DFIR Summary Commands

aws-ct-summary command

This command creates a summary of the following information based on user ARNs: - Total number of events (API calls) (NumOfEvents) - Timestamp of the first API call found in the logs (FirstTimestamp) - Timestamp of the last API call found in the logs (LastTimestamp) - Commonly abused API calls that were successful (AbusedAPIs-Success) - Commonly abused API calls that were attempted but failed (AbusedAPIs-Failed) - Other API calls that are not on the list of commonly abused API calls that were successful (OtherAPIs-Success) - Other API calls that are not on the list of commonly abused API calls that were attempted but failed (OtherAPIs-Failed) - AWS regions of where the API calls were made (AWS-Regions) - Source IP addresses of the API call (SrcIPs) - User types (UserTypes) - User access key IDs (UserAccessKeyIDs) - User agents of the source who made the API call (UserAgents)

※ Note: the API calls that are commonly abused come from the config file hosted at https://github.com/Yamato-Security/suzaku-rules/blob/main/config/abused_aws_api_calls.csv. This file will be updated over time and will be locally synced every time you run the update-rules command.

These results are intended to provide analysts with information to discover compromised accounts or attacks that does not rely on specific signatures. For example, you can check to see if certain users are calling suspicious API calls that they shouldn't be calling, using regions that are not typically used, logging in from suspicious source IP addresses or with suspicious user agents, etc... After you found any suspicious API calls being called, source IP addresses or user agents, you can quickly determine which access keys might have been abused during that timeframe and pivot on those keywords in the original JSON logs to create a timeline of attacker activity.

Warning: there will be a lot of data in the cells and will most likely not display well in programs like Excel. Please use Numbers on a Mac, Timeline Explorer on Windows, etc...

See also: this command breaks the information down per user ARN. To see the same fields (source IPs, user agents, regions, access key IDs, ...) aggregated across all events, with one row per value that you can sort and filter, use the aws-ct-metrics command.

AbusedAPIs-Success example:

Unique APIs: 11 | Total APIs 477,373
415,552 - RunInstances (ec2.amazonaws.com) - Spin up EC2 instances (crypto mining, tools) (2019-08-23 06:00:07 ~ 2019-08-23 06:00:07)
28,907 - GetBucketAcl (s3.amazonaws.com) - S3 recon (2019-08-21 08:03:03 ~ 2019-10-21 13:59:40)
10,095 - GetCallerIdentity (sts.amazonaws.com) - Current credentials recon (2019-08-23 06:00:07 ~ 2019-08-23 06:04:14)
9,936 - ListBuckets (s3.amazonaws.com) - S3 recon (2019-08-23 06:00:07 ~ 2019-08-23 06:14:53)
9,168 - DescribeInstances (ec2.amazonaws.com) - EC2 and network layout recon (2019-08-23 06:00:07 ~ 2019-08-23 06:04:20)
3,658 - DescribeVpcs (ec2.amazonaws.com) - EC2 and network layout recon (2019-08-21 08:03:03 ~ 2019-09-12 20:00:44)
19 - ListGroups (greengrass.amazonaws.com) - IAM enumeration (2019-08-21 08:03:03 ~ 2019-10-19 23:49:25)
14 - DescribeInstances (opsworks.amazonaws.com) - EC2 and network layout recon (2019-08-21 08:03:03 ~ 2019-10-19 23:49:22)
12 - GetBucketPolicy (s3.amazonaws.com) - S3 recon (2019-01-08 20:30:01 ~ 2020-03-29 09:06:56)
7 - ListGroups (resource-groups.amazonaws.com) - IAM enumeration (2019-01-08 20:30:01 ~ 2020-03-29 09:06:56)
5 - StartInstances (ec2.amazonaws.com) - Spin up EC2 instances (crypto mining, tools) (2019-08-21 08:03:03 ~ 2019-12-12 07:07:31)

AbusedAPIs-Failed example:

