Detects the first successful creation of a Model Context Protocol (MCP) server integration name in an Anthropic organization within the rule history window. MCP servers add external data pathways into Claude and can expose organizational data to third-party infrastructure. This is a New Terms rule keyed on organization.id and anthropic.audit.mcp_server_name so the same connector name can still alert in another tenant, while routine re-creation of an already-seen name in the same organization does not.

Read More

Detects when a Claude project owner or editor role is granted through a role_assignment_granted event. Project owners and editors can access project chats, artifacts, and knowledge bases that may hold sensitive data. An attacker with organization access can grant these roles to persist access to high-value project content without holding organization admin privileges.

Read More

This rule detects network events that may indicate inbound Windows file sharing (SMB or CIFS) traffic originating from the Internet. SMB should never be directly reachable from the Internet, as it is a primary target for exploitation by threat actors seeking initial access. Inbound SMB from a public IP is a direct precondition for attacks such as EternalBlue (MS17-010) and related SMB remote code execution vulnerabilities.

Read More

This rule detects network events that may indicate the use of Windows file sharing (also called SMB or CIFS) traffic to the Internet. SMB is commonly used within networks to share files, printers, and other system resources amongst trusted systems. It should almost never be directly exposed to the Internet, as it is frequently targeted and exploited by threat actors as an initial access or backdoor vector or for data exfiltration.

Read More

Identifies the creation of a Process ID (PID), lock or reboot file created in temporary file storage paradigm (tmpfs) directory /var/run. On Linux, the PID files typically hold the process ID to track previous copies running and manage other tasks. Certain Linux malware use the /var/run directory for holding data, executables and other tasks, disguising itself or these files as legitimate PID files.

Read More

This rule detects successful authentications via PAM grantors that are not commonly used. This could indicate an attacker is attempting to escalate privileges or maintain persistence on the system by modifying the default PAM configuration.

Read More

Identifies the first time, within a lookback window, an identity performs AWS Organizations or IAM account enumeration APIs. Attackers with compromised credentials often map the organization (accounts, OUs, roots, delegated admins) and account-level metadata (aliases, summary) using the AWS CLI or SDKs. This is a New Terms rule detecting a rare occurrence of the cloud.account.id and user.name pair for these actions.

Read More

Identifies AWS API activity originating from uncommon desktop client applications based on the user agent string. This rule detects S3 Browser and Cyberduck, which are graphical S3 management tools that provide bulk upload/download capabilities. While legitimate, these tools are rarely used in enterprise environments and have been observed in use by threat actors for data exfiltration. Any activity from these clients should be validated against authorized data transfer workflows.

Read More

Detects the first time an AWS identity submits an AWS Batch job with a container command override ("containerOverrides.command"), indicating a runtime-modified execution environment. Command overrides allow the submitter to replace the default command of a job definition at submission time. This flexibility is commonly abused by adversaries to inject malicious commands or exfiltration logic into otherwise legitimate Batch compute environments without modifying the underlying job definition — making the malicious activity harder to detect through configuration review alone.

Read More

Detects modification of deployed Amazon Bedrock agents and their action groups, collaborators, or aliases via the Bedrock Agent control plane. Adversaries with access to an AWS account can tamper with an existing, trusted agent by altering its instructions (UpdateAgent), adding or changing action groups that wire the agent to Lambda functions or APIs (CreateAgentActionGroup, UpdateAgentActionGroup), attaching or modifying collaborators (AssociateAgentCollaborator, UpdateAgentCollaborator), or repointing an alias to a tampered version (CreateAgentAlias, UpdateAgentAlias). A PrepareAgent call is required to make a tampered configuration live. By implanting malicious behavior into an agent that legitimate users continue to invoke, an attacker can maintain durable access through a trusted component. Creation of brand-new agents (CreateAgent) is intentionally excluded as lower-signal activity.

Read More

Identifies an Amazon Bedrock AgentCore execution role (an AssumedRole identity whose role name begins with "AgentCore-" or contains "BedrockAgentCore") making an AWS API call to a service it has not previously called. AgentCore runtimes normally interact only with Bedrock inference, AgentCore data-plane, and observability services (CloudWatch Logs, X-Ray, CloudWatch metrics), so an execution role suddenly calling STS, EC2, IAM, Secrets Manager, or other services is a strong indicator that the role's temporary credentials were exfiltrated from the agent's microVM (for example, via the Code Interpreter instance-metadata-service credential theft) and are being used outside the runtime for reconnaissance, privilege escalation, or lateral movement. Because the stolen credentials are recorded in CloudTrail under the execution role's own identity, the anomalous service usage, not the identity, is the detectable signal.

Read More

Detects control-plane mutations to AWS Bedrock knowledge bases and their backing RAG data sources via CloudTrail. An adversary with access to Bedrock Agent APIs can poison the corpus that RAG-enabled models treat as authoritative by ingesting attacker-controlled documents (IngestKnowledgeBaseDocuments, StartIngestionJob), deleting legitimate documents (DeleteKnowledgeBaseDocuments), or repointing/altering the data source itself (CreateDataSource, UpdateDataSource, DeleteDataSource, UpdateKnowledgeBase). Because downstream applications and users trust model answers grounded in this stored data, tampering with the corpus is a stored data manipulation that can drive misinformation, fraud, or manipulated decisions at inference time. This is a New Terms rule that looks for the first time a given identity ARN performs one of these knowledge base or data source mutations within the history window.

Read More

Detects modification or deletion of resource-based access policies on AWS Bedrock resources via the PutResourcePolicy and DeleteResourcePolicy API calls. Resource-based policies govern which principals (including external accounts) may access Bedrock resources such as agents, knowledge bases, and custom models. An adversary may attach a resource policy granting an external or unexpected principal access to a Bedrock resource to establish persistence or enable cross-account access, or may delete an existing policy to weaken access controls. These changes should be validated for principal ownership and least-privilege intent.

Read More

Detects when an Amazon Bedrock agent is associated with, or updated to use, a knowledge base via the AssociateAgentKnowledgeBase, or UpdateAgentKnowledgeBase API actions. Bedrock agents consume knowledge base (RAG) content as trusted context for the model. By wiring an agent to an externally controlled or third-party knowledge base, or by swapping in an attacker-controlled knowledge base, an adversary can redraw the agent's trust boundary toward an untrusted source. This is a software-supply-chain compromise and an indirect prompt-injection delivery vector: poisoned or adversarial content served from the associated knowledge base is treated as authoritative by the agent. Validate that the associated knowledge base, and any underlying data source, is owned and controlled by your organization.

