detect_exchange_web_shell.yml

name: Detect Exchange Web Shell
id: 8c14eeee-2af1-4a4b-bda8-228da0f4862a
version: 4
date: '2023-07-10'
author: Michael Haag, Shannon Davis, David Dorsey, Splunk
status: production
type: TTP
description: 'The following query identifies suspicious .aspx created in 3 paths identified
  by Microsoft as known drop locations for Exchange exploitation related to HAFNIUM
  group and recently disclosed vulnerablity named ProxyShell and ProxyNotShell. Paths
  include: `\HttpProxy\owa\auth\`, `\inetpub\wwwroot\aspnet_client\`, and `\HttpProxy\OAB\`.
  Upon triage, the suspicious .aspx file will likely look obvious on the surface.
  inspect the contents for script code inside. Identify additional log sources, IIS
  included, to review source and other potential exploitation. It is often the case
  that a particular threat is only applicable to a specific subset of systems in your
  environment. Typically analytics to detect those threats are written without the
  benefit of being able to only target those systems as well. Writing analytics against
  all systems when those behaviors are limited to identifiable subsets of those systems
  is suboptimal. Consider the case ProxyShell vulnerability on Microsoft Exchange
  Servers. With asset information, a hunter can limit their analytics to systems that
  have been identified as Exchange servers. A hunter may start with the theory that
  the exchange server is communicating with new systems that it has not previously.
  If this theory is run against all publicly facing systems, the amount of noise it
  will generate will likely render this theory untenable. However, using the asset
  information to limit this analytic to just the Exchange servers will reduce the
  noise allowing the hunter to focus only on the systems where this behavioral change
  is relevant.'
data_source:
- Sysmon Event ID 1
search: '| tstats `security_content_summariesonly` count FROM datamodel=Endpoint.Processes
  where Processes.process_name=System  by _time span=1h Processes.process_id Processes.process_name
  Processes.dest | `drop_dm_object_name(Processes)` | join process_guid, _time [|
  tstats `security_content_summariesonly` count min(_time) as firstTime max(_time)
  as lastTime FROM datamodel=Endpoint.Filesystem where Filesystem.file_path IN ("*\\HttpProxy\\owa\\auth\\*",
  "*\\inetpub\\wwwroot\\aspnet_client\\*", "*\\HttpProxy\\OAB\\*") Filesystem.file_name
  IN( "*.aspx", "*.ashx") by _time span=1h Filesystem.user Filesystem.dest Filesystem.file_create_time
  Filesystem.file_name Filesystem.file_path | `drop_dm_object_name(Filesystem)` |
  fields _time dest file_create_time file_name file_path process_name process_path
  process] | dedup file_create_time | table dest file_create_time, file_name, file_path,
  process_name | `detect_exchange_web_shell_filter`'
how_to_implement: To successfully implement this search you need to be ingesting information
  on process that include the name of the process responsible for the changes from
  your endpoints into the `Endpoint` datamodel in the `Processes` node and `Filesystem`
  node.
known_false_positives: The query is structured in a way that `action` (read, create)
  is not defined. Review the results of this query, filter, and tune as necessary.
  It may be necessary to generate this query specific to your endpoint product.
references:
- https://raw.githubusercontent.com/Azure/Azure-Sentinel/master/Sample%20Data/Feeds/MSTICIoCs-ExchangeServerVulnerabilitiesDisclosedMarch2021.csv
- https://www.zerodayinitiative.com/blog/2021/8/17/from-pwn2own-2021-a-new-attack-surface-on-microsoft-exchange-proxyshell
- https://www.youtube.com/watch?v=FC6iHw258RI
- https://www.huntress.com/blog/rapid-response-microsoft-exchange-servers-still-vulnerable-to-proxyshell-exploit#what-should-you-do
tags:
  analytic_story:
  - ProxyNotShell
  - ProxyShell
  - CISA AA22-257A
  - HAFNIUM Group
  - BlackByte Ransomware
  asset_type: Endpoint
  confidence: 90
  impact: 90
  message: A file - $file_name$ was written to disk that is related to IIS exploitation
    previously performed by HAFNIUM. Review further file modifications on endpoint
    $dest$ by user $user$.
  mitre_attack_id:
  - T1505
  - T1505.003
  - T1190
  - T1133
  observable:
  - name: user
    type: User
    role:
    - Victim
  - name: dest
    type: Hostname
    role:
    - Victim
  - name: file_name
    type: File Name
    role:
    - Victim
  product:
  - Splunk Enterprise
  - Splunk Enterprise Security
  - Splunk Cloud
  required_fields:
  - _time
  - Filesystem.file_path
  - Filesystem.process_id
  - Filesystem.file_name
  - Filesystem.file_hash
  - Filesystem.user
  risk_score: 81
  security_domain: endpoint
tests:
- name: True Positive Test
  attack_data:
  - data: https://media.githubusercontent.com/media/splunk/attack_data/master/datasets/attack_techniques/T1505.003/windows-sysmon_proxylogon.log
    source: XmlWinEventLog:Microsoft-Windows-Sysmon/Operational
    sourcetype: xmlwineventlog