external/anthropic-cybersecurity-skills/skills/detecting-aws-guardduty-findings-automation/SKILL.md
Automate AWS GuardDuty threat detection findings processing using EventBridge and Lambda to enable real-time incident response, automatic quarantine of compromised resources, and security notification workflows.
npx skillsauth add seikaikyo/dash-skills detecting-aws-guardduty-findings-automationInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Amazon GuardDuty is a threat detection service that continuously monitors AWS accounts for malicious activity and unauthorized behavior. By integrating GuardDuty with Amazon EventBridge and AWS Lambda, security teams achieve automated, real-time responses to threats, reducing mean time to response (MTTR) from hours to seconds. GuardDuty analyzes VPC Flow Logs, CloudTrail management and data events, DNS logs, EKS audit logs, and S3 data events.
# Enable GuardDuty
aws guardduty create-detector --enable --finding-publishing-frequency FIFTEEN_MINUTES
# Enable additional data sources
aws guardduty update-detector \
--detector-id DETECTOR_ID \
--data-sources '{
"S3Logs": {"Enable": true},
"Kubernetes": {"AuditLogs": {"Enable": true}},
"MalwareProtection": {"ScanEc2InstanceWithFindings": {"EbsVolumes": true}},
"RuntimeMonitoring": {"Enable": true}
}'
{
"source": ["aws.guardduty"],
"detail-type": ["GuardDuty Finding"],
"detail": {
"severity": [{"numeric": [">=", 7.0]}]
}
}
aws events put-rule \
--name "guardduty-high-severity" \
--event-pattern '{
"source": ["aws.guardduty"],
"detail-type": ["GuardDuty Finding"],
"detail": {
"severity": [{"numeric": [">=", 7.0]}]
}
}'
aws events put-targets \
--rule "guardduty-high-severity" \
--targets "Id"="lambda-handler","Arn"="arn:aws:lambda:us-east-1:123456789012:function:guardduty-response"
import boto3
import json
import os
ec2 = boto3.client('ec2')
sns = boto3.client('sns')
QUARANTINE_SG = os.environ.get('QUARANTINE_SECURITY_GROUP')
SNS_TOPIC = os.environ.get('SNS_TOPIC_ARN')
def lambda_handler(event, context):
finding = event['detail']
finding_type = finding['type']
severity = finding['severity']
account_id = finding['accountId']
region = finding['region']
# Extract resource information
resource = finding.get('resource', {})
resource_type = resource.get('resourceType', '')
if resource_type == 'Instance':
instance_id = resource['instanceDetails']['instanceId']
instance_tags = {t['key']: t['value']
for t in resource['instanceDetails'].get('tags', [])}
# Skip if already quarantined
if instance_tags.get('SecurityStatus') == 'Quarantined':
return {'statusCode': 200, 'body': 'Already quarantined'}
# Get current security groups for forensics
instance = ec2.describe_instances(InstanceIds=[instance_id])
current_sgs = [sg['GroupId'] for sg in
instance['Reservations'][0]['Instances'][0]['SecurityGroups']]
# Tag instance with finding info and original SGs
ec2.create_tags(
Resources=[instance_id],
Tags=[
{'Key': 'SecurityStatus', 'Value': 'Quarantined'},
{'Key': 'GuardDutyFinding', 'Value': finding_type},
{'Key': 'OriginalSecurityGroups', 'Value': ','.join(current_sgs)},
{'Key': 'QuarantineTime', 'Value': finding['updatedAt']}
]
)
# Move to quarantine security group (blocks all traffic)
if QUARANTINE_SG:
ec2.modify_instance_attribute(
InstanceId=instance_id,
Groups=[QUARANTINE_SG]
)
# Create EBS snapshots for forensics
volumes = ec2.describe_volumes(
Filters=[{'Name': 'attachment.instance-id', 'Values': [instance_id]}]
)
for vol in volumes['Volumes']:
ec2.create_snapshot(
VolumeId=vol['VolumeId'],
Description=f'GuardDuty forensic snapshot - {finding_type}',
TagSpecifications=[{
'ResourceType': 'snapshot',
'Tags': [
{'Key': 'Purpose', 'Value': 'ForensicCapture'},
{'Key': 'SourceInstance', 'Value': instance_id},
{'Key': 'FindingType', 'Value': finding_type}
]
}]
)
# Notify security team
sns.publish(
TopicArn=SNS_TOPIC,
Subject=f'[GuardDuty] {finding_type} - Instance {instance_id} Quarantined',
Message=json.dumps({
'action': 'instance_quarantined',
'instance_id': instance_id,
'finding_type': finding_type,
'severity': severity,
'account': account_id,