Read More

Detects the use of the AWS CLI with the "--endpoint-url" argument, which allows users to specify a custom endpoint URL for AWS services. This can be leveraged by adversaries to redirect API requests to non-standard or malicious endpoints, potentially bypassing typical security controls and logging mechanisms. This behavior may indicate an attempt to interact with unauthorized or compromised infrastructure, exfiltrate data, or perform other malicious activities under the guise of legitimate AWS operations.

Read More

Identifies AWS CloudTrail data events where an unauthenticated identity successfully retrieves temporary AWS credentials from a Cognito Identity Pool via GetCredentialsForIdentity. Cognito Identity Pools can be configured to allow unauthenticated (guest) access, intended for scenarios like anonymous app analytics, but a pool that grants those anonymous identities meaningful IAM permissions becomes a public, unauthenticated path to real AWS credentials. Adversaries who discover an identity pool ID (often embedded in mobile app binaries, web app JavaScript, or public source repositories) can call GetId followed by GetCredentialsForIdentity with no login token at all to obtain temporary credentials, then use them to access whatever the pool's unauthenticated role permits. This is a New Terms rule that limits alerting to identity pools that have not been observed issuing credentials to an unauthenticated caller before, since some applications intentionally and continuously use guest access as part of normal operation.

Read More

Flags the first time a given IAM principal invokes a narrow set of high-signal discovery APIs (credential check, account and IAM enumeration, bucket and compute inventory, logging introspection) from a source IP whose autonomous system number (ASN) matches a curated set commonly associated with consumer VPN brands, VPN-heavy hosting, and provider networks referenced in public reporting on TeamPCP activity (for example 31173 Services AB AS39351 and Oy Crea Nova Hosting Solution Ltd). Broad List*/Describe* patterns are intentionally omitted to reduce noise. Hosting ASNs are heavily dual-use; validate source.as.number in your data and extend event.action only when your baseline allows it.

Read More

Identifies when an AWS DynamoDB table is scanned by a user who does not typically perform this action. Adversaries may use the Scan operation to collect sensitive information or exfiltrate data from DynamoDB tables. This rule detects unusual user activity by monitoring for the Scan action in CloudTrail logs. This is a New Terms rule that only flags when this behavior is observed by a user or role for the first time.

Read More

Identifies when an AWS DynamoDB table is exported to S3. Adversaries may use the ExportTableToPointInTime operation to collect sensitive information or exfiltrate data from DynamoDB tables. This rule detects unusual user activity by monitoring for the ExportTableToPointInTime action in CloudTrail logs. This is a New Terms rule that only flags when this behavior is observed by a user or role for the first time.

Read More

Identifies the first time a given IAM principal successfully creates an EC2 key pair when the request is sourced from a network whose autonomous system organization is not attributed to common cloud or hyperscaler providers in your GeoIP data. Adversaries may call CreateKeyPair to stage SSH access material before launching or accessing instances. A new terms baseline on user_identity.arn suppresses repeated noise from the same principal while still surfacing the initial suspicious creation from an unusual egress label.

Read More

Detects a principal account creating or replacing - or attempts to create or replace - an AWS Network Access Control List (NACL) entry using protocol -1 (all traffic). Both successful and failed outcomes are included. A NACL entry with protocol -1 passes all traffic regardless of port, which would disable network-layer controls for the affected subnets. Monitoring for new identities performing this change helps surface freshly compromised credentials or unauthorized principals removing a defense-in-depth layer to facilitate lateral movement or data exfiltration. This signal only flags if this behavior was not observed historically in a specific time window.

Read More

Identifies the first time an EC2 instance role session calls AWS STS GetCallerIdentity from a given source autonomous system (AS) organization name within the lookback window. Adversaries who steal instance role credentials often verify them with GetCallerIdentity from infrastructure outside your normal egress paths. Baseline learning on the pairing of identity and source network reduces noise from stable NAT or AWS-classified egress compared to alerting on every call from a non-Amazon ASN.

Read More

Identifies when an EC2 Route Table has been created. Route tables can be used by attackers to disrupt network traffic, reroute communications, or maintain persistence in a compromised environment. This is a New Terms rule that detects the first instance of this behavior by a user or role.

Read More

Identifies AWS CloudTrail events where an EC2 route table or association has been modified or deleted. Route table or association modifications can be used by attackers to disrupt network traffic, reroute communications, or maintain persistence in a compromised environment. This is a New Terms rule that detects the first instance of this behavior by a user or role.

Read More

Identifies the first occurrence of an unauthorized attempt by an AWS role to use GetPassword to access the administrator password of an EC2 instance. Adversaries may use this API call to escalate privileges or move laterally within EC2 instances.

Read More

Identifies discovery request DescribeInstanceAttribute with the attribute userData and instanceId in AWS CloudTrail logs. This may indicate an attempt to retrieve user data from an EC2 instance. Adversaries may use this information to gather sensitive data from the instance such as hardcoded credentials or to identify potential vulnerabilities. This is a New Terms rule that identifies the first time an IAM user or role requests the user data for a specific EC2 instance.

Read More

Identifies the first occurrence of an AWS Security Token Service (STS) GetFederationToken request made by a user. The GetFederationToken API call allows users to request temporary security credentials to access AWS resources. The maximum expiration period for these tokens is 36 hours and they can be used to create a console signin token even for identities that don't already have one. Adversaries may use this API to obtain temporary credentials for persistence and to bypass IAM API call limitations by gaining console access.

Read More

Detects sensitive AWS IAM API operations executed using temporary session credentials (access key IDs beginning with "ASIA"). Temporary credentials are commonly issued through sts:GetSessionToken, sts:AssumeRole, or AWS SSO logins and are meant for short-term use. It is unusual for legitimate users or automated processes to perform privileged IAM actions (e.g., creating users, updating policies, or enabling/disabling MFA) with session tokens. This behavior may indicate credential theft, session hijacking, or the abuse of a privileged role’s temporary credentials.

Read More

Detects the creation of an AWS Identity and Access Management (IAM) user initiated by an assumed role on an EC2 instance. Assumed roles allow users or services to temporarily adopt different AWS permissions, but the creation of IAM users through these roles, particularly from within EC2 instances, may indicate a compromised instance. Adversaries might exploit such permissions to establish persistence by creating new IAM users under unauthorized conditions.

Read More

Detects when an AWS Identity and Access Management (IAM) customer-managed policy is attached to a role by an unusual or unauthorized user. Customer-managed policies are policies created and controlled within an AWS account, granting specific permissions to roles or users when attached. This rule identifies potential privilege escalation by flagging cases where a customer-managed policy is attached to a role by an unexpected actor, which could signal unauthorized access or misuse. Attackers may attach policies to roles to expand permissions and elevate their privileges within the AWS environment. This is a New Terms rule that uses the "cloud.account.id", "user.name" and "entity.target.id" fields to check if the combination of the actor identity and target role name has not been seen before.

Read More

Identifies the first time, within the configured history window, that a long-term IAM access key ID (prefix AKIA) is used successfully from a given source.ip in AWS CloudTrail. Long-term access keys belong to IAM users or the account root user. They are a common target after credential theft or leakage, including supply-chain and exposed-key scenarios. Temporary security credentials (prefix ASIA) and console sessions are excluded so the signal emphasizes programmatic access patterns.

Read More

Detects when an uncommon user or role creates an OpenID Connect (OIDC) Identity Provider in AWS IAM. OIDC providers enable web identity federation, allowing users authenticated by external identity providers (such as Google, GitHub, or custom OIDC-compliant providers) to assume IAM roles and access AWS resources. Adversaries who have gained administrative access may create rogue OIDC providers to establish persistent, federated access that survives credential rotation. This technique allows attackers to assume roles using tokens from an IdP they control. While OIDC provider creation is benign in some environments, it should still be validated against authorized infrastructure changes.

Read More

Detects the first time an AWS identity successfully deletes an IAM managed policy whose ARN contains guardrail-related keywords (for example Boundary, Deny, Restrict, Guard, SCP, Guardrail). Adversaries who have obtained elevated IAM privileges may delete policies to remove restrictive permissions boundaries, eliminate deny-based guardrails, or clean up after a privilege escalation operation. Infrastructure-as-code tools (Terraform, CloudFormation, Pulumi, and Ansible) are excluded because policy lifecycle management is a routine part of automated deployments. A policy deletion by an identity not seen performing this activity during the prior seven days may indicate newly compromised credentials being used to modify the account's permission structure.

Read More

Identifies the first observed occurrence, within the configured New Terms history window, of a regular IAM user successfully signing in to the AWS Management Console without multi-factor authentication. A password alone is a weaker control than password-plus-MFA, and an adversary who has phished, guessed, or otherwise obtained a user's password can sign in directly if MFA is not enforced for that user. This rule is scoped to standard IAM users only; it excludes the AWS root user (covered by a dedicated rule) and federated/SSO sign-ins (covered by a dedicated rule that also accounts for IdP-side MFA), since MFAUsed: No is expected in both of those cases for reasons unrelated to this gap.

Read More

Identifies the first time within the prior 14 days that a principal directly invokes an AWS Lambda function in an account, excluding invocations made on behalf of AWS services (normal event-source triggers). Adversaries who compromise credentials or move laterally may directly invoke functions to execute code, retrieve data returned by a function, or abuse an over-permissioned execution role. Direct, ad hoc invocation by a principal that does not normally call Lambda deviates from the usual event-driven invocation pattern and is worth reviewing. This rule relies on AWS Lambda data event logging, which is not enabled by default.

Read More

Identifies an AWS Lambda function invoked directly by a principal from a source network (ASN) not seen for that principal in the prior 10 days, excluding common cloud provider networks. Direct invocation from an unfamiliar external network can indicate use of stolen execution-role or user credentials from attacker-controlled infrastructure to execute functions or retrieve the data they return. This rule relies on AWS Lambda data event logging, which is not enabled by default.

Read More

Detects a principal modifying an S3 bucket ACL to grant public read or write access that has not been observed doing so within the history window, using canned ACLs such as public-read or public-read-write. ACL-based public access is a distinct API path (PutBucketAcl) that can bypass some Block Public Access controls. Monitoring for new identities performing this change helps surface freshly compromised credentials being used to stage data for exfiltration or inadvertently expose sensitive content.

Read More

Identifies AWS CloudTrail events where an unauthenticated source is attempting to access an S3 bucket. This activity may indicate a misconfigured S3 bucket policy that allows public access to the bucket, potentially exposing sensitive data to unauthorized users. Adversaries can specify --no-sign-request in the AWS CLI to retrieve objects from an S3 bucket without authentication. This is a New Terms rule, which means it will trigger for each unique combination of the source.address and targeted bucket name that has not been seen making this API request.

Read More

Detects the first time an AWS identity requests a service quota increase via the AWS Service Quotas API within a 7-day history window. Service quota increases are submitted to AWS Support and, when approved, raise the limits on EC2 instances, Lambda concurrency, VPC resources, and other services. An adversary who obtains AWS credentials may request quota increases as infrastructure preparation for large-scale cryptomining, DDoS amplification, phishing campaigns, or data exfiltration operations that require compute or network resources beyond the account's current limits.

Read More

Detects the first occurrence in 7 days of an AWS identity attaching the managed policy AmazonSESFullAccess to an IAM user, role, or group. AmazonSESFullAccess grants unrestricted permission to send email, manage identities and templates, manage suppression lists, and access SES account-level settings. Granting this policy to an unexpected IAM entity, particularly a newly created user or a role not previously associated with email operations, is a documented technique used by threat actors to establish phishing infrastructure on compromised AWS accounts, enabling them to send email on behalf of the victim organization's trusted sending domain. Using new terms on the calling identity suppresses recurring attachments by known email automation while surfacing identities performing this action for the first time.

Read More

Identifies when a user subscribes to an SNS topic using a new protocol type (ie. email, http, lambda, etc.). SNS allows users to subscribe to recieve topic messages across a broad range of protocols like email, sms, lambda functions, http endpoints, and applications. Adversaries may subscribe to an SNS topic to collect sensitive information or exfiltrate data via an external email address, cross-account AWS service or other means. This rule identifies a new protocol subscription method for a particular user.

Read More

Identifies when an SNS topic is created by a user who does not typically perform this action. Adversaries may create SNS topics to stage capabilities for data exfiltration or other malicious activities. This is a New Terms rule that only flags when this behavior is observed for the first time by a user or role.

Read More

Identifies when an SNS topic message is published by a rare user in AWS. Adversaries may publish messages to SNS topics for phishing campaigns, data exfiltration, or lateral movement within the AWS environment. SNS topics are used to send notifications and messages to subscribed endpoints such as applications, mobile devices or email addresses, making them a valuable target for adversaries to distribute malicious content or exfiltrate sensitive data. This is a New Terms rule that only flags when this behavior is observed for the first time by a user or role.

