Security Incident Survey Cheat Sheet for Server Administrators
This cheat sheet helps server administrators examine suspect systems to decide whether to escalate for incident response. Avoid actions that access many files; look at logs, network connections, users, processes, scheduled jobs, and auto-start programs. Includes specific commands for Windows and Unix systems.
This cheat sheet captures tips for examining a suspect server to decide whether to escalate for formal incident response. To print, use the one-sheet PDF version; you can also edit the Word version for you own needs. The steps presented in this cheat sheet aim at minimizing the adverse effect that the initial survey will have on the system, to decrease the likelihood that the attacker’s footprints will be inadvertently erased. If you are an incident handler looking to take on the management of a qualified incident, see the related incident questionnaire cheat sheet.
Assessing the Suspicious Situation
- To retain attacker’s footprints, avoid taking actions that access many files or installing tools.
- Look at system, security, and application logs for unusual events.
- Look at network configuration details and connections; note anomalous settings, sessions or ports.
- Look at the list of users for accounts that do not belong or should have been disabled.
- Look at a listing of running processes or scheduled jobs for those that do not belong there.
- Look for unusual programs configured to run automatically at system’s start time.
- Check ARP and DNS settings; look at contents of the hosts file for entries that do not belong there.
- Look for unusual files and verify integrity of OS and application files.
- Use a network sniffer, if present on the system or available externally, to observe for unusual activity.
- A rootkit might conceal the compromise from tools; trust your instincts if the system just doesn’t feel right.
- Examine recently reported problems, intrusion detection and related alerts for the system.
If You Believe a Compromise is Likely…
- Involve an incident response specialist for next steps, and notify your manager.
- Do not panic or let others rush you; concentrate to avoid making careless mistakes.
- If stopping an on-going attack, unplug the system from the network; do not reboot or power down.
- Take thorough notes to track what you observed, when, and under what circumstances.
Windows Initial System Examination
Look at event logs
eventvwr
Examine network configuration
arp -a, netstat -nr
List network connections and related details
netstat -nao, netstat -vb, net session, net use
List users and groups
lusrmgr, net users, net localgroup administrators, net group administrators
Look at scheduled jobs
schtasks
Look at auto-start programs
msconfig
List processes
taskmgr, wmic process list full
List services
net start, tasklist /svc
Check DNS settings and the hosts file
ipconfig /all, more %SystemRoot%System32Driversetchosts, ipconfig /displaydns
Verify integrity of OS files (affects lots of files!)
sigverif
Research recently-modified files (affects lots of files!)
dir /a/o-d/p %SystemRoot%System32
Avoid using Windows Explorer, as it modifies useful file system details; use command-line.
Unix Initial System Examination
| Task | Command |
|---|---|
| Look at event log files in directories (locations vary) | /var/log/, /var/adm/, /var/spool/ |
| List recent security events | wtmp, who, last, lastlog |
| Examine network configuration | arp -an, route print |
| List network connections and related details | netstat -nap (Linux), netstat -na (Solaris), lsof -i |
| List users | more /etc/passwd |
| Look at scheduled jobs | more /etc/crontab, ls /etc/cron.*, ls /var/at/jobs |
| Check DNS settings and the hosts file | more /etc/resolv.conf, more /etc/hosts |
| Verify integrity of installed packages (affects lots of files!) | rpm -Va (Linux), pkgchk (Solaris) |
| Look at auto-start services | chkconfig —list (Linux), ls /etc/rc*.d (Solaris), smf (Solaris 10+) |
| List processes | ps aux (Linux, BSD), ps -ef (Solaris), lsof +L1 |
| Find recently-modified files (affects lots of files!) | ls -lat /, find / -mtime -2d -ls |
Incident Response Communications
- Do not share incident details with people outside the team responding to the incident.
- Avoid sending sensitive data over email or instant messenger without encryption.
- If you suspect the network was compromised, communicate out-of-band, e.g. non-VoIP phones.
Key Incident Response Steps
- Preparation: Gather and learn the necessary tools, become familiar with your environment.
- Identification: Detect the incident, determine its scope, and involve the appropriate parties.
- Containment: Contain the incident to minimize its effect on neighboring IT resources.
- Eradication: Eliminate compromise artifacts, if necessary, on the path to recovery.
- Recovery: Restore the system to normal operations, possibly via reinstall or backup.
- Wrap-up: Document the incident’s details, retail collected data, and discuss lessons learned.
Other Incident Response Resources
- Windows Intrusion Discovery Cheat Sheet
- Checking Windows for Signs of Compromise
- Linux Intrusion Discovery Cheat Sheet
- Checking Unix/Linux for Signs of Compromise
Post-Scriptum
Special thanks for feedback to Lorna Hutcheson, Patrick Nolan, Raul Siles, Ed Skoudis, Donald Smith, Koon Yaw Tan, Gerard White, and Bojan Zdrnja. If you have suggestions for improving this cheat sheet, please let me know. Creative Commons v3 “Attribution” License for this cheat sheet v. 1.8. Take a look at my other security cheat sheets.