Explicitly fetch http payloads for analysis
Many of the exploit attempts cause ES to attempt to wget a URL. This URL should be fetched, md5'ed, and ideally included in the payload for analysis. Maybe this should be configurable.
I've thought about this exact thing. Kippo does this quite well. My only concern would be how to parse the wget command sent to the honeypot.
This is how we do it for shockpot: https://github.com/threatstream/shockpot/blob/master/commands.py. We just grabbed a bunch of example exploit attempts and made sure that we covered most of them. The bad guys will be able to evade but I haven't seen many examples of this in the wild.
Opened a PR to address this enhancement https://github.com/jordan-wright/elastichoney/pull/6
Thanks for this everyone! Let me review the PR and I'll see what I can do to merge. My biggest holdup is reviewing exactly what's going on with the parsing to make sure there isn't a way to exploit the honeypot.
Give me a week or so and bug me if I haven't merged by then :smile:
Thanks again!
Thanks @jordan-wright - I have one more commit coming shortly to address handling form-encoded vs json encoded post requests.