Impact
This is low impact and limited XSS, because code for XSS payload is always visible, but attacker can use other techniques to hide the code the victim sees.
Also if the application use execHash option and execute code from URL the attacker can use this URL to execute his code. The scope is limited because the javascript code inside html attribute used is added to span tag, so no automatic execution like with onerror
on images is possible.
Patches
Fixed version 2.31.1
Workarounds
The user can use formatting that wrap whole user input and it's no op.
$.terminal.new_formatter([/([\s\S]+)/g, '[[;;]$1]']);
The fix will only work when user of the library is not using different formatters (e.g. to highlight code in different way).
References
The issue was reported here jcubic/jquery.terminal#727
For more information
If you have any questions or comments about this advisory:
Impact
This is low impact and limited XSS, because code for XSS payload is always visible, but attacker can use other techniques to hide the code the victim sees.
Also if the application use execHash option and execute code from URL the attacker can use this URL to execute his code. The scope is limited because the javascript code inside html attribute used is added to span tag, so no automatic execution like with
onerror
on images is possible.Patches
Fixed version 2.31.1
Workarounds
The user can use formatting that wrap whole user input and it's no op.
The fix will only work when user of the library is not using different formatters (e.g. to highlight code in different way).
References
The issue was reported here jcubic/jquery.terminal#727
For more information
If you have any questions or comments about this advisory: