add TLS signatures to the sybil scoring algorithm (useful for sybil resistence for new users in china)

```
wocki hey zach was it you that told me we could find sybil attacks with TLS signatures?
8:06 AM
Zach Wolff yep, I was talking about TLS fingerprinting as a means of identifying website visitors. It could def help with sybil attacks but its no sliver bullet
8:07 AM
it's an emerging practice for web application firewalls to collect metadata on site visitors (as a means of identifying botnets and even just users creating duplicate accounts)
8:55 AM
Zach Wolff I can get deep on the various techniques and methods used under the umbrella of 'web2' style metadata (JS fingerprinting, etc) for identifying site visitors. It gets a bit tricky with an fully open source platform I guess but there is probably a middle ground as you wouldn't need to reveal every Web application firewall rule being used. Anyway, as I'm sure you understand, it can be super valuable info (i'm sure we'd be able to connect some fraudulent grants with their more public personas using these techniques) but ultimately it's a game of cat and mouse in the long run.
9:25 AM
Owocki is there an OSS package/python package where i can play with some of this stuff on the existing app? seems really promising, i wonder if theres an easy win here. (sybil resistence is super important, and with all these new users from china, we are kind of flying blind rn)
New
9:29 AM
Zach Wolff so the TLS FP stuff is hard to do at the web app level itself because it takes into account information related to the TLS handshake itself. Leaving that one aside for now, this project is little old but might be a decent starting point - https://pypi.org/project/shadowd/
PyPIPyPI
shadowd
Python connector for the Shadow Daemon web application firewall (13 kB)
https://pypi.org/static/images/twitter.90915068.jpg
9:30 AM
tbh, we are probably already getting enough info to get started on this as is though. Is anyone looking at doing any kind of correlations agains the web log data right now?
9:30 AM
specially, i'd be interested in a couple key points like, when a user account is first 'created', when a grant is created, and when a grant is funded
9:31 AM
simply looking at the existing data for those things (depending on what is actually being collected and logged) would probably be a good start without any new code needed
9:32 AM
Owocki
Is anyone looking at doing any kind of correlations agains the web log data right now?
nope its all in the webapp right now
9:32 AM
damn. ok well it probably wouldnt be fair to ask you to work on this given its not really QL related.
9:32 AM
but will add it to the backlog
```