Base64 Decoding and Unpacking

This module allows you to extract base64 encoded content from a string or columns of a Pandas DataFrame. The library returns the following information:

  • decoded string (if decodable to utf-8 or utf-16)

  • hashes of the decoded segment (MD5, SHA1, SHA256)

  • string of printable byte values (e.g. for submission to a disassembler)

  • the detected decoded file type (limited)

If the results of the decoding contain further encoded strings these will be decoded recursively. If the encoded string appears to be a zip, gzip or tar archive, the contents will be decompressed after decoding. In the case of zip and tar, the contents of the archive will also be checked for base64 encoded content and decoded/decompressed if possible.

See b64decode

# Imports
import sys
MIN_REQ_PYTHON = (3,6)
if sys.version_info < MIN_REQ_PYTHON:
      print('Check the Kernel->Change Kernel menu and ensure that Python 3.6')
      print('or later is selected as the active kernel.')
      sys.exit("Python %s.%s or later is required.\n" % MIN_REQ_PYTHON)


from IPython.display import display
import pandas as pd

# Import Base64 module
from msticpy.nbtools import *
from msticpy.sectools import *
# Load test data
process_tree = pd.read_csv('data/process_tree.csv',
                           parse_dates=["TimeGenerated"],
                           infer_datetime_format=True)
process_tree[['CommandLine']].head()
                                         CommandLine
0                   .\ftp  -s:C:\RECYCLER\xxppyy.exe
1  .\reg  not /domain:everything that /sid:shines...
2                 cmd  /c "systeminfo && systeminfo"
3                           .\rundll32  /C 12345.exe
4       .\rundll32  /C c:\users\MSTICAdmin\12345.exe

Base64 decode an input string

unpack(input_string=cmdline)

See unpack

Items that decode to utf-8 or utf-16 strings will be returned as decoded strings replaced in the original string. If the encoded string is a known binary type it will identify the file type and return the hashes of the file. If any binary types are known archives (zip, tar, gzip) it will unpack the contents of the archive. For any binary it will return the decoded file as a byte array, and as a printable list of byte values.

The function a tuple of the decoded string and a pandas DataFrame of metadata (hashes, list of byte values, etc.)

# get a commandline from our data set
cmdline = process_tree['CommandLine'].loc[39]
cmdline
'.\\powershell  -enc JAB0ACAAPQAgACcAZABpAHIAJwA7AA0ACgAmACAAKAAnAEkAbgB2AG8AawBlACcAKwAnAC0ARQB4AHAAcgBlAHMAcwBpAG8AbgAnACkAIAAkAHQA'
# Decode the string
base64_dec_str = base64.unpack(input_string=cmdline)

# Print decoded string
print(base64_dec_str)
(".\\powershell  -enc <decoded type='string' name='[None]' index='1' depth='1'>$\x00t\x00 \x00=\x00 \x00'\x00d\x00i\x00r\x00'\x00;\x00\r\x00\n\x00&\x00 \x00(\x00'\x00I\x00n\x00v\x00o\x00k\x00e\x00'\x00+\x00'\x00-\x00E\x00x\x00p\x00r\x00e\x00s\x00s\x00i\x00o\x00n\x00'\x00)\x00 \x00$\x00t\x00</decoded>",    reference                                    original_string file_name  \
0  (, 1., 1)  JAB0ACAAPQAgACcAZABpAHIAJwA7AA0ACgAmACAAKAAnAE...   unknown

  file_type                                        input_bytes  \
0      None  b"$\x00t\x00 \x00=\x00 \x00'\x00d\x00i\x00r\x0...

                                      decoded_string encoding_type  \
0  $t='dir';\r\n&('Invok...         utf-8

                                         file_hashes  \
0  {'md5': '6cd1486db221e532cc2011c9beeb4ffc', 's...

                                md5                                      sha1  \
0  6cd1486db221e532cc2011c9beeb4ffc  6e485467d7e06502046b7c84a8ef067cfe1512ad

                                              sha256  \
0  d3291dab1ae552b91e6b50d7460ceaa39f6f92b2cda433...

