Impact

Versions from 1.2.0 to 1.3.1 of Astro-Shield allow to bypass the allow-lists for cross-origin resources by introducing valid integrity attributes to the injected code. This implies that the injected SRI hash would be added to the generated CSP header, which would lead the browser to believe that the injected resource is legit.

To exploit this vulnerability, the attacker needs to first inject code into the rendered pages by exploiting other not-related potential vulnerabilities.

Patches

Version 1.3.2 provides a patch.

Workarounds - To not use the middleware functionality of Astro-Shield. - To use the middleware functionality of Astro-Shield ONLY for content that cannot be controlled in any way by external users.

References Are there any links users can visit to find out more?

Impact

When the following conditions are met: - Automated CSP headers generation for SSR content is enabled - The web application serves content that can be partially controlled by external users

Then it is possible that the CSP headers generation feature might be "allow-listing" malicious injected resources like inlined JS, or references to external malicious scripts.

Patches Available in version 1.3.0 .

Workarounds - Do not enable CSP headers generation. - Use it only for dynamically generated content that cannot be controlled by external users in any way.

References Are there any links users can visit to find out more?