'region': region,
'original_security_groups': current_sgs,
'description': finding.get('description', '')
}, indent=2)
)
return {
'statusCode': 200,
'body': f'Instance {instance_id} quarantined and snapshots created'
}
return {'statusCode': 200, 'body': 'Non-EC2 finding processed'}
import boto3
import json
import os
iam = boto3.client('iam')
sns = boto3.client('sns')
SNS_TOPIC = os.environ.get('SNS_TOPIC_ARN')
def lambda_handler(event, context):
finding = event['detail']
finding_type = finding['type']
if 'IAMUser' not in finding_type and 'UnauthorizedAccess' not in finding_type:
return {'statusCode': 200, 'body': 'Not an IAM finding'}
resource = finding.get('resource', {})
access_key_details = resource.get('accessKeyDetails', {})
user_name = access_key_details.get('userName', '')
access_key_id = access_key_details.get('accessKeyId', '')
if not user_name:
return {'statusCode': 200, 'body': 'No user identified'}
actions_taken = []
# Deactivate the compromised access key
if access_key_id and access_key_id != 'GeneratedFindingAccessKeyId':
try:
iam.update_access_key(
UserName=user_name,
AccessKeyId=access_key_id,
Status='Inactive'
)
actions_taken.append(f'Deactivated access key {access_key_id}')
except Exception as e:
actions_taken.append(f'Failed to deactivate key: {str(e)}')
# Attach deny-all policy to user
deny_policy = {
"Version": "2012-10-17",
"Statement": [{
"Effect": "Deny",
"Action": "*",
"Resource": "*"
}]
}
try:
iam.put_user_policy(
UserName=user_name,
PolicyName='GuardDuty-DenyAll-Quarantine',
PolicyDocument=json.dumps(deny_policy)
)
actions_taken.append(f'Applied deny-all policy to {user_name}')
except Exception as e:
actions_taken.append(f'Failed to apply deny policy: {str(e)}')
# Notify
sns.publish(
TopicArn=SNS_TOPIC,
Subject=f'[GuardDuty] IAM Compromise - {user_name}',
Message=json.dumps({
'finding_type': finding_type,
'user': user_name,
'access_key': access_key_id,
'actions_taken': actions_taken,
'severity': finding['severity']
}, indent=2)
)
return {'statusCode': 200, 'body': json.dumps(actions_taken)}
resource "aws_guardduty_detector" "main" {
enable = true
finding_publishing_frequency = "FIFTEEN_MINUTES"
datasources {
s3_logs { enable = true }
kubernetes { audit_logs { enable = true } }
malware_protection {
scan_ec2_instance_with_findings {
ebs_volumes { enable = true }
}
}
}
}
resource "aws_cloudwatch_event_rule" "guardduty_high" {
name = "guardduty-high-severity"
description = "GuardDuty high severity findings"
event_pattern = jsonencode({
source = ["aws.guardduty"]
detail-type = ["GuardDuty Finding"]
detail = {
severity = [{ numeric = [">=", 7.0] }]
}
})
}
resource "aws_cloudwatch_event_target" "lambda" {
rule = aws_cloudwatch_event_rule.guardduty_high.name
arn = aws_lambda_function.guardduty_response.arn
}
| Category | Severity Range | Examples | |----------|---------------|---------| | Backdoor | 5.0 - 8.0 | Backdoor:EC2/C&CActivity | | CryptoCurrency | 5.0 - 8.0 | CryptoCurrency:EC2/BitcoinTool | | Trojan | 5.0 - 8.0 | Trojan:EC2/BlackholeTraffic | | UnauthorizedAccess | 5.0 - 8.0 | UnauthorizedAccess:IAMUser/ConsoleLogin | | Recon | 2.0 - 5.0 | Recon:EC2/PortProbeUnprotected | | Persistence | 5.0 - 8.0 | Persistence:IAMUser/AnomalousBehavior |
# Designate GuardDuty administrator
aws guardduty enable-organization-admin-account \
--admin-account-id 111111111111
# Auto-enable for new accounts
aws guardduty update-organization-configuration \
--detector-id DETECTOR_ID \
--auto-enable
development
拋棄式 HTML mockup 比稿:產出 2 到 3 個設計立場不同的變體(密度 / 版式 / 強調軸,不是換色),各附取捨說明,最後給有立場的對比結論。適用:「畫個草圖」「比較 A 版 B 版」「先看方向再做」「給我看幾種做法」。要 production 元件或設計已定案時不適用。
tools
需求不明時的意圖萃取訪談:一次一題、每題附上自己的猜測、聽出「真正想要 vs 覺得應該要」,直到能預測使用者反應(約 95% 信心)才動工。適用:需求缺少對象 / 動機 / 成功標準 / 約束,或使用者點名「訪談我」「先確認一下」「我們確定嗎」。明確自足的指示、純資訊查詢、機械性操作不適用。
development
對非平凡決策啟動新鮮 context 對抗審查(找碴不背書),在修正還便宜的時候抓出錯誤方向。適用:高風險改動(production、資安敏感邏輯、不可逆操作)、不熟的程式碼、要宣稱「這樣是安全的 / 可行的」之前。機械性操作與一行修改不適用。
testing
Reference for writing and editing agent skills well — the vocabulary and principles that make a skill predictable. Consult when authoring, reviewing, or pruning a SKILL.md.