CVE-2026-62324

Jodit Editor is a WYSIWYG editor with a built-in file browser & image editor. Prior to 4.12.31, Jodit's sanitizeHTMLElement method fails to use isDangerousUrl to normalize javascript: href values before checking the scheme, allowing case variants, control-byte prefixes, and embedded tabs or newlines to bypass filtering and execute attacker-controlled script when a victim clicks a stored link rendered by an application. This issue is fixed in version 4.12.31.

Jodit Editor is a WYSIWYG editor with a built-in file browser & image editor. Prior to 4.12.31, Jodit's sanitizeHTMLElement method fails to use isDangerousUrl to normalize javascript: href values before checking the scheme, allowing case variants, control-byte prefixes, and embedded tabs or newlines to bypass filtering and execute attacker-controlled script when a victim clicks a stored link rendered by an application. This issue is fixed in version 4.12.31.

Summary

jodit's sanitizeHTMLElement neutralizes a javascript: href using a bare href.trim().indexOf('javascript') === 0 check. This omits the normalization jodit applies to every other URL attribute: isDangerousUrl strips control bytes with value.replace(/[\u0000-\u0020]+/g, '') and lowercases the value before testing the scheme. Because the href check does neither, it is bypassed by three obfuscation classes, all confirmed firing on click against the shipped 4.12.30 build:

1. Case variants: JAVASCRIPT:, Javascript:, jaVaScRiPt: (the check is case-sensitive).
2. A leading C0 control byte, e.g. a \x01 prefix before lowercase javascript: (trim() does not remove bytes in the \x00-\x08 / \x0e-\x1f range, but the browser strips a leading control byte before resolving the scheme).
3. An embedded tab or newline inside the scheme, e.g. java\tscript: or java\nscript: (the browser strips tab/newline from a URL, but indexOf('javascript') sees the broken word and does not match).

The dangerous href survives editor.value = assignment and the on-change LazyWalker, persisting in the stored editor value. A victim who clicks the link in any consumer that renders the stored value (readonly editor, server-rendered page, innerHTML consumer) runs attacker-controlled JS in that page's origin.

Details

The check is in sanitizeHTMLElement at src/core/helpers/html/safe-html.ts:213:

if (safeJavaScriptLink && href && href.trim().indexOf('javascript') === 0) {
    attr(elm, 'href', location.protocol + '//' + href);
    effected = true;
}

href.trim() removes leading/trailing ASCII whitespace only, and indexOf('javascript') is case-sensitive and literal. So the check fails to fire whenever the scheme is upper/mixed-case, prefixed by a non-whitespace control byte, or split by an embedded tab/newline - all of which a browser still resolves to javascript: on click (URI schemes are case-insensitive per RFC 3986 section 3.1; leading control bytes, tabs and newlines are stripped from a URL during parsing).

The same file already contains the correct routine, isDangerousUrl() (line 176), used for every other URL attribute (src, data, action, formaction, poster, background, xlink:href):

function isDangerousUrl(value, tagName) {
    const normalized = value.replace(/[\u0000-\u0020]+/g, '').toLowerCase();
    if (/^(?:javascript|vbscript|livescript|mocha):/.test(normalized)) {
        return true;
    }
    // ...
}

isDangerousUrl strips every control byte and ASCII space (/[\u0000-\u0020]+/g) and lowercases before testing the scheme, so it resists all three obfuscations. But href never goes through it: the attribute list isDangerousUrl is applied to (URL_ATTRIBUTES) is commented "besides href", and href is handled only by the weaker indexOf check. Both the synchronous value-set path (onBeforeSetNativeEditorValue -> safeHTML -> sanitizeHTMLElement) and the asynchronous on-change path (sanitizeAttributes -> sanitizeHTMLElement) use that same weak check.

Positive controls (filter is otherwise live): a plain lowercase javascript: href IS neutralized: jodit rewrites the value to location.protocol + '//' + href, so it reads about://javascript:... on an about:blank test page and https://javascript:... on an https page. A leading ASCII space or tab IS caught by trim(); the bypass is specific to the un-normalized forms above.

Proof of concept

Default configuration. Assign a payload to the editor and read back the stored value:

const editor = Jodit.make('#editor');
editor.value = '<a href="JAVASCRIPT:alert(document.domain)">click me</a>';
// editor.value getter returns the href unchanged:
//   <p><a href="JAVASCRIPT:alert(document.domain)">click me</a></p>
document.getElementById('view').innerHTML = editor.value;
// Clicking "click me" runs alert(document.domain) in the consumer's origin.

