Promptcraft Forge Studio 的不完善的 URL 检查容易受到 SVG 注入的 XSS 攻击

The vulnerability stems from inadequate sanitization of URLs within Promptcraft Forge Studio's validation process. Specifically, the application fails to properly filter data: URLs, allowing attackers to embed malicious JavaScript code within image or other data URI schemes. When these URLs are processed and displayed, the embedded script executes in the user's browser, potentially leading to session hijacking, data theft, or defacement of the application. Given the tool's purpose in managing LLM applications, a successful exploit could compromise sensitive prompts, training data, or even the underlying LLM models themselves, significantly expanding the attack surface.

Developers and organizations utilizing Promptcraft Forge Studio for building and managing LLM-powered applications are at significant risk. This includes teams working with sensitive data, deploying applications in production environments, or relying on the tool for critical workflows. Shared hosting environments where multiple users share the same instance of Promptcraft Forge Studio are particularly vulnerable.

• javascript / web: Inspect network traffic for requests containing data: URLs, particularly in href or src attributes.

// Example: Check for data: URLs in a webpage
const links = document.querySelectorAll('a');
links.forEach(link => {
  if (link.href.startsWith('data:')) {
    console.warn('Potential XSS vulnerability: ', link.href);
  }
});

• generic web: Monitor access logs for unusual patterns involving URLs with embedded scripts or data URIs.

 grep 'data:image/svg+xml' access.log

1. 2025-09-04Disclosure

   disclosure

1. 已保留2025-08-29
2. 发布日期2025-09-04
3. EPSS 更新日期2026-03-23

The primary mitigation for CVE-2025-58361 is to immediately upgrade Promptcraft Forge Studio to version 0.0.1, which includes the necessary URL validation fixes. If upgrading is not immediately feasible, consider implementing a Web Application Firewall (WAF) rule to block requests containing data: URLs or other suspicious URL schemes. Additionally, carefully review any user-supplied URLs before incorporating them into the application to ensure they are safe and properly sanitized. Thorough input validation on all user-provided data is crucial to prevent similar vulnerabilities in the future.