Hunting words
Some words about hunting including some perspectives from different sources
- “Random acts of incident response”
- Orgs that don’t have mature/effective IR may quest about on indefensible networks for signs of misbehaviour and compromise
- On networks without basic hygiene in place (CSC 1-5 / ASD quick wins) this will always succeed and never really help
- Using data in a structured manner to disprove assumptions about your network and defences (via FireEye team @ BSides Augusta (Harry Potter themed)
- http://www.irongeek.com/i.php?page=videos/bsidesaugusta2016/living-in-america05-hunting-defense-against-the-dark-arts-jacqueline-stokes-danny-akacki-and-stephen-hinck
- EG: We don’t have unauthorised software installs.
- EG: SSH outbound to Internet isn’t possible from corporate networks
- Proving these false with data can motivate/empower organisational change / prove points to clients
- Like most data science getting, cleaning, and normalising data is significant work
- Assume compromise and try to detect post-compromise/next attack phase activities (sniff, spray, pivot, escalate, backdoor, exfil, etc) (SEC511)
- EG: Pick a network segment and look for outbound file transfers. Are any legitimate activity?
- EG: Looking for beacons and C2 in web proxy or DNS logs
- This can be fruitful, but needs some structure to be practical.
- And of course without functional IR and the defensible network hygiene it’s not of much use.
- Have your more skilled responders manually looking through collected data for potential misbehaviour that can be configured for automated detection
- This is a step towards @bammv’s idea of three detection types: “RealTime, Batch, and Hunting”
- This also aligns with the two playbook report types in Cisco team’s Crocodile Book: High Fidelity and Investigation Lead
- Using collected techniques (public or private) for post-compromise detections / incident discovery:
- EG https://github.com/ThreatHuntingProject/ThreatHunting
- https://detect-respond.blogspot.com/2015/10/a-simple-hunting-maturity-model.html
Written on February 12, 2017