                                     printable_bytes
0  24 00 74 00 20 00 3d 00 20 00 27 00 64 00 69 0...  )

Using a DataFrame as input

You can use unpack_df to pass a DataFrame as an argument. Use the column parameter to specify which column to process.

In the case of DataFrame input, the output DataFrame contains these additional columns:

  • src_index - the index of the row in the input dataframe from which the data came.

  • full_decoded_string - the full decoded string with any decoded replacements. This is only really useful for top-level items, since nested items will only show the ‘full’ string representing the child fragment.

Base64 decode strings taken from a pandas dataframe.

Items that decode to utf-8 or utf-16 strings will be returned as decoded strings replaced in the original string. If the encoded string is a known binary type it will identify the file type and return the hashes of the file. If any binary types are known archives (zip, tar, gzip) it will unpack the contents of the archive. For any binary it will return the decoded file as a byte array, and as a printable list of byte values.

# specify the data and column parameters
dec_df = base64.unpack_df(data=process_tree, column='CommandLine')
dec_df
   reference                                    original_string file_name  \
0  (, 1., 1)  JAB0ACAAPQAgACcAZABpAHIAJwA7AA0ACgAmACAAKAAnAE...   unknown
1  (, 1., 1)                   aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa   unknown
2  (, 1., 1)                   aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa   unknown
3  (, 1., 1)                   81ed03caf6901e444c72ac67d192fb9c   unknown

  file_type                                        input_bytes  \
0      None  b"$\x00t\x00 \x00=\x00 \x00'\x00d\x00i\x00r\x0...
1      None  b'i\xa6\x9ai\xa6\x9ai\xa6\x9ai\xa6\x9ai\xa6\x9...
2      None  b'i\xa6\x9ai\xa6\x9ai\xa6\x9ai\xa6\x9ai\xa6\x9...
3      None  b'\xf3W\x9d\xd3w\x1a\x7f\xaft\xd5\xee8\xe1\xce...

                                      decoded_string encoding_type  \
0  $t='dir';\r\n&('Invok...         utf-8
1                                       ꙩ榚骦ꙩ榚骦ꙩ榚骦ꙩ榚骦        utf-16
2                                       ꙩ榚骦ꙩ榚骦ꙩ榚骦ꙩ榚骦        utf-16
3                                       埳펝᩷꽿해㣮컡槶믎彷絶岿        utf-16

                                         file_hashes  \
0  {'md5': '6cd1486db221e532cc2011c9beeb4ffc', 's...
1  {'md5': '9a45b2520e930dc9186f6d93a7798a13', 's...
2  {'md5': '9a45b2520e930dc9186f6d93a7798a13', 's...
3  {'md5': '1c8cc6299bd654bbcd85710968d6a87c', 's...

                                md5                                      sha1  \
0  6cd1486db221e532cc2011c9beeb4ffc  6e485467d7e06502046b7c84a8ef067cfe1512ad
1  9a45b2520e930dc9186f6d93a7798a13  f526c90fa0744e3a63d84421ff25e3f5a3d697cb
2  9a45b2520e930dc9186f6d93a7798a13  f526c90fa0744e3a63d84421ff25e3f5a3d697cb
3  1c8cc6299bd654bbcd85710968d6a87c  55377391141f59a2ff5ae4765d9f0b4438adfd73

                                              sha256  \
0  d3291dab1ae552b91e6b50d7460ceaa39f6f92b2cda433...
1  c1f6c05bdbe28a58557a9477cd0fa96fbc5e7c54ceb605...
2  c1f6c05bdbe28a58557a9477cd0fa96fbc5e7c54ceb605...
3  fd80ceba7cfb49d296886c10d9a3497d63c89a589587cd...