Read More

Detects the execution of commands or scripts on EC2 instances using AWS Systems Manager (SSM), such as RunShellScript, RunPowerShellScript or custom documents. While legitimate users may employ these commands for management tasks, they can also be exploited by attackers with credentials to establish persistence, install malware, or execute reverse shells for further access to compromised instances. This is a New Terms rule that looks for the first instance of this behavior by a user or role.

Read More

Identifies the use of the AWS Systems Manager (SSM) SendCommand API with the either AWS-RunShellScript or AWS-RunPowerShellScript parameters. The SendCommand API call allows users to execute commands on EC2 instances using the SSM service. Adversaries may use this technique to execute commands on EC2 instances without the need for SSH or RDP access. This behavior may indicate an adversary attempting to execute commands on an EC2 instance for malicious purposes. This is a New Terms rule that only flags when this behavior is observed for the first time on a host in the last 7 days.

Read More

Identifies when an AWS Systems Manager (SSM) command document is created by a user or role who does not typically perform this action. Adversaries may create SSM command documents to execute commands on managed instances, potentially leading to unauthorized access, command and control, data exfiltration and more.

Read More

Detects the rare occurrence of a user or role accessing AWS Systems Manager (SSM) inventory APIs or running the AWS-GatherSoftwareInventory job. These APIs reveal detailed information about managed EC2 instances including installed software, patch compliance status, and command execution history. Adversaries may use these calls to collect software inventory while blending in with legitimate AWS operations. This is a New Terms rule that detects when a user accesses these reconnaissance APIs for the first time.

Read More

Identifies the first occurrence of an AWS user or role establishing a session via SSM to an EC2 instance. Adversaries may use AWS Session Manager to establish a session to an EC2 instance to execute commands on the instance. This can be used to gain access to the instance and perform actions such as privilege escalation.

Read More

Identifies when a user has assumed a role using a new MFA device. Users can assume a role to obtain temporary credentials and access AWS resources using the AssumeRole API of AWS Security Token Service (STS). While a new MFA device is not always indicative of malicious behavior it should be verified as adversaries can use this technique for persistence and privilege escalation.

Read More

Identifies when the STS AssumeRoot action is performed by a rare user in AWS. The AssumeRoot action allows users to assume the root member account role, granting elevated but specific permissions based on the task policy specified. Adversaries who have compromised user credentials can use this technique to escalate privileges and gain unauthorized access to AWS resources. This is a New Terms rule that identifies when the STS AssumeRoot action is performed by a user that rarely assumes this role against a specific member account.

Read More

An adversary with access to a set of compromised credentials may attempt to verify that the credentials are valid and determine what account they are using. This rule looks for the first time an identity has called the STS GetCallerIdentity API, which may be an indicator of compromised credentials. A legitimate user would not need to perform this operation as they should know the account they are using.

Read More

Identifies when a service has assumed a role in AWS Security Token Service (STS). Services can assume a role to obtain temporary credentials and access AWS resources. Adversaries can use this technique for credential access and privilege escalation. This is a New Terms rule that identifies when a service assumes a role in AWS Security Token Service (STS) to obtain temporary credentials and access AWS resources. While often legitimate, adversaries may use this technique for unauthorized access, privilege escalation, or lateral movement within an AWS environment.

Read More

Identifies when a user or role has assumed a role in AWS Security Token Service (STS). Users can assume a role to obtain temporary credentials and access AWS resources. Adversaries can use this technique for credential access and privilege escalation. This is a New Terms rule that identifies when a user assumes a role in AWS Security Token Service (STS) to obtain temporary credentials and access AWS resources. While often legitimate, adversaries may use this technique for unauthorized access, privilege escalation, or lateral movement within an AWS environment.

Read More

Identifies role chaining activity. Role chaining is when you use one assumed role to assume a second role through the AWS CLI or API. While this a recognized functionality in AWS, role chaining can be abused for privilege escalation if the subsequent assumed role provides additional privileges. Role chaining can also be used as a persistence mechanism as each AssumeRole action results in a refreshed session token with a 1 hour maximum duration. This is a new terms rule that looks for the first occurance of one role (aws.cloudtrail.user_identity.session_context.session_issuer.arn) assuming another (aws.cloudtrail.resources.arn).

Read More

Detects the first occurrence of a user identity accessing AWS Systems Manager (SSM) SecureString parameters using the GetParameter or GetParameters API actions with credentials in the request parameters. This could indicate that the user is accessing sensitive information. This rule detects when a user accesses a SecureString parameter with the withDecryption parameter set to true. This is a New Terms rule that detects the first occurrence of an AWS identity accessing SecureString parameters with decryption.

Read More

Identifies the first time, in a historical window, a host runs AzCopy copy or sync to Azure Blob, Data Lake, or File storage, or starts Azure Storage Explorer. These Microsoft utilities are legitimate data-transfer tools; ransomware and cloud-ransomware operators drop portable copies and use SAS-authenticated jobs to pull data from victim storage and push it to attacker-controlled accounts.

Read More

Identifies Azure AD Graph (graph.windows.net) requests where the combination of calling OAuth client ("azure.aadgraphactivitylogs.properties.app_id") and signed-in user ("user.id") has not been observed in the tenant in a historical window. A user appearing against AAD Graph under an OAuth client that has not previously authenticated that user is a sign of a FOCI swap, a phished refresh token being redeemed for a new client, or an adversary running tooling under a client identity the user does not normally use.

Read More

Identifies Azure AD Graph (graph.windows.net) requests originating from network sources outside the major public-cloud and Microsoft ASNs that legitimate first-party callers normally come from. Adversary tooling typically rides on commodity hosting (residential ISPs, VPS providers, anonymisers) which produces an ASN distribution very different from the Microsoft / AWS / GCP / Akamai / Cloudflare ranges that dominate legitimate AAD Graph traffic.

Read More

Detects when a service principal or user performs an Azure Arc cluster credential listing operation from a source IP not previously associated with that identity. The listClusterUserCredential action retrieves credentials for the Arc Cluster Connect proxy, enabling kubectl access through the Azure ARM API. An adversary using stolen service principal credentials will typically call this operation from infrastructure not previously seen for that SP. By tracking the combination of caller identity and source IP, this rule avoids false positives from backend services and CI/CD pipelines that rotate IPs but maintain consistent identity-to-IP patterns over time.

Read More

Identifies the deletion of Azure Restore Point Collections by a user who has not previously performed this activity. Restore Point Collections contain recovery points for virtual machines, enabling point-in-time recovery capabilities. Adversaries may delete these collections to prevent recovery during ransomware attacks or to cover their tracks during malicious operations.

