Attack Chain

1. The attacker crafts malicious HTML content containing a <pre> tag with embedded backticks and HTML-like text.
2. The attacker supplies this crafted HTML to the JustHTML() constructor with sanitize=True.
3. The to_markdown() method is called on the JustHTML object.
4. The to_markdown() method processes the <pre> tag and generates a fenced code block with a fixed delimiter of ``````` (three backticks).
5. Due to the fixed delimiter, the attacker-controlled content, containing a backtick run of length 3 or more, terminates the code block prematurely.
6. The remaining HTML-like text, such as <img src=x onerror=alert(1)>, appears outside the fenced code block in the resulting Markdown output.
7. The vulnerable markdown is rendered by a CommonMark/GFM-style renderer.
8. The injected HTML is treated as raw HTML, leading to the execution of the embedded JavaScript code.

Impact

Successful exploitation of this vulnerability allows attackers to inject arbitrary HTML and JavaScript code into applications that use JustHTML(..., sanitize=True).to_markdown() and render the output in Markdown contexts. This may lead to XSS, where the attacker can execute malicious scripts in the context of the victim's browser. The number of affected applications is currently unknown, but any application using JustHTML for Markdown generation with user-supplied HTML is potentially vulnerable.

Recommendation

• Upgrade to a version of justhtml greater than 1.12.0 to remediate the vulnerability as advised in the advisory.
• Sanitize user-provided HTML input before passing it to JustHTML to mitigate the risk of malicious code injection.
• If upgrading is not immediately feasible, consider implementing custom escaping logic for <pre> content before passing it to to_markdown() to prevent code fence breakouts.