                                     printable_bytes  src_index  \
0  24 00 74 00 20 00 3d 00 20 00 27 00 64 00 69 0...         39
1  69 a6 9a 69 a6 9a 69 a6 9a 69 a6 9a 69 a6 9a 6...         40
2  69 a6 9a 69 a6 9a 69 a6 9a 69 a6 9a 69 a6 9a 6...         41
3  f3 57 9d d3 77 1a 7f af 74 d5 ee 38 e1 ce f6 6...         44

                                 full_decoded_string
0  .\powershell  -enc <decoded type='string' name...
1  cmd  /c "echo # <decoded type='string' name='[...
2  cmd  /c "echo # <decoded type='string' name='[...
3  implant.exe  <decoded type='string' name='[Non...

Interpreting the DataFrame output

For simple strings the Base64 decoded output is straightforward. However for nested encodings this can get a little complex and difficult to represent in a tabular format.

Output columns

  • reference - The index of the row item in dotted notation, in depth.seq pairs (e.g. 1.2.2.3 would be the 3 item at depth 3 that is a child of the 2nd item found at depth 1). This may not always be an accurate notation - it is mainly use to allow you to associate an individual row with the reference value contained in the full_decoded_string column of the topmost item).

  • original_string - the original string before decoding.

  • file_name - filename, if any (only if this is an item in zip or tar file).

  • file_type - a guess at the file type (this is currently elementary and only includes a few file types).

  • input_bytes - the decoded bytes as a Python bytes string.

  • decoded_string - the decoded string if it can be decoded as a UTF-8 or UTF-16 string. Note: binary sequences may often successfully decode as UTF-16 strings but, in these cases, the decodings are meaningless.

  • encoding_type - encoding type (UTF-8 or UTF-16) if a decoding was possible, otherwise ‘binary’.

  • file_hashes - collection of file hashes for any decoded item.

  • md5 - md5 hash as a separate column.

  • sha1 - sha1 hash as a separate column.

  • sha256 - sha256 hash as a separate column.

  • printable_bytes - printable version of input_bytes as a string of \xNN values

  • src_index - the DataFrame index of the input row.

The src_index column allows you to merge the results with the input DataFrame.

Where an input row results in multiple decoded elements, (e.g. a nested encoding or encoded archive file), the output of this merge will result in duplicate rows from the input (one per decoded element). The row index of the input is preserved in the src_index.

Note

In order to merge output with input you may need to explictly force the type of the SourceIndex column. In the example below case we are matching with the default numeric index so we force the type to be numeric. In cases where you are using an index of a different dtype you will need to convert the SourceIndex (dtype=object) to match the type of your index column.

Note

the output of unpack_items() may have multiple rows (for nested encodings). In this case merged DF will have duplicate rows from the source.

# Set the type of the SourceIndex column.
dec_df['SourceIndex'] = pd.to_numeric(dec_df['src_index'])
merged_df = (process_tree
             .merge(right=dec_df, how='left', left_index=True, right_on='SourceIndex')
             .drop(columns=['Unnamed: 0'])
             .set_index('SourceIndex'))

# Show the result of the merge (only those rows that have a value in original_string)
merged_df.dropna(subset=['original_string'])

# Note the output of unpack_items() may have multiple rows (for nested encodings)
# In this case merged DF will have duplicate rows from the source.
                                         TenantId                     Account  \
SourceIndex
39           802d39e1-9d70-404d-832c-2de5e2478eda  MSTICAlertsWin1\MSTICAdmin
40           802d39e1-9d70-404d-832c-2de5e2478eda  MSTICAlertsWin1\MSTICAdmin
41           802d39e1-9d70-404d-832c-2de5e2478eda  MSTICAlertsWin1\MSTICAdmin
44           802d39e1-9d70-404d-832c-2de5e2478eda  MSTICAlertsWin1\MSTICAdmin

             EventID           TimeGenerated         Computer  \
SourceIndex
39              4688 2019-01-15 05:15:13.567  MSTICAlertsWin1
40              4688 2019-01-15 05:15:13.683  MSTICAlertsWin1
41              4688 2019-01-15 05:15:13.793  MSTICAlertsWin1
44              4688 2019-01-15 05:15:12.003  MSTICAlertsWin1