Read More

Identifies when an Azure disk snapshot is deleted by an unusual user in a specific resource group. Snapshots are critical for backup, disaster recovery, and forensic analysis. Adversaries may delete snapshots to prevent data recovery, eliminate forensic evidence, or disrupt backup strategies before executing ransomware or other destructive attacks. Monitoring snapshot deletions is essential for detecting potential attacks targeting backup and recovery capabilities.

Read More

Identifies synchronous command execution on a virtual machine (VM) or virtual machine scale set (VMSS) in Azure via the action-based Run Command ("runCommand/action"). A Virtual Machine Contributor role lets you manage virtual machines, but not access them, nor access the virtual network or storage account they’re connected to. However, commands can be run on the VM via the Run Command feature, which execute as System (Windows) or root (Linux). Other roles, such as certain Administrator roles, may be able to execute commands on a VM as well.

Read More

Identifies the deletion of diagnostic settings in Azure, which send platform logs and metrics to different destinations. An adversary may delete diagnostic settings in an attempt to evade defenses.

Read More

Identifies modifications to a Key Vault in Azure. The Key Vault is a service that safeguards encryption keys and secrets like certificates, connection strings, and passwords. Because this data is sensitive and business critical, access to key vaults should be secured to allow only authorized applications and users. This is a New Terms rule that detects when this activity hasn't been seen by the user in a specified time frame.

Read More

Identifies secrets, keys, or certificates retrieval operations from Azure Key Vault by a user principal that has not been seen previously doing so in a certain amount of days. Azure Key Vault is a cloud service for securely storing and accessing secrets, keys, and certificates. Unauthorized or excessive retrievals may indicate potential abuse or unauthorized access attempts.

Read More

Identifies when Azure Storage Account Blob public access is enabled, allowing external access to blob containers. This technique was observed in cloud ransom-based campaigns where threat actors modified storage accounts to expose non-remotely accessible accounts to the internet for data exfiltration. Adversaries abuse the Microsoft.Storage/storageAccounts/write operation to modify public access settings.

Read More

Identifies when an Azure Storage Account is deleted. Adversaries may delete storage accounts to disrupt operations, destroy evidence, or cause denial of service. This activity could indicate an attacker attempting to cover their tracks after data exfiltration or as part of a destructive attack. Monitoring storage account deletions is critical for detecting potential impact on business operations and data availability.

Read More

Identifies unusual high-privileged access to Azure Storage Account keys by users with Owner, Contributor, or Storage Account Contributor roles. This technique was observed in STORM-0501 ransomware campaigns where compromised identities with high-privilege Azure RBAC roles retrieved access keys to perform unauthorized operations on Storage Accounts. Microsoft recommends using Shared Access Signature (SAS) models instead of direct key access for improved security. This rule detects when a user principal with high-privilege roles accesses storage keys for the first time in 7 days.

Read More

Identifies the first time an Azure Storage resource receives an anonymous data-plane read (GetBlob and related Get or List operations). Anonymous requests are used to probe public containers and to test stolen blob URLs before a SAS is appended. First-seen resource ID keeps volume down while still covering WireServer-related probes of status or extension blobs.

Read More

Identifies successful GetBlob operations on Azure Storage Accounts using AzCopy user agent with SAS token authentication. AzCopy is a command-line utility for copying data to and from Azure Storage. While legitimate for data migration, adversaries may abuse AzCopy with compromised SAS tokens to exfiltrate data from Azure Storage Accounts. This rule detects the first occurrence of GetBlob operations from a specific storage account using this pattern.

Read More

Identifies create, read, update, or delete (CRUD) operations against Azure VM or VM scale set extensions ("MICROSOFT.COMPUTE/VIRTUALMACHINES/EXTENSIONS/*" or the scale set equivalent) where the combination of the targeted extension resource name and the source autonomous system (AS) number has not been observed recently. VM extensions such as CustomScript and DSC run with high privilege on the guest (SYSTEM on Windows, root on Linux), so writing, modifying, or removing them is a common code-execution and persistence primitive. By keying a new terms approach on the extension resource name and the source AS number, this rule surfaces extension operations originating from networks that have not historically managed that extension, while routine first-party Microsoft automation (which originates from well-known Microsoft AS numbers) is excluded.

Read More

Identifies the creation or update of a managed Azure Run Command resource ("MICROSOFT.COMPUTE/VIRTUALMACHINES/RUNCOMMANDS/WRITE" or the virtual machine scale set equivalent) by an identity that has not performed this operation recently. Unlike the action-based Run Command ("runCommand/action"), the managed Run Command is a persistent resource on the VM whose creation or update executes the supplied script as System (Windows) or root (Linux). Because creating a managed run command both executes code and leaves a durable object, adversaries can use it as an alternative to the action invocation to evade detections that only watch "runCommand/action". Alerting on the first time a given principal performs this operation surfaces unusual or unauthorized use while suppressing routine automation that repeatedly manages the same run commands.

Read More

Identifies a connection to the Azure Serial Console of a virtual machine (VM) by an identity and source network combination that has not been observed recently. The Serial Console provides text-based console access to a VM through the boot diagnostics serial port, independent of the VM's network state. Because it does not traverse the VM's network interface, a Serial Console session bypasses Network Security Groups (NSGs), Just-in-Time (JIT) access policies, and other network controls. An adversary with a privileged Azure RBAC role (for example Virtual Machine Contributor) and boot diagnostics enabled on the target can use the Serial Console to obtain an interactive session as SYSTEM (Windows) or root (Linux).

Read More

Identifies shells, LOLBins, GTFOBins, and scripting runtimes connecting to the Azure WireServer / HostGAPlugin address 168.63.129.16 on ports 80 or 32526. The guest agent uses this fabric endpoint for GoalState, certificates, and vmSettings. Adversaries with code execution on an Azure VM (including via Run Command) use curl, PowerShell, openssl, bun, or similar tools to enumerate versions, pull transport certificates, and read HostGAPlugin /vmSettings. Azure guest-agent binaries and system python used by waagent are excluded. Descendants of the guest agent are not excluded: Run Command payloads execute in that tree.

Read More

Detects modifications to the msDS-ManagedAccountPrecededByLink attribute of a delegated managed service account by an unusual subject account. Attackers can abuse this attribute to inherit a target account's permissions and further elevate privileges.

