Can I change the CAPTCHA word library?

MadelineBarr

How does Confluence Server select the words for CAPTCHA? 

We have had a couple interesting or questionable words appear in the CAPTCHA. Some users were concerned (thought we were hacked) and some were even a little offended. 

Are these words in a library somewhere or is it two parts of words randomly put together that just happen to form a word? Answer: I did a little testing and realize they are not really "words." I'm assuming they are randomly generate letters or more likely parts of words put together. I think in this case it just happened to form a real word.
Still wouldn't mind clarification on this though. 

Is there a way we can remove or prevent some of the words? Do I need to put in a request to Atlassian to consider to have the words removed? Edit: Wondering if there is something like a word blacklist.

Thanks in advance.
Madeline

 

Sattesh M

Hello Madaline, thank you for contacting us about this.

The answer to this is that the CAPTCHAs really are random, where it takes a section of one word, a section of another word and then tosses them together.

We've had issues with this before, where certain "words" were formed, that should not have.
Since then, starting Confluence 6.0.6, we added a "bad-list" filter, to check the words that were formed, and to filter them if they match the words in the "bad-list".

• CONFSERVER-22963 - Confluence CAPTCHA includes offensive words

There is no way to edit this "bad-list" at the moment, but there is en enhancement request for this exact feature.

• CONFSERVER-52217 - Updating and maintaining offensive captcha word list

If you are already on a version of Confluence, over 6.0.6, and the CAPTCHA threw out together an offensive word, do let us know in the CONFSERVER-22963 ticket above.

If you feel that the word or the content is sensitive, just get in touch with the Confluence support team, here, and we'll make sure the issue is raised confidentially. 

Let me know if you have any questions.

Cheers,
Sattesh