Stored XSS via SVG Upload in ComfyUI /view Endpoint

ComfyUI is a modular diffusion model GUI, api and backend with a graph/nodes interface. Prior to 0.28.0, the /view endpoint served uploaded SVG files inline because image/svg+xml and related XML content types were absent from the dangerous-content-type handling, allowing stored cross-site scripting in the ComfyUI origin. This issue is fixed in version 0.28.0.

Casky was already ahead

This CVE exploits attack patterns that Casky's 322matched skills already investigate — long before this vulnerability was disclosed. Claude's reasoning model maps these techniques to MITRE ATT&CK, so practitioners who ran these skills have already seen the threat behaviour in their findings.

Analysis

ComfyUI versions prior to 0.28.0 contain a stored cross-site scripting (XSS) vulnerability in the /view endpoint that fails to properly sanitize uploaded SVG files. SVG and related XML content types were not included in the platform's dangerous-content-type filtering, allowing attackers to upload malicious SVG files containing embedded JavaScript that executes in the ComfyUI origin when viewed. This affects any organization or individual using ComfyUI for diffusion model operations, particularly those with multi-user deployments or public-facing instances. The vulnerability enables persistent attacks where a single malicious upload can compromise multiple users' sessions, steal credentials, or facilitate lateral movement within the broader infrastructure.

Casky's 322 mapped security skills leverage Claude AI with extended reasoning to detect the attack patterns underlying this vulnerability across MITRE ATT&CK's Initial Access (TA0001) and Persistence (TA0043) techniques. Practitioners using Casky would identify this threat through detection of suspicious file upload behaviors, improper content-type validation patterns, and SVG/XML handling in web endpoints. The platform's skills enable practitioners to recognize how attackers exploit CWE-79 (Improper Neutralization of Input During Web Page Generation) by examining request-response chains where user-controlled SVG content reaches the DOM unfiltered. Security teams would see findings highlighting the absence of content-security-policy headers, unsafe inline script execution in SVG contexts, and inadequate input validation—all precursors to successful exploitation that Casky's reasoning engine correlates with real-world attack chains.

Live Threat IntelligenceComing soon

Casky Risk Score

—

EPSS Probability

—

Shodan Exposure

—

GreyNoise Signal

—

Composite risk scoring from EPSS, CISA KEV, Shodan, and GreyNoise — 21 security APIs correlated into a single Casky Risk Score. Coming in Casky Pro. Join early access →

322 Casky skills cover this attack pattern

These skills use Claude AI's reasoning model to surface findings in the same attack categories as CVE-2026-56670.

Abuse of Public-Facing API: Mobile API Abuse

mobile security · low

Run

abusing-dpapi-for-credential-access

red teaming · high

Run

abusing-shadow-credentials-for-privesc

red teaming · high

Run

MITRE ATT&CK Techniques

TA0001 TA0043

Investigate the attack patterns behind CVE-2026-56670

Casky has 322 skills that investigate the attack patterns behind CVE-2026-56670. Run one and get CVSS-scored findings in 3 minutes.

Run the skill that detects this →

X Instagram LinkedIn