Read More

Identifies successful outbound TLS sessions that negotiate deprecated protocol versions (SSLv3, TLS 1.0, or TLS 1.1) or weak cipher suites such as RC4, 3DES, NULL, EXPORT, or anonymous Diffie-Hellman. Adversaries-in-the-middle and legacy malware often force these negotiations to decrypt or intercept traffic. Modern clients and services should negotiate TLS 1.2 or 1.3 with strong ciphers on internet-bound connections.

Read More

This is a New Terms rule that identifies the first occurrence of setsid or nohup being used to directly execute a process on a host. Attackers may leverage these tools to execute commands in a new session and/or to ignore signals.

Read More

Detects creation of a delegated Managed Service Account by an unusual subject account. Attackers can abuse weak child-object or msDS-DelegatedManagedServiceAccount rights during account migration to elevate privileges.

Read More

This rule detects the installation of a Debian package (dpkg) by an unusual parent process. The dpkg command is used to install, remove, and manage Debian packages on a Linux system. Attackers can abuse the dpkg command to install malicious packages on a system.

Read More

Identifies the first observed instance of a Microsoft first-party public client application acquiring a Microsoft Graph token using single-factor (password-only) authentication while an MFA Conditional Access grant control went unenforced, for a given user, application, and source autonomous system (ASN). This pattern is associated with the Conditional Access "resource exclusion" bypass: when a tenant's "all resources" Conditional Access policy contains at least one application exclusion, Entra ID issues tokens for low-privilege baseline scopes (User.Read, openid, profile, email) to any resource, including Microsoft Graph, without enforcing the policy's grant controls (such as MFA). An adversary holding only a stolen password can therefore obtain a Graph token through a trusted first-party public client (for example, Microsoft Bing Search) and enumerate directory objects, even though the tenant requires MFA. Critically, the overall conditional_access_status is never "failure" for this technique (the sign-in is not blocked); it is reported as "success" or "notApplied" depending on what other policies exist in the tenant, so detections that key on Conditional Access failures will not observe it. The reliable fingerprint is in the per-policy results: a policy whose enforced grant control is MFA reports a result of "notApplied" for this sign-in, meaning the MFA requirement was silently not enforced while the single-factor, password-only sign-in still succeeded.

Read More

Identifies a modification to a conditional access policy (CAP) in Microsoft Entra ID. Adversaries may modify existing CAPs to loosen access controls and maintain persistence in the environment with a compromised identity or entity.

Read More

Identifies the first occurrence of a Microsoft Entra ID device, surfaced through the Entra ID Entity Analytics device inventory, whose host name follows the default "DESKTOP-" pattern and whose operating system build is 10.0.19041.928. This combination is the default device profile that ROADtools (roadtx) uses when registering a device, and the OS build typically differs from the patched OS versions of legitimate hosts in the environment. Adversaries register rogue devices in Entra ID to acquire a Primary Refresh Token (PRT), establish persistence, and obtain trusted, programmatic access to the tenant. Because the OS build is a tool default, this is a high-fidelity but evadable indicator; baseline approved device builds and naming conventions before relying on it.

Read More

Detects a first-party FOCI tooling client (Azure CLI, PowerShell, VS Code, Graph CLI, Azure AD PowerShell, or Visual Studio) redeeming a device-bound Primary Refresh Token (PRT) for Microsoft Graph, SharePoint/OneDrive, or Exchange Online from a source IP that has not been seen with that deviceid. Replay events are limited to compliant or Intune-managed devices. Adversaries who steal a WAM PRT SSO cookie replay it off-box; the token keeps the workstation deviceid, so this pair is new even when Windows Sign-In for that device is outside a correlation window.

Read More

Identifies when a user has elevated their access to User Access Administrator for their Azure Resources. The User Access Administrator role allows users to manage user access to Azure resources, including the ability to assign roles and permissions. Adversaries may target an Entra ID Global Administrator or other privileged role to elevate their access to User Access Administrator, which can lead to further privilege escalation and unauthorized access to sensitive resources. This is a New Terms rule that only signals if the user principal name has not been seen doing this activity in the last 14 days.

Read More

Identifies when an external authentication method (EAM) is added or modified in Entra ID. EAM may allow adversaries to bypass multi-factor authentication (MFA) requirements, potentially leading to unauthorized access to user accounts and sensitive resources by using bring-your-own IdP (BYOIDP) methods.

Read More

Identifies the first occurrence of an OAuth 2.0 authorization code grant flow for a specific combination of client application, target resource, and user principal in Microsoft Entra ID. Developer tools like Azure CLI, Visual Studio Code, and Azure PowerShell accessing Microsoft Graph or legacy Azure AD are flagged for infrequent or first time usage by a user. Additionally, any FOCI (Family of Client IDs) application accessing the deprecated Windows Azure Active Directory for the first time is flagged since this resource is rarely accessed legitimately. This pattern is indicative of OAuth phishing attacks like ConsentFix, where attackers steal authorization codes and exchange them for tokens from attacker controlled infrastructure.

Read More

Identifies when a user is observed for the first time authenticating using the device code authentication workflow. This authentication workflow can be abused by attackers to phish users and steal access tokens to impersonate the victim. By its very nature, device code should only be used when logging in to devices without keyboards, where it is difficult to enter emails and passwords. This rule only applies to Entra ID user types and detects new users leveraging this flow.

Read More

Detects unusual resource owner password credential (ROPC) login attempts by a user principal in Microsoft Entra ID. ROPC is a legacy authentication flow that allows applications to obtain tokens by directly providing user credentials. This method is less secure and can be exploited by adversaries to gain access to user accounts without requiring multi-factor authentication (MFA), especially during enumeration or password spraying. This is a New Terms rule that identifies when user principals are involved in ROPC login attempts, not seen before in the last 10 days, indicating potential abuse or unusual activity.

Read More

Identifies rare occurrences of OAuth workflow for a user principal that is single factor authenticated, with an OAuth scope containing user_impersonation for a token issued by Entra ID. Adversaries may use this scope to gain unauthorized access to user accounts, particularly when the sign-in session status is unbound, indicating that the session is not associated with a specific device or session. This behavior is indicative of potential account compromise or unauthorized access attempts. This rule flags when this pattern is detected for a user principal that has not been seen in the last 10 days, indicating potential abuse or unusual activity.