                                           SubjectUserSid SubjectUserName  \
SourceIndex
39           S-1-5-21-996632719-2361334927-4038480536-500      MSTICAdmin
40           S-1-5-21-996632719-2361334927-4038480536-500      MSTICAdmin
41           S-1-5-21-996632719-2361334927-4038480536-500      MSTICAdmin
44           S-1-5-21-996632719-2361334927-4038480536-500      MSTICAdmin

            SubjectDomainName SubjectLogonId NewProcessId  ...  \
SourceIndex                                                ...
39            MSTICAlertsWin1       0xfaac27       0x1684  ...
40            MSTICAlertsWin1       0xfaac27       0x16b8  ...
41            MSTICAlertsWin1       0xfaac27       0x16ec  ...
44            MSTICAlertsWin1       0xfaac27       0x1250  ...

                                                   input_bytes  \
SourceIndex
39           b"$\x00t\x00 \x00=\x00 \x00'\x00d\x00i\x00r\x0...
40           b'i\xa6\x9ai\xa6\x9ai\xa6\x9ai\xa6\x9ai\xa6\x9...
41           b'i\xa6\x9ai\xa6\x9ai\xa6\x9ai\xa6\x9ai\xa6\x9...
44           b'\xf3W\x9d\xd3w\x1a\x7f\xaft\xd5\xee8\xe1\xce...

                                                decoded_string encoding_type  \
SourceIndex
39           $t='dir';\r\n&('Invok...         utf-8
40                                                ꙩ榚骦ꙩ榚骦ꙩ榚骦ꙩ榚骦        utf-16
41                                                ꙩ榚骦ꙩ榚骦ꙩ榚骦ꙩ榚骦        utf-16
44                                                埳펝᩷꽿해㣮컡槶믎彷絶岿        utf-16

                                                   file_hashes  \
SourceIndex
39           {'md5': '6cd1486db221e532cc2011c9beeb4ffc', 's...
40           {'md5': '9a45b2520e930dc9186f6d93a7798a13', 's...
41           {'md5': '9a45b2520e930dc9186f6d93a7798a13', 's...
44           {'md5': '1c8cc6299bd654bbcd85710968d6a87c', 's...

                                          md5  \
SourceIndex
39           6cd1486db221e532cc2011c9beeb4ffc
40           9a45b2520e930dc9186f6d93a7798a13
41           9a45b2520e930dc9186f6d93a7798a13
44           1c8cc6299bd654bbcd85710968d6a87c

                                                 sha1  \
SourceIndex
39           6e485467d7e06502046b7c84a8ef067cfe1512ad
40           f526c90fa0744e3a63d84421ff25e3f5a3d697cb
41           f526c90fa0744e3a63d84421ff25e3f5a3d697cb
44           55377391141f59a2ff5ae4765d9f0b4438adfd73

                                                        sha256  \
SourceIndex
39           d3291dab1ae552b91e6b50d7460ceaa39f6f92b2cda433...
40           c1f6c05bdbe28a58557a9477cd0fa96fbc5e7c54ceb605...
41           c1f6c05bdbe28a58557a9477cd0fa96fbc5e7c54ceb605...
44           fd80ceba7cfb49d296886c10d9a3497d63c89a589587cd...

                                               printable_bytes src_index  \
SourceIndex
39           24 00 74 00 20 00 3d 00 20 00 27 00 64 00 69 0...      39.0
40           69 a6 9a 69 a6 9a 69 a6 9a 69 a6 9a 69 a6 9a 6...      40.0
41           69 a6 9a 69 a6 9a 69 a6 9a 69 a6 9a 69 a6 9a 6...      41.0
44           f3 57 9d d3 77 1a 7f af 74 d5 ee 38 e1 ce f6 6...      44.0

                                           full_decoded_string
SourceIndex
39           .\powershell  -enc <decoded type='string' name...
40           cmd  /c "echo # <decoded type='string' name='[...
41           cmd  /c "echo # <decoded type='string' name='[...
44           implant.exe  <decoded type='string' name='[Non...

