What is the CVSS score of CVE-2026-50758?

CVE-2026-50758 has a CVSS 3.1 base score of 8.1 (High). CVSS vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:N. EPSS exploitation probability: 0.3%.

Which versions of next-ai-draw-io are affected by CVE-2026-50758?

DayuanJiang next-ai-draw-io 0.4.13 contains a reflected cross-site scripting vulnerability allowing attackers to execute arbitrary JavaScript in users' browsers through malicious links.

How to fix or mitigate CVE-2026-50758?

Within 24 hours, identify all instances of next-ai-draw-io 0.4.13 and alert users to avoid clicking suspicious links to the application. Within 7 days, implement Content Security Policy (CSP) headers to prevent inline script execution, apply input validation and output encoding to the mcp URL parameter, and review application logs for exploitation attempts. Within 30 days, contact the vendor for patch status; if available, plan and execute an upgrade to the patched version; if unavailable, plan decommissioning of this tool and migration to a maintained alternative.

next-ai-draw-io CVE-2026-50758

HIGH

Cross-site Scripting (XSS) (CWE-79)

2026-07-21 mitre

XSS RCE N A

8.1

CVSS 3.1 · Vendor: mitre

Severity by source

Vendor (mitre) PRIMARY

8.1 HIGH

AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:N

vuln.today AI

6.1 MEDIUM

Reflected XSS is remotely reachable without auth but needs victim interaction (UI:R); browser execution crosses into the user's context (S:C) with limited confidentiality/integrity impact (C:L/I:L), typical for XSS rather than the input's C:H/I:H.

3.1 AV:N/AC:L/PR:N/UI:R/S:C/C:L/I:L/A:N

4.0 AV:N/AC:L/AT:N/PR:N/UI:P/VC:L/VI:L/VA:N/SC:L/SI:L/SA:N

Primary rating from Vendor (mitre).

CVSS VectorVendor: mitre

Attack Vector

Network

Attack Complexity

Low

Privileges Required

None

User Interaction

Required

Scope

Unchanged

Confidentiality

High

Integrity

High

Availability

None

Lifecycle Timeline

Analysis Generated

Jul 22, 2026 - 19:35 vuln.today

CVSS changed

Jul 22, 2026 - 18:22 NVD

8.1 (HIGH)

CVE Published

Jul 21, 2026 - 00:00 cve.org

HIGH 8.1

CVE Published

Jul 21, 2026 - 00:00 cve.org

UNKNOWN (no severity yet)

DescriptionCVE.org

Cross Site Scripting vulnerability in DayuanJiang next-ai-draw-io 0.4.13 allows a remote attacker to execute arbitrary code via the mcp parameter

AnalysisAI

Cross-site scripting in DayuanJiang next-ai-draw-io 0.4.13 lets a remote unauthenticated attacker inject arbitrary script through the mcp parameter, executing in the victim's browser when they open a crafted link. The 'arbitrary code' claim and RCE tag in the source data most plausibly refer to arbitrary JavaScript in the browser (CWE-79 reflected XSS) rather than server-side code execution. EPSS is low (0.29%, 21st percentile), there is no CISA KEV listing, and no public exploit is confirmed, though third-party GitHub writeups referencing this CVE exist.

Technical ContextAI

next-ai-draw-io is a small open-source Next.js web application pairing draw.io-style diagramming with AI/LLM assistance, where 'mcp' likely denotes a Model Context Protocol integration parameter. The flaw is CWE-79 (Improper Neutralization of Input During Web Page Generation): the mcp request parameter is reflected into rendered HTML/JS without adequate contextual output encoding, so attacker-controlled markup is interpreted by the browser. The supplied CPE (cpe:2.3:a:n/a:n/a) is a placeholder providing no usable product/version identification, so scoping relies solely on the description value 0.4.13.

RemediationAI

No vendor-released patched version is identified at time of analysis, so treat fix status as unconfirmed and monitor the upstream issue (github.com/DayuanJiang/next-ai-draw-io/issues/755) for a tagged release beyond 0.4.13. As compensating controls, apply strict server-side contextual output encoding of the mcp parameter before rendering and deploy a restrictive Content-Security-Policy disallowing inline script to blunt reflected execution - noting CSP may break legitimate inline handlers and require app changes. Additionally allowlist/validate accepted mcp values, and if the MCP feature is unused, disable or restrict that endpoint at a reverse proxy; this reduces exposure but removes MCP functionality for dependent users.