Read More

Identifies the first occurrence of a Microsoft Entra ID device, surfaced through the Entra ID Entity Analytics device inventory, whose host name follows the default "DESKTOP-" pattern and whose operating system build is "10.0.19045.2006". This is the frozen default device profile observed when adversary-in-the-middle (AiTM) phishing kits such as Tycoon2FA and Kali365 register Azure AD-joined devices after capturing a victim session, in order to acquire a Primary Refresh Token (PRT) and establish persistence. The build is hardcoded by the tooling and differs from legitimate hosts: a patched Windows 10 22H2 device reports a far higher "10.0.19045." value, so a device frozen at ".2006" with a default name is a high-fidelity, though evadable, indicator.

Read More

Identifies potential OAuth client ID spoofing in Microsoft Entra ID sign-in logs. Adversaries submit fabricated or non-existent application identifiers (client IDs) in Resource Owner Password Credentials (ROPC) authentication requests to the token endpoint. Because Entra ID validates the submitted credential before rejecting the request on application resolution, the resulting error code acts as a credential- and account-validity oracle while never producing a successful sign-in. By fragmenting attempts across many fictional application identifiers, adversaries evade per-application detections, rate limiting, and application-scoped Conditional Access. This rule detects a ROPC authentication that fails with error code 700016 (application not found in the directory) where no application display name resolves, which is characteristic of a spoofed client ID used for stealthy user enumeration and password spraying.

Read More

Identifies when new Service Principal credentials have been added in Microsoft Entra ID. In most organizations, credentials will be added to service principals infrequently. Hijacking an application (by adding a rogue secret or certificate) with granted permissions will allow the attacker to access data that is normally protected by MFA requirements.

Read More

Identifies when a service principal authenticates using a federated identity credential for the first time in the historical window. This indicates that Entra ID validated a JWT token potentially against an external OIDC identity provider and issued an access token. While legitimate for CI/CD workflows (GitHub Actions, Azure DevOps), adversaries may abuse this by configuring rogue identity providers (BYOIDP) to authenticate as compromised applications. First-time federated credential usage for a service principal warrants investigation to determine if the external identity provider is legitimate.

Read More

Identifies Entra ID service principal sign-ins where the workload identity and source autonomous system number (ASN) together have not appeared in recent history. Attackers who obtain application secrets or tokens often authenticate from unfamiliar hosting providers, residential or VPN egress, or networks outside normal automation footprints, which can precede data access, lateral movement, or ransomware activity in the tenant. The detection emphasizes first-seen network context for non-interactive workload identities.

Read More

Identifies when an application accesses SharePoint Online or OneDrive for Business for the first time in the tenant within a specified timeframe. This detects successful OAuth phishing campaigns, illicit consent grants, or compromised third-party applications gaining initial access to file storage. Adversaries often use malicious OAuth applications or phishing techniques to gain consent from users, allowing persistent access to organizational data repositories without traditional credential theft.

Read More

Detects a successful sign-in by a Member user principal through a legacy authentication client (such as Authenticated SMTP, IMAP4, POP3, Exchange ActiveSync, Exchange Web Services, or other basic-authentication clients) in Microsoft Entra ID, where the user principal has not been seen using a legacy client in the last 7 days. Legacy authentication clients rely on basic authentication, do not support modern authentication or interactive multi-factor authentication, and are frequently abused by adversaries for password spraying and account takeover because they translate into single-factor Resource Owner Password Credentials (ROPC) grants. This is a New Terms rule that surfaces the first occurrence of legacy client authentication for a given user, which is unusual in most modern environments.

Read More

Identifies rare instances of authentication methods for Microsoft Entra ID principal users. An adversary with stolen credentials may attempt to authenticate with an unusual method, which may indicate an attempt to bypass conditional access policies (CAP) and multi-factor authentication (MFA) requirements. The authentication method may not be commonly used by the user based on their historical sign-in activity.

Read More

Detects rare non-interactive sign-ins where an Entra ID client application authenticates on behalf of a principal user using an application (client) ID that is not commonly associated with that user's historical sign-in behavior. Adversaries with stolen credentials or OAuth tokens may abuse Entra ID–managed or first-party client IDs to perform on-behalf-of (OBO) authentication, blending into legitimate cloud traffic while avoiding traditional interactive sign-in flows. This technique is commonly observed in OAuth phishing, token theft, and access broker operations, and may precede lateral movement, persistence, or data access via Microsoft Graph or other cloud resources. The rule uses a New Terms approach to identify first-seen combinations of the UPN and Client ID within a defined history window, helping surface unexpected client usage that may indicate compromised identities, malicious automation, or unauthorized application impersonation.

Read More

Identifies when a Microsoft Entra ID user signs in from a device that is not typically used by the user and is not managed, which may indicate potential compromise or unauthorized access attempts. This rule detects unusual sign-in activity by comparing the device used for the sign-in against the user's typical device usage patterns. Adversaries may create and register a new device to obtain a Primary Refresh Token (PRT) and maintain persistent access.

Read More

Identifies the first-seen registration of a Windows Hello for Business (WHfB) credential for a Microsoft Entra ID user from a given source ASN in a tenant, based on a prefixed historic window. Enrollment is commonly part of legitimate onboarding or passwordless rollout and is not inherently malicious. Adversaries who have obtained a token that satisfies fresh (NGC) multi-factor authentication, for example by borrowing an existing WHfB key or passkey, can also enroll their own WHfB credential to establish durable, phishing-resistant persistence that survives password resets and standard session revocation. Correlate first-seen enrollments with the sign-in and device state that preceded them.

Read More

Identifies a Microsoft Entra ID sign-in that is satisfied by a phishing-resistant, device-bound credential (Windows Hello for Business, FIDO2 security key, or passkey) while carrying no device identifier. Windows Hello for Business (WHfB) and passkey credentials are bound to a device's TPM, so a genuine sign-in with one of these methods is normally accompanied by the registered device it lives on. A WHfB or passkey assertion that authenticates with an empty device_detail.device_id indicates the underlying key material is being used away from its bound device, for example by an adversary who extracted the key (or signed an assertion with it) and replayed it from attacker infrastructure to mint device-agnostic tokens. This is the core primitive of the "borrowing Windows Hello keys" technique and is a strong precursor to attacker device registration and Primary Refresh Token (PRT) issuance. Cross-tenant (B2B) sign-ins are excluded because they are a common benign source of empty device identifiers.

Read More

Identifies instances of an unusual process enumerating built-in Windows privileged local groups membership like Administrators or Remote Desktop users.

Read More

Identifies execution via MSSQL xp_cmdshell stored procedure. Malicious users may attempt to elevate their privileges by using xp_cmdshell, which is disabled by default, thus, it's important to review the context of it's use.