[4 rows x 36 columns]

Decoding Nested Base64/Archives

The module will try to follow nested encodings. It uses the following algorithm:

  1. Search for a pattern in the input that looks like a Base64 encoded string

  2. If not a known undecodable_string, try to decode the matched pattern.

    • If the base 64 string matches a known archive type (zip, tar, gzip) also decompress or unpack

      • For multi-item archives (zip, tar) process each contained item recursively (i.e. go to item 1. with child item as input)

    • For anything that decodes to a UTF-8 or UTF-16 string replace the input pattern with the decoded string

    • Recurse over resultant output (i.e. submit decoded/replaced string to 1.)

  3. If decoding fails, add to list of undecodable_strings (prevents infinite looping over something that looks like a base64 string but isn’t)

encoded_cmd = '''
powershell.exe  -nop -w hidden -encodedcommand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'''

import re
dec_string, dec_df = base64.unpack(input_string=encoded_cmd)
print(dec_string.replace('<decoded', '\n<decoded'))
powershell.exe  -nop -w hidden -encodedcommand
<decoded value='multiple binary' type='multiple' index='1' depth='1'>
<decoded type='string' name='[zip] Filename: PeDll.dll' index='1.1' depth='2'>笾뿴䅐〳⌁㾝䲍ǔ庰퀈쐿Č熠倀℈ꚜᨒ腦㽋ꚩ黓恓⊀́€᠀菀ﳶ态ꨭꪪꪪꪪꪪꪪꨊ᪪耚ⶡꪪꪪꪪꪪ⪪ᆢ죺ſﵠ쀇׿タ쀇׿`䘁대䀃蜨榑v⃈Σ ።</decoded>
<decoded type='string' name='[zip] Filename: b64inzip.foo' index='1.2' depth='2'>Base64 encoded string in zip file</decoded>
<decoded type='string' name='[zip] Filename: PlainTextInZip.dll' index='1.3' depth='2'>Unencoded text file in zip</decoded></decoded>

IPython magic

You can use the cell magic %%b64 to decode text directly in a cell

The b64 magic supports the following options:

