CVE intelligence and bounded remediation

CVE-2025-57203 — Liquidlabs Magicai security vulnerability

Medium CVSS 4.8

MagicProject AI version 9.1 is affected by a Cross-Site Scripting (XSS) vulnerability within the chatbot generation feature available to authenticated admin users. The vulnerability resides in the prompt parameter submitted to the /dashboard/user/generator/generate-stream endpoint via a multipart/form-data POST request. Due to insufficient input sanitization, attackers can inject HTML-based JavaScript payloads. This payload is stored and rendered unsanitized in subsequent views, leading to execution in other users' browsers when they access affected content. This issue allows an authenticated attacker to execute arbitrary JavaScript in the context of another user, potentially leading to session hijacking, privilege escalation, data exfiltration, or administrative account takeover. The application does not implement a Content Security Policy (CSP) or adequate input filtering to prevent such attacks. A fix should include proper sanitization, output encoding, and strong CSP enforcement to mitigate exploitation.

Severity
Medium
CVSS
4.8 (3.1)
Published
2025-09-22
CISA KEV
Not currently listed
Ecosystem
javascript/npm
Weaknesses
CWE-79

Affected products

  • liquidlabs / magicai / 9.1

Matched remediation archetype

Cross-site scripting and unsafe browser output

This catalog composition supplies bounded fallback guidance. Explicitly reviewed curated workflows load with the complete record below.

Check exposure

  • Trace reflected, stored, and DOM-derived untrusted values into HTML, attributes, URLs, styles, scripts, and client-side template sinks.
  • Identify affected origins, authenticated user roles, sensitive browser capabilities, and where content is shared across tenants.
  • Review framework escaping, rich-text sanitization, legacy templates, and client-side rendering paths.

Remediate safely

  • Use context-aware framework output encoding and safe DOM APIs; keep untrusted data out of executable contexts.
  • Sanitize intentionally supported markup with a maintained allowlist policy and validate URLs and attributes separately.
  • Update affected rendering components and add tests for every output context using inert sentinel markup.

Authoritative sources

Complete CVE record and remediation plan

The detailed catalog view below loads this exact record, its source evidence, and the full seven-phase agentic change plan.