Unique APIs: 23 | Total APIs 20,464
11,603 - AssumeRole (sts.amazonaws.com) - Lateral movement via roles (2019-08-21 08:03:03 ~ 2019-09-18 07:04:12)
7,279 - GetBucketAcl (s3.amazonaws.com) - S3 recon (2019-08-21 08:03:03 ~ 2019-09-09 09:01:26)
515 - GetBucketPolicy (s3.amazonaws.com) - S3 recon (2019-08-21 08:03:03 ~ 2019-10-01 19:11:07)
331 - ListUsers (iam.amazonaws.com) - IAM enumeration (2019-08-21 08:03:03 ~ 2019-08-29 14:53:14)
210 - ListSecrets (secretsmanager.amazonaws.com) - Find credential storage locations (2019-08-21 08:03:03 ~ 2019-10-19 23:49:30)
153 - ListGroups (iam.amazonaws.com) - IAM enumeration (2019-08-21 08:03:03 ~ 2019-09-12 15:24:39)
148 - ListRoles (iam.amazonaws.com) - IAM enumeration (2019-08-21 08:03:03 ~ 2019-09-12 15:20:56)
112 - ListAccessKeys (iam.amazonaws.com) - Enumerates keys on IAM users (2019-08-21 08:03:03 ~ 2019-09-16 14:28:15)
31 - ListGroups (greengrass.amazonaws.com) - IAM enumeration (2019-08-21 08:03:03 ~ 2020-02-25 14:41:24)
...

OtherAPIs-Success example:

Unique APIs: 289 | Total APIs 143,759
98,689 - DescribeSnapshots (ec2.amazonaws.com) (2019-08-23 06:00:07 ~ 2019-08-23 06:50:59)
10,679 - DescribeSpotPriceHistory (ec2.amazonaws.com) (2019-08-21 08:03:03 ~ 2019-09-12 20:07:32)
3,740 - DescribeReservedInstancesOfferings (ec2.amazonaws.com) (2019-08-21 08:03:03 ~ 2019-09-12 20:07:30)
2,372 - DescribeSnapshotAttribute (ec2.amazonaws.com) (2019-08-21 08:03:03 ~ 2019-08-24 12:38:34)
2,307 - CreateDefaultVpc (ec2.amazonaws.com) (2019-08-23 06:00:07 ~ 2019-08-23 06:04:17)
1,532 - DescribeKeyPairs (ec2.amazonaws.com) (2019-08-23 06:00:07 ~ 2019-08-23 06:04:16)
1,504 - DescribeSecurityGroups (ec2.amazonaws.com) (2019-08-21 08:03:03 ~ 2019-09-12 20:00:40)
1,495 - DescribeImages (ec2.amazonaws.com) (2019-08-21 08:03:03 ~ 2019-09-12 20:07:26)
1,438 - CreateKeyPair (ec2.amazonaws.com) (2019-08-23 06:00:07 ~ 2019-08-23 06:04:16)
1,402 - DescribeVolumes (ec2.amazonaws.com) (2019-08-21 08:03:03 ~ 2019-09-03 12:06:20)
1,217 - DescribeSubnets (ec2.amazonaws.com) (2019-08-21 08:03:03 ~ 2019-09-12 20:00:43)
...

AWS-Regions example:

Total regions: 16
167,155 - us-west-2 (2019-08-23 06:00:07 ~ 2019-08-23 06:14:53)
113,328 - us-east-1 (2019-08-23 06:00:07 ~ 2019-08-23 06:04:14)
65,718 - ap-northeast-2 (2019-08-23 06:00:07 ~ 2019-08-23 06:22:55)
64,787 - ap-northeast-1 (2019-08-23 06:00:07 ~ 2019-08-23 06:34:57)
...

SrcIPs example:

Total source IPs: 5,293
634,454 - 5.205.62.253 (2019-08-23 06:00:07 ~ 2019-08-23 06:00:07)
23,498 - 193.29.252.218 (2019-08-21 08:03:00 ~ 2019-10-17 09:11:22)
15,925 - 155.63.17.217 (2018-04-17 10:09:00 ~ 2020-09-21 21:06:22)
9,067 - 253.0.255.253 (2018-04-17 10:09:00 ~ 2020-09-21 21:06:22)
7,078 - 163.21.250.220 (2018-04-17 10:09:00 ~ 2020-09-21 21:06:22)
6,575 - 236.9.245.88 (2018-04-17 10:09:00 ~ 2020-09-21 21:06:22)
5,138 - 84.252.252.117 (2019-01-08 20:30:01 ~ 2020-03-29 09:06:56)
4,946 - 24.251.252.2 (2019-08-21 08:03:00 ~ 2019-09-30 06:36:13)
4,225 - 211.111.151.81 (2019-08-21 08:03:00 ~ 2019-09-12 19:53:35)
...

