to read-only file access; attackers cannot modify, delete, or create files on the server. The endpoint must be network-accessible, but no special configuration or non-default settings are described as prerequisites - the vulnerability exists in the default Model Preview Endpoint functionality. Attackers are constrained by file system permissions of the ComfyUI process owner; files readable by the ComfyUI service can be accessed, but files with stricter permissions cannot. The vulnerability does NOT allow unauthenticated access; an attacker must first obtain or compromise valid ComfyUI credentials. Non-default configurations (such as restricting the Preview Endpoint to localhost-only or behind authentication layers beyond ComfyUI's built-in system) would limit real-world exploitability but are not standard deployments. |
| Impact | CVSS score of 4.3 with network attack vector and low complexity indicates a moderate baseline risk. However, several factors reduce real-world severity: the vulnerability requires authenticated access (PR:L), limiting the attack surface to users with valid credentials or account compromise, and the impact is read-only (Confidentiality:Low, Integrity:None, Availability:None), preventing direct data modification or service disruption. The presence of public exploit code (E:P) and VulDB reporting elevates practical risk, as malicious actors with credentials can now easily extract sensitive files. The vendor's non-response to early disclosure is concerning for timely patch availability. CVSS Environmental Score (RL:X, RC:R) reflects remediation uncertainty. This is a real but contextual priority: high-risk in multi-tenant or shared ComfyUI deployments where untrusted users have accounts, lower-risk in isolated or private installations. The authentication requirement significantly limits the attack surface compared to a remote unauthenticated equivalent. |
| Exploit ScenarioDerived | An authenticated user with a valid ComfyUI account crafts a malicious request to the Model Preview Endpoint, injecting a path traversal payload such as '../../../etc/passwd' or similar sequences into the model preview parameter. The endpoint, failing to sanitize the path, resolves the traversal sequence and returns the contents of the target file. The attacker downloads sensitive files including configuration files (containing API keys or database credentials), model weights, or system information, potentially enabling further compromise or data exfiltration. Public exploit code (available on the referenced Gist) automates this process. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Apr 20, 2026
Path traversal in ComfyUI up to version 0.13.0 allows authenticated remote attackers to read arbitrary files via manipulation of the get_model_preview function in the Model Preview Endpoint. An attacker with valid credentials can traverse the file system to access sensitive configuration files, model weights, or other data outside intended directories. Public exploit code is available, and the vendor has not provided a patched version despite early disclosure notification.
RemediationAI
No vendor-released patch has been identified at time of analysis, and the vendor did not respond to early disclosure. Immediate mitigation steps include: upgrade ComfyUI to the latest available version beyond 0.13.0 if a newer build is available from the official repository, though patch confirmation is lacking. As a compensating control, restrict network access to the ComfyUI Model Preview Endpoint using a Web Application Firewall (WAF) or reverse proxy to block requests containing path traversal sequences ('..', '/', backslashes); this prevents exploitation but may impact legitimate preview functionality if the endpoint uses relative paths for valid operations. Alternatively, implement strict input validation within the model manager to reject any model preview requests containing path traversal characters before file access. For shared or multi-tenant deployments, immediately audit file access logs for suspicious path traversal attempts and revoke credentials for accounts that may have been compromised. Monitor the official ComfyUI GitHub repository for security updates or community-maintained forks that may provide patches. Trade-off: WAF blocking may cause false positives if the application requires forward slashes in legitimate model names; coordinate with ComfyUI developers or community maintainers to clarify safe character sets.
Affected VersionsENISA
Affected Packages
Source: OSV (Open Source Vulnerabilities)
Affects your stack
Affects 1 package in euvd
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
ComfyUI is a node-based UI for Stable Diffusion and other generative AI models. The vulnerability resides in the app/model_manager.py file within the get_model_preview endpoint, which is designed to retrieve preview images for AI models. The root cause is CWE-22 (Improper Limitation of a Pathname to a Restricted Directory), meaning the endpoint fails to properly validate or sanitize file path parameters before using them in file system operations. An attacker can inject path traversal sequences (such as '../../../') into the model preview request to escape the intended model directory and access arbitrary files on the server. The vulnerability is classified under path traversal attacks and affects the Model Preview Endpoint specifically.
Know when Comfyui gets its next vulnerability
- Watch products and get an email when something new is published
- Check this CVE against your own stack and servers
- Exploit and PoC references, threat intelligence from 22 sources
No credit card · 7-day full trial
Severity by source
CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:L/VI:N/VA:N/SC:N/SI:N/SA:N/E:P
Primary rating from NVD · only source for this CVE.
CVSS VectorNVD
Lifecycle Timeline
5DescriptionCVE.org
A vulnerability was detected in ComfyUI up to 0.13.0. This impacts the function get_model_preview of the file app/model_manager.py of the component Model Preview Endpoint. The manipulation results in path traversal. The attack may be launched remotely. The exploit is now public and may be used. The vendor was contacted early about this disclosure but did not respond in any way.
Threat Intelligence Overview
How this decision was reached
The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.
What every action level means
SSVC AssessmentENISA
Priority Score
Attack Technique
Path traversal exploits occur when applications use user-controlled input to construct file system paths without proper validation.
Related Vulnerabilities in ComfyUI Multiple Vulnerabilities
View all 5Arbitrary file write via path traversal in ComfyUI before 0.30.0 lets an attacker place attacker-controlled content anyw
comfyanonymous/comfyui version v0.2.4 suffers from a non-blind Server-Side Request Forgery (SSRF) vulnerability. Rated h
A stored cross-site scripting (XSS) vulnerability exists in comfyanonymous/comfyui version 0.2.2 and possibly earlier. R
Remote code execution in ComfyUI v0.23.0 (and all prior 0.x releases) lets unauthenticated attackers run arbitrary Pytho
Stored cross-site scripting in ComfyUI (Comfy-Org) before 0.28.0 lets an attacker upload HTML or SVG to the per-user use
Stored cross-site scripting in ComfyUI before 0.28.0 lets an attacker upload a malicious SVG that the /view endpoint ser
Unauthenticated information disclosure in ComfyUI (comfy-org/comfyui) before 0.28.0 lets remote attackers read arbitrary
Path traversal arbitrary file read in ComfyUI (comfy-org) before 0.28.0 lets remote unauthenticated attackers probe arbi
EUVD-2026-23733
NVD
CVE.org
EPSS
Vulnrichment
GitHub
VulDB
Exploit-DB