smert.net

There are quite a few applications on Facebook one of the more popular ones is (fluff)Friends. Mike Sego is the creator and I guess he works at Google (must be nice). Fluff Friends has about 340,000 daily active users. So what is the purpose of this game? It is kind of like a virtual pet that you feed, pet and race. You can buy it different foods and backgrounds. The various foods increase the speed of your Fluff. I have a cowe (not a cow but a cowe, it is the same animal, I want to make a baseball glove out of mine) that is saying not very cleverly “I’M RICH BITCH!”. Well because in Fluff terms I’m pretty damn rich (57480 Munny). I wanted to get even more than before mentioning the fact that I had a ton of money but I’ve lost interest mainly to Warbook. It is also very pointless since all I care about is MUNNY. I liek Munny…you like Munny too? No way. We should totally hang out. Ok moving on…

How can you legitimately get that much money? Besides being extremely lucky? For me, I didn’t get rich winning bets. Besides betting on your Fluff you do get 5 Munny for each Fluff Friend you pet (not including yours) every 24 hours. Once you have enough Munny you can experiment more and find out how betting works. Least that is what I started to do. Petting does have some limitations. Taken directly from the FAQ.

Yes, in order to discourage abuse, we have some limits in place. Specifically, you can only
pet a (fluff)Friend once every 5 seconds, you are limited to earning 500 munny per hour
from petting, and any particular (fluff)Friend cannot be pet more than 2500 times in 30 days.

By the time I started playing Jordan already had a very reliable yet not very efficient way of automating petting. The tool of choice again greasemonkey. Just bookmark a list of Fluff Friends and open them all in tabs. You will no doubt go over the limit but you can hit refresh and wait 5 seconds and go on to the next one. We used another method to launch Firefox with the URL of a fluff friend, then kill it off. Here are two scripts one for petting and one for betting. Now where to get a list of Fluff Friends? Well someone actually posted a list of about 1,100 URLs in the forum. But this wasn’t good enough for me I wanted more than just Park Place, I wanted Boardwalk too. So I decided that I would visit the entire forum and collect URLs automatically of course. Now 2102 topics later I had a decent list to start with. Yet, this was only the beginning.

From forum:There are 2102 topics total.
There are 36948 total lines containing fluff URLs.
There are 73320 total extracted fluff URLs.
There are 2542 unique extracted fluff URLs.
There are 361 invalid topics.

But this still wasn’t good enough I WANTED TO BUILD A HOTEL. Unlike the Fluff Forum, Facebook requires you to login and it uses cookies to keep you authenticated. This was a bit of a new area for me and I started by just stealing the cookie file from the Firefox profile. But this didn’t work at first because Facebook had cookies that began with #HttpOnly_ and for some reason Wget didn’t like the taste of that. Once I removed that from the cookies file I began visiting my newly acquired Fluff Friends. I found that most of the people had left links to other Fluff Friends asking them to be pet. The URLs coming from the first few hundred people made my list grow very rapidly. There was another problem however, sometimes while I was using Firefox the cookies changed and logged me out. I decided to ditch Firefox completely and make it separate.

File: make_cookie.sh
COOKIE_FILE="/home/kovertops/.mozilla/firefox/<changethis>.default/cookies.txt"

cat ${COOKIE_FILE} | grep facebook | sed -e 's/#HttpOnly_//' > cookies.txt

File: do_login.sh
#!/bin/bash
USER=emailaddress@domain.com
PASS=password

wget --save-cookies="cookies.txt" --user-agent="Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8.1.6) Gecko/20061201 Firefox/2.0.0.6 (Ubuntu-feisty)" -O login/browse.php http://www.facebook.com > wget/login.get 2>&1

cat login/browse.php | grep -o '<form id="app2219808235_pet_friend".*>.*<input type="hidden" name="type" value="pet" />'

CHALLENGE=`cat login/browse.php | grep -o '<input type="hidden" name="challenge" value=".*' | awk '{ print $4 }' | awk -F "=" '{ print $2 }' | sed -e 's/"//g'`

DATA="challenge=${CHALLENGE}&md5pass=&email=${USER}&noerror=1&pass=${PASS}&persistent=1&doquicklogin=Login&charset_test=€´€´???"

wget --header="Content-Type: application/x-www-form-urlencoded" --keep-session-cookies --load-cookies="cookies.txt" --post-data="${DATA}" --save-cookies="cookies.txt" --user-agent="Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8.1.6) Gecko/20061201 Firefox/2.0.0.6 (Ubuntu-feisty)" -O login/login.php https://login.facebook.com/login.php > wget/login.post 2>&1

That is what I came up with to get me to login. You might have to view the charset_test from Facebook since it won’t save properly. Once that was out of the way my collection began. I downloaded and pet the first 2542 unique Fluff Friends, then parsed them and continued to download new ones. This continued over a few days. At one point my list got too big and broke because Bash can handle only a certain number of arguments. After switching to xargs I was able to generate again. Once it slowed down and I only started to get a few hundred new URLs I cleared my results and began from the beginning. You can get the list of Fluff Friends from here.

From Facebook:
There are 12751 topics total.
There are 12928 total lines containing fluff URLs.
There are 304237 total extracted fluff URLs.
There are 11686 unique extracted fluff URLs.
There are 82 invalid friends.

There are 11686 total URLs from facebook.
There are 2542 total URLs from the forum.
There are 14228 total extracted fluff URLs.
There are 12858 unique extracted fluff URLs.

I think that 12k Fluff Friends should be sufficient for the time being. About a week ago they made changes to the page and added hidden forms. This broke my application as it blindly submitted the fake information. My plan was to rewrite this in PHP and make a better petter (kinda like the poller from Cacti). This would require a MySQL database to store everything. I would increase the error checking and make it very difficult to submit bogus information. It does sound like a nice idea but I have lost interest. Anyways enjoy the real list of URLs. It would have been better if I released this on Halloween (trick and a treat) BTW my fluff folder with all of my output is about a 1GB in size and this was before I cleared out my results (1.4GB before). The list has a small number of invalid Fluff Friends and I think there are still 5 duplicates. I don’t know why uniq didn’t remove them all. Anyways to the 12k people that I clicked on thanks for the Munny

Update (26JAN08)

I decided to release the rest of my bash scripts used for automation/petting. I don’t plan on making any other changes to the code. That goes for any Greasemonkey scripts too. You can download it here.