UserType example:

IAMUser

UserAccessKeyIDs example:

Total access key ids: 629
667,476 - AKIA01U43UX3RBRDXF4Q (2019-08-23 06:00:07 ~ 2019-08-23 06:00:07)
218,544 - ASIARF55FBMFZBXLKDFW (2019-08-21 11:31:47 ~ 2019-08-23 13:00:28)
12,677 - AKIA1ZBTOEKWKVHP6GHZ (2017-02-12 21:15:12 ~ 2020-09-21 21:06:22)
8,822 - ASIAGD2JRX0V6RJGWR59 (2018-04-17 10:09:00 ~ 2020-09-21 21:06:22)
4,940 - ASIAUNHV6EHIK5MNPEKF (2019-08-21 08:03:00 ~ 2019-09-30 06:36:17)
...

Note that since there will be a large amount of temporary AWS STS access key IDs, these are filtered out by default. Add the -s, --include-sts-keys option if you want to include these.

UserAgents example:

Total user agents: 7,760
351,022 - Boto3/1.9.201 Python/2.7.12 Linux/4.4.0-159-generic Botocore/1.12.201 (2019-08-23 06:00:07 ~ 2019-08-23 06:00:07)
283,430 - Boto3/1.9.201 Python/2.7.12 Linux/4.4.0-157-generic Botocore/1.12.201 (2019-08-21 11:31:47 ~ 2019-08-23 13:00:28)
23,467 - [Boto3/1.15.13 Python/3.8.5 Darwin/19.6.0 Botocore/1.18.13 Resource] (2017-02-12 21:15:12 ~ 2020-10-07 16:05:52)
15,924 - Boto3/1.7.4 Python/2.7.12 Linux/4.4.0-119-generic Botocore/1.10.4 (2018-04-17 10:09:00 ~ 2020-09-21 21:06:22)
9,400 - aws-sdk-java/1.11.301 Linux/4.9.93-linuxkit-aufs Java_HotSpot(TM)_64-Bit_Server_VM/25.172-b11 java/1.8.0_172 (2018-04-17 10:09:00 ~ 2020-09-21 21:06:22)
6,372 - [Boto3/1.7.65 Python/3.5.2 Linux/4.13.0-37-generic Botocore/1.10.65 Resource] (2018-04-17 10:09:00 ~ 2020-09-21 21:06:22)
5,352 - AWSPowerShell/3.3.365.0 .NET_Runtime/4.0 .NET_Framework/4.0 OS/Microsoft_Windows_NT_10.0.01985.0 WindowsPowerShell/5.0 ClientSync (2019-01-08 20:30:01 ~ 2020-03-29 09:06:56)
4,945 - Boto3/1.9.231 Python/2.7.15+ Linux/4.15.0-64-generic Botocore/1.12.231 (2019-08-21 08:03:00 ~ 2019-09-30 06:36:13)
4,599 - [aws-cli/1.16.301 Python/3.7.6 Linux/5.4.0-kali3-amd64 botocore/1.13.37] (2019-08-21 08:03:00 ~ 2020-02-09 22:00:32)
3,909 - Boto3/1.14.28 Python/3.8.5 Linux/5.7.0-kali1-amd64 Botocore/1.17.28 (2019-01-08 20:30:01 ~ 2020-09-11 17:35:39)
3,450 - Boto3/1.4.2 Python/2.7.13+ Linux/4.9.0-3-amd64 Botocore/1.5.19 (2017-02-12 21:15:12 ~ 2020-09-21 21:06:22)
3,198 - Boto3/1.4.2 Python/2.7.14 Linux/4.13.0-1-amd64 Botocore/1.5.19 (2017-02-12 21:15:12 ~ 2020-09-21 21:06:22)
...