--out OUT, -o OUT  The variable to return the results in the variable `OUT`
                   Note: the output is a tuple of decoded string and pandas DataFrame
--pretty, -p       Print formatted version of output (if you `print` the output)
--clean, -c        Print decoded string with no formatting
%%b64 --pretty --out dec_xml
powershell.exe  -nop -w hidden -encodedcommand
UEsDBBQAAAAIAGBXkk3LfdszdwAAAIoAAAAJAAAAUGVEbGwuZGxss6v+sj/A0diA
UXmufa/PFcYNcRwX7I/wMC4oZAjgUJyzTEgqrdHbfuWyy/OCExqUGJkZGBoYoEDi
QPO3P4wJuqsQgGvVKimphoUIIa1Fgr9OMLyoZ0z4y37gP2vDfxDp8J/RjWEzs4NG
+8TMMoYTCouZGRSShAFQSwMEFAAAAAAAYYJrThx8YzUhAAAAIQAAAAwAAABiNjRp
bnppcC5mb29CYXNlNjQgZW5jb2RlZCBzdHJpbmcgaW4gemlwIGZpbGVQSwMEFAAA
AAAAi4JrTvMfsJUaAAAAGgAAABIAAABQbGFpblRleHRJblppcC5kbGxVbmVuY29k
ZWQgdGV4dCBmaWxlIGluIHppcFBLAQIUABQAAAAIAGBXkk3LfdszdwAAAIoAAAAJ
AAAAAAAAAAAAIAAAAAAAAABQZURsbC5kbGxQSwECFAAUAAAAAABhgmtOHHxjNSEA
AAAhAAAADAAAAAAAAAABACAAAACeAAAAYjY0aW56aXAuZm9vUEsBAhQAFAAAAAAA
i4JrTvMfsJUaAAAAGgAAABIAAAAAAAAAAQAgAAAA6QAAAFBsYWluVGV4dEluWmlw
LmRsbFBLBQYAAAAAAwADALEAAAAzAQAAAAA=
'<?xml version="1.0" encoding="utf-8"?>n<decoded_string>n powershell.exe  -nop -w hidden -encodedcommandn <decoded depth="1" index="1" type="multiple" value="multiple binary">n  <decoded depth="2" index="1.1" name="[zip] Filename: PeDll.dll" type="zip" value="binary">n   3e 7b f4 bf 50 41 33 30 01 23 9d 3f 8d 4c d4 01 b0 5e 08 d0 3f c4 0c 01 a0 71 00 50 08 21 9c a6 12 1a 66 81 4b 3f a9 a6 d3 9e 53 60 80 22 01 03 00 00 80 00 00 00 00 00 00 00 00 18 c0 83 f6 fc 01 60 2d aa aa aa aa aa aa aa aa aa aa aa 0a aa aa 1a 1a 80 a1 2d aa aa aa aa aa aa aa aa aa 2a a2 11 fa c8 00 e8 7f 01 60 fd 07 c0 ff 05 80 ff 07 c0 ff 05 40 ff 01 46 00 b3 03 40 28 87 91 69 76 00 c8 20 a3 03 00 20 62 13n  </decoded>n  <decoded depth="2" index="1.2" name="[zip] Filename: b64inzip.foo" type="string">n   Base64 encoded string in zip filen  </decoded>n  <decoded depth="2" index="1.3" name="[zip] Filename: PlainTextInZip.dll" type="string">n   Unencoded text file in zipn  </decoded>n </decoded>n</decoded_string>'

Display the pretty-printed version of the decoded string.

print(dec_xml[0])
<?xml version="1.0" encoding="utf-8"?>
<decoded_string>
 powershell.exe  -nop -w hidden -encodedcommand
 <decoded depth="1" index="1" type="multiple" value="multiple binary">
  <decoded depth="2" index="1.1" name="[zip] Filename: PeDll.dll" type="zip" value="binary">
   3e 7b f4 bf 50 41 33 30 01 23 9d 3f 8d 4c d4 01 b0 5e 08 d0 3f c4 0c 01 a0 71 00 50 08 21 9c a6 12 1a 66 81 4b 3f a9 a6 d3 9e 53 60 80 22 01 03 00 00 80 00 00 00 00 00 00 00 00 18 c0 83 f6 fc 01 60 2d aa aa aa aa aa aa aa aa aa aa aa 0a aa aa 1a 1a 80 a1 2d aa aa aa aa aa aa aa aa aa 2a a2 11 fa c8 00 e8 7f 01 60 fd 07 c0 ff 05 80 ff 07 c0 ff 05 40 ff 01 46 00 b3 03 40 28 87 91 69 76 00 c8 20 a3 03 00 20 62 13
  </decoded>
  <decoded depth="2" index="1.2" name="[zip] Filename: b64inzip.foo" type="string">
   Base64 encoded string in zip file
  </decoded>
  <decoded depth="2" index="1.3" name="[zip] Filename: PlainTextInZip.dll" type="string">
   Unencoded text file in zip
  </decoded>
 </decoded>
</decoded_string>

Pandas Extension

The decoding functionality is also available in a pandas extension mp_b64. This supports a single method extract().

This supports the same syntax as unpack_df (described earlier). It returns a DataFrame with the decoded contents (this may be multiple output lines for inputs where there is a nested encoding.

process_tree.mp_b64.extract(column='CommandLine')

To-Do Items

  • Use more comprehensive list of binary magic numbers and match on byte values after decoding to get better file typing

  • Output nested decodings in a more readable output

  • Add a pandas pipe() partial function to allow inline decoding in a pands pipeline. E.g.

my_df = pd.read_cs('input.csv').b64decode(column='CommandLine').drop_duplicates().some_func()