The same persists for the…

Jodit Editor is a WYSIWYG editor with a built-in file browser & image editor. Prior to 4.12.31, Jodit's sanitizeHTMLElement method fails to use isDangerousUrl to normalize javascript: href values before checking the scheme, allowing case variants, control-byte prefixes, and embedded tabs or newlines to bypass filtering and execute attacker-controlled script when a victim clicks a stored link rendered by an application. This issue is fixed in version 4.12.31.

Jodit Editor is a WYSIWYG editor with a built-in file browser & image editor. Prior to 4.12.31, Jodit's sanitizeHTMLElement method fails to use isDangerousUrl to normalize javascript: href values before checking the scheme, allowing case variants, control-byte prefixes, and embedded tabs or newlines to bypass filtering and execute attacker-controlled script when a victim clicks a stored link rendered by an application. This issue is fixed in version 4.12.31.

Summary

jodit's sanitizeHTMLElement neutralizes a javascript: href using a bare href.trim().indexOf('javascript') === 0 check. This omits the normalization jodit applies to every other URL attribute: isDangerousUrl strips control bytes with value.replace(/[\u0000-\u0020]+/g, '') and lowercases the value before testing the scheme. Because the href check does neither, it is bypassed by three obfuscation classes, all confirmed firing on click against the shipped 4.12.30 build:

1. Case variants: JAVASCRIPT:, Javascript:, jaVaScRiPt: (the check is case-sensitive).
2. A leading C0 control byte, e.g. a \x01 prefix before lowercase javascript: (trim() does not remove bytes in the \x00-\x08 / \x0e-\x1f range, but the browser strips a leading control byte before resolving the scheme).
3. An embedded tab or newline inside the scheme, e.g. java\tscript: or java\nscript: (the browser strips tab/newline from a URL, but indexOf('javascript') sees the broken word and does not match).

The dangerous href survives editor.value = assignment and the on-change LazyWalker, persisting in the stored editor value. A victim who clicks the link in any consumer that renders the stored value (readonly editor, server-rendered page, innerHTML consumer) runs attacker-controlled JS in that page's origin.

Details

The check is in sanitizeHTMLElement at src/core/helpers/html/safe-html.ts:213:

if (safeJavaScriptLink && href && href.trim().indexOf('javascript') === 0) {
    attr(elm, 'href', location.protocol + '//' + href);
    effected = true;
}

href.trim() removes leading/trailing ASCII whitespace only, and indexOf('javascript') is case-sensitive and literal. So the check fails to fire whenever the scheme is upper/mixed-case, prefixed by a non-whitespace control byte, or split by an embedded tab/newline - all of which a browser still resolves to javascript: on click (URI schemes are case-insensitive per RFC 3986 section 3.1; leading control bytes, tabs and newlines are stripped from a URL during parsing).

The same file already contains the correct routine, isDangerousUrl() (line 176), used for every other URL attribute (src, data, action, formaction, poster, background, xlink:href):

function isDangerousUrl(value, tagName) {
    const normalized = value.replace(/[\u0000-\u0020]+/g, '').toLowerCase();
    if (/^(?:javascript|vbscript|livescript|mocha):/.test(normalized)) {
        return true;
    }
    // ...
}

isDangerousUrl strips every control byte and ASCII space (/[\u0000-\u0020]+/g) and lowercases before testing the scheme, so it resists all three obfuscations. But href never goes through it: the attribute list isDangerousUrl is applied to (URL_ATTRIBUTES) is commented "besides href", and href is handled only by the weaker indexOf check. Both the synchronous value-set path (onBeforeSetNativeEditorValue -> safeHTML -> sanitizeHTMLElement) and the asynchronous on-change path (sanitizeAttributes -> sanitizeHTMLElement) use that same weak check.

Positive controls (filter is otherwise live): a plain lowercase javascript: href IS neutralized: jodit rewrites the value to location.protocol + '//' + href, so it reads about://javascript:... on an about:blank test page and https://javascript:... on an https page. A leading ASCII space or tab IS caught by trim(); the bypass is specific to the un-normalized forms above.

Proof of concept

Default configuration. Assign a payload to the editor and read back the stored value:

const editor = Jodit.make('#editor');
editor.value = '<a href="JAVASCRIPT:alert(document.domain)">click me</a>';
// editor.value getter returns the href unchanged:
//   <p><a href="JAVASCRIPT:alert(document.domain)">click me</a></p>
document.getElementById('view').innerHTML = editor.value;
// Clicking "click me" runs alert(document.domain) in the consumer's origin.

The same persists for the…