Note that the aws client tool will include the OS information in the user agent so it is possible to detect if there were API calls made from attacker OSes like kali.

Command usage

Usage:
  suzaku aws-ct-summary <INPUT> [OPTIONS]

Input:
  -d, --directory <DIR>  Directory of multiple gz/json files
  -f, --file <FILE>      File path to one gz/json file

Filtering:
  -s, --include-sts-keys       Include temporary AWS STS access key IDs
      --timeline-start <DATE>  Start time of the events to load (ex: "2022-02-22T23:59:59Z)
      --timeline-end <DATE>    End time of the events to load (ex: "2020-02-22T00:00:00Z")
      --time-offset <OFFSET>   Scan recent events based on an offset (ex: 1y, 3M, 30d, 24h, 30m)
      --file-date-from <DATE>  Filter files by start date based on AWSLogs S3 path date structure (ex: "20240101")
      --file-date-to <DATE>    Filter files by end date based on AWSLogs S3 path date structure (ex: "20241231")

Output:
  -C, --clobber                   Overwrite files when saving
  -D, --hide-descriptions         Hide description of the commonly abused API calls
  -G, --geo-ip <MAXMIND-DB-DIR>   Add GeoIP (ASN, city, country) info to IP addresses [alias: --GeoIP]
  -o, --output <FILE>             Save the results to a file
  -t, --output-type <FORMAT,...>  Output format(s): csv (default), json, jsonl, duckdb. Comma-separate or repeat to write several at once, e.g. -t csv,duckdb [default: csv] [possible values: csv, json, jsonl, duckdb]

General Options:
  -h, --help  Show the help menu

Display Settings:
  -K, --no-color  Disable color output
  -q, --quiet     Quiet mode: do not display the launch banner

Output formats

-t, --output-type takes format names and accepts several at once, comma-separated or repeated. Each format is written to <output>.<ext>, so a single -o base path can produce more than one file:

./suzaku aws-ct-summary -d ../suzaku-sample-data -o summary -t csv,duckdb

• csv (default) — one row per principal, with each principal's API calls and attributes folded into multi-line cells. Best read in Timeline Explorer or Numbers rather than Excel.
• json / jsonl — the same data with those cells kept as nested arrays.
• duckdb — a self-contained DuckDB database. Because the CSV's multi-line cells cannot be queried, the nested data is stored relationally across three tables, plus the suzaku_meta provenance table:

Table	One row per	Columns
summary	principal	UserARN, NumOfEvents, FirstTimestamp, LastTimestamp, UserTypes
summary_api_calls	principal + API	UserARN, IsAbused, Outcome, API, EventSource, Description, Count, FirstSeen, LastSeen
summary_attributes	principal + value	UserARN, Attribute, Value, Count, FirstSeen, LastSeen

Values are typed rather than rendered:

• FirstTimestamp / LastTimestamp / FirstSeen / LastSeen are TIMESTAMP, and NumOfEvents / Count are BIGINT.
• IsAbused is a BOOLEAN and Outcome an ENUM('success','failed') — the two independent facts the old single Category string packed together.
• API holds the action only (RunInstances); the service moved to its own EventSource column, spelled like the timeline column holding the same fact.
• UserTypes is a VARCHAR[] listing every identity type the principal was seen with.
• Attribute is one of AwsRegion, SrcIP, UserAccessKeyID, UserAgent — again the same spelling as the timeline columns.
• - and empty placeholders are written as NULL.

This makes questions the CSV cannot answer into ordinary joins — for example, which source IPs were used by the principals that called an abused API:

SELECT a.API, SUM(a.Count) AS calls, COUNT(DISTINCT t.Value) AS src_ips
FROM summary_api_calls a
JOIN summary_attributes t
  ON t.UserARN = a.UserARN AND t.Attribute = 'SrcIP'
WHERE a.IsAbused
GROUP BY a.API
ORDER BY calls DESC;

aws-ct-metrics command example

• Save results to a CSV file: ./suzaku aws-ct-summary -d ../suzaku-sample-data -o sample-summary.csv