Python script created to parse Windows Prefetch files: Supports XP - Windows 10 Prefetch files
The Windows application prefetch mechanism was put in place to offer performance benefits when launching applications. It just so happens to be one of the more beneficial forensic artifacts regarding evidence of applicaiton execution as well. prefetch.py provides functionality for parsing prefetch files for all current prefetch file versions: 17, 23, 26, and 30.
- Specify a single prefetch file or a directory of prefetch files
- CSV output support
- (Limited) Windows 10 support - Windows 10 prefetch files must be parsed from a Windows 8+ workstation
- Sort a directory of Prefetch files by all execution timestamps
- Output a directory full of .pf files into SQLite tables for aggregate analysis
For now, prefetch.py requires one of two command-line options: --file
specifies a single prefetch to point the script at. --directory
specifies an entire directory of prefetch files which will be parsed and printed to stdout. When using --directory / -d
, remember to include the trailing slash:
dev@computer:~$ ./prefetch.py -h
usage: prefetch.py [-h] [-c] [-d DIRECTORY] [-e EXECUTED] [-f FILE]
optional arguments:
-h, --help show this help message and exit
-c, --csv Present results in CSV format
-d DIRECTORY, --directory DIRECTORY
Parse all PF files in a given directory
-e EXECUTED, --executed EXECUTED
Sort PF files by ALL execution times
-f FILE, --file FILE Parse a given Prefetch file
-s SQLITE, --sqlite SQLITE
Output data to SQLite database file
--file
Using the --file / -f
switch with a single prefetch file results in the output below:
dev@computer:~$ python prefetch.py -f CMD.EXE-4A81B364.pf
=====================
CMD.EXE-4A81B364.pf
=====================
Executable Name: CMD.EXE
Run count: 2
Last Executed: 2016-01-16 20:26:42.515108
Volume Information:
Volume Name: \DEVICE\HARDDISKVOLUME2
Creation Date: 2016-01-16 21:15:18.109374
Serial Number: 88008c2f
Directory Strings:
\DEVICE\HARDDISKVOLUME2\WINDOWS
\DEVICE\HARDDISKVOLUME2\WINDOWS\BRANDING
\DEVICE\HARDDISKVOLUME2\WINDOWS\BRANDING\BASEBRD
\DEVICE\HARDDISKVOLUME2\WINDOWS\GLOBALIZATION
\DEVICE\HARDDISKVOLUME2\WINDOWS\GLOBALIZATION\SORTING
\DEVICE\HARDDISKVOLUME2\WINDOWS\SYSTEM32
Resources loaded:
1: \DEVICE\HARDDISKVOLUME2\WINDOWS\SYSTEM32\NTDLL.DLL
2: \DEVICE\HARDDISKVOLUME2\WINDOWS\SYSTEM32\KERNEL32.DLL
3: \DEVICE\HARDDISKVOLUME2\WINDOWS\SYSTEM32\APISETSCHEMA.DLL
4: \DEVICE\HARDDISKVOLUME2\WINDOWS\SYSTEM32\KERNELBASE.DLL
5: \DEVICE\HARDDISKVOLUME2\WINDOWS\SYSTEM32\LOCALE.NLS
6: \DEVICE\HARDDISKVOLUME2\WINDOWS\SYSTEM32\CMD.EXE
7: \DEVICE\HARDDISKVOLUME2\WINDOWS\SYSTEM32\MSVCRT.DLL
8: \DEVICE\HARDDISKVOLUME2\WINDOWS\SYSTEM32\WINBRAND.DLL
9: \DEVICE\HARDDISKVOLUME2\WINDOWS\SYSTEM32\USER32.DLL
10: \DEVICE\HARDDISKVOLUME2\WINDOWS\SYSTEM32\GDI32.DLL
11: \DEVICE\HARDDISKVOLUME2\WINDOWS\SYSTEM32\LPK.DLL
12: \DEVICE\HARDDISKVOLUME2\WINDOWS\SYSTEM32\USP10.DLL
13: \DEVICE\HARDDISKVOLUME2\WINDOWS\SYSTEM32\IMM32.DLL
14: \DEVICE\HARDDISKVOLUME2\WINDOWS\SYSTEM32\MSCTF.DLL
15: \DEVICE\HARDDISKVOLUME2\WINDOWS\BRANDING\BASEBRD\BASEBRD.DLL
16: \DEVICE\HARDDISKVOLUME2\WINDOWS\GLOBALIZATION\SORTING\SORTDEFAULT.NLS
--directory
By invoking the --directory / -d
flag, the Analyst is able to parse an entire directory of Prefetch files at once.
--sqlite
Use the --sqlite / -s
option with the --directory / -d
option to output the prefetch file header, directory list, and file list into SQLite tables.
--executed
Sort a directory of Prefetch files by execution time. This sort will include ALL timestamps in Windows 8+ Prefetch files (up to eight per file):
dev@computer:~$ python prefetch.py -e Prefetch/
Execution Time, File Executed
2015-11-15 00:02:39.781250, WUAUCLT.EXE-399A8E72
2015-11-15 00:02:26.281250, VERCLSID.EXE-3667BD89
2015-11-15 00:02:24.343750, WMIPRVSE.EXE-28F301A9
2015-11-15 00:02:07.453124, RUNDLL32.EXE-451FC2C0
2015-11-15 00:01:50.765626, GOOGLEUPDATE.EXE-1E123D86
2015-11-15 00:01:08, NTOSBOOT-B00DFAAD
...
...
...
--csv
Using the --csv / -c
flag will provide results in CSV format:
Last Executed, Executable Name, Run Count
2016-01-20 16:01:27.680128, ADOBEIPCBROKER.EXE-c8d02fab, 1
2016-01-20 16:59:42.077480, CREATIVE CLOUD UNINSTALLER.EX-216b8ea8, 1
2016-01-19 18:07:18.101626, MSIEXEC.EXE-a2d55cb6, 37237
2016-01-20 16:11:15.818394, ACRODIST.EXE-782bc2b2, 1
Testing on the prefetch file types below has been completed successfully:
- Windows XP (version 17)
- Windows 7 (version 23)
- Windows 8.1 (version 26)
- Windows 10 (version 30)
This project would not have been possible without the work of others much smarter than I. The prefetch file format is not officially documented by Microsoft and has been understood through reverse engineering, and trial-and-error.
Additionally, Without the excellent work by Francesco Picasso in understanding the Windows 10 prefetch compression method, I would not have been able to get Windows 10 parsed here. I use a modified version of his decompression script in prefetch.py. Francesco's original script can be found at the link below:
To gain a better understanding of the prefetch file format, check out the following resources; which were all used as references for the creation of my script:
ForensicsWiki: Windows Prefetch File Format
Zena Forensics: A first look at Windows 10 Prefetch files
- from argparse import ArgumentParser
- import binascii
- import collections
- import ctypes
- from datetime import datetime,timedelta
- import json
- import os
- import struct
- import sys
- import tempfile