Read More

This rule detects the creation of files in the /var/log/ directory via process executables located in world-writeable locations or via hidden processes. Attackers may attempt to hide their activities by creating files in the /var/log/ directory, which is commonly used for logging system events.

Read More

This rule detects the creation of files in world-writable directories by an unusual process. Attackers may attempt to hide their activities by creating files in world-writable directories, which are commonly used for temporary file storage. This behavior is often associated with lateral movement and can be an indicator of an attacker attempting to move laterally within a network.

Read More

Identifies file permission modifications in common writable directories by a non-root user. Adversaries often drop files or payloads into a writable directory and change permissions prior to execution.

Read More

Identifies the first occurrence of an Okta user session started via a proxy.

Read More

Identifies a successful SonicWall VPN- or WAN-zone administrator or remote-user login from a source IP that was not previously observed with the same user on the same appliance during the prior 14 days. This may indicate stolen credentials, compromised administrator access, or unauthorized remote access.

Read More

This rule detects the first time a principal calls AWS CloudFormation CreateStack, CreateStackSet or CreateStackInstances API. CloudFormation is used to create a collection of cloud resources called a stack, via a defined template file. An attacker with the appropriate privileges could leverage CloudFormation to create specific resources needed to further exploit the environment. This is a new terms rule that looks for the first instance of this behavior for a role or IAM user within a particular account.

Read More

Detects the first time a Python process accesses sensitive credential files on a given host. This behavior may indicate post-exploitation credential theft via a malicious Python script, compromised dependency, or malicious model file deserialization. Legitimate Python processes do not typically access credential files such as SSH keys, AWS credentials, browser cookies, Kerberos tickets, or keychain databases, so a first occurrence is a strong indicator of compromise.

Read More

Detects the first time a Python process creates or modifies a LaunchAgent or LaunchDaemon plist file on a given host. Malicious Python scripts, compromised dependencies, or model file deserialization can establish persistence on macOS by writing plist files to LaunchAgent or LaunchDaemon directories. Legitimate Python processes do not typically create persistence mechanisms, so a first occurrence is a strong indicator of compromise.

Read More

Detects the first time a Python process spawns a shell on a given host. Malicious Python scripts, compromised dependencies, or model file deserialization can result in shell spawns that would not occur during normal workflows. Since legitimate Python processes rarely shell out to interactive shells, a first occurrence of this behavior on a host is a strong signal of potential compromise.

Read More

This rule identifies when a User Account starts the Active Directory Replication Process for the first time. Attackers can use the DCSync technique to get credential information of individual accounts or the entire domain, thus compromising the entire domain.

Read More

An adversary with access to a compromised AWS service such as an EC2 instance, Lambda function, or other service may attempt to leverage the compromised service to access secrets in AWS Secrets Manager. This rule looks for the first time a specific user identity has programmatically retrieved a secret value from Secrets Manager using the GetSecretValue action. This rule assumes that AWS services such as Lambda functions and EC2 instances are setup with IAM role's assigned that have the necessary permissions to access the secrets in Secrets Manager. An adversary with access to a compromised AWS service would rely on its' attached role to access the secrets in Secrets Manager.

Read More

Identifies the load of a driver with an original file name and signature values that were observed for the first time during the last 30 days. This rule type can help baseline drivers installation within your environment.

Read More

Detects the first time a user authorizes a third-party Google OAuth application that requests identity or sign-in scopes. Adversaries may abuse compromised credentials or phishing-linked consent flows to register novel OAuth clients, obtain refresh tokens, and authenticate as valid users while evading password-only detections.

Read More

Identifies the first successful or no-reply Memcached store command from a client to a server. Memcached commonly has no authentication, so an unauthorized writer can overwrite session tokens, poison cached application content, or alter security-sensitive state. This behavior can enable session hijacking such as the exposure described by CVE-2026-29093.

Read More

Identifies a new credentials logon type performed by an unusual process. This may indicate the existence of an access token forging capability that are often abused to bypass access control restrictions.

Read More

Identifies the first source and destination IP pair observed in a five-day history window where an NFS client asserts AUTH_SYS (RPC UNIX) credentials with UID 0 (root). NFSv3 and NFSv4 clients can claim arbitrary UIDs through AUTH_SYS, and weak export controls may honor root-equivalent access from unexpected hosts. This is a common precursor to unauthorized mounts, sensitive file reads, and remote encryption of exported shares.

Read More

Adversaries may install legitimate remote monitoring and management (RMM) tools or remote access software on compromised endpoints for command-and-control (C2), persistence, and execution of native commands. This rule detects when a process is started whose name or code signature matches commonly abused RMM or remote access tools. New Terms type: the host.id and process.name pair has not been seen before within the configured 7-day history window.

Read More

Identifies newly seen removable devices by device friendly name using registry modification events. While this activity is not inherently malicious, analysts can use those events to aid monitoring for data exfiltration over those devices.

Read More

Identifies a newly observed RMM-related code-signature subject across the Windows Elastic Defend hosts. Attackers often use RMM tools to gain remote access to victim machines and deploy malware.

Read More

Identifies the first client IP observed issuing MongoDB commands that can drop databases, collections, indexes, users, or roles within a five-day history window. Adversaries with access to an exposed or compromised MongoDB service may use these commands to destroy data, disrupt applications, or prepare a wipe-and-extort attack.

Read More

This rule detects FortiGate administrator account creation from a source IP address not previously seen performing admin operations on the device. Threat actors exploiting CVE-2026-24858 (FG-IR-26-060) authenticate via FortiCloud SSO bypass and immediately create local administrator accounts for persistence, typically from infrastructure not associated with normal administrative activity.

Read More

Identifies when a service account impersonation role is granted on a Google Cloud Platform (GCP) service account via a SetIamPolicy operation. Roles such as "roles/iam.serviceAccountTokenCreator", "roles/iam.serviceAccountUser", and "roles/iam.serviceAccountOpenIdTokenCreator" allow a principal to mint access or identity tokens for the target service account, or to act as it when deploying resources. Adversaries who have obtained sufficient privileges may grant themselves or an attacker-controlled principal one of these roles to impersonate a higher-privileged service account, escalating privileges and establishing durable, key-less persistence that survives credential rotation. This is a New Terms rule that alerts when the granting principal has not been observed performing this action in the last weeks.

Read More

Detects GenAI tools connecting to unusual domains on macOS. Adversaries may compromise GenAI tools through prompt injection, malicious MCP servers, or poisoned plugins to establish C2 channels or exfiltrate sensitive data to attacker-controlled infrastructure. AI agents with network access can be manipulated to beacon to external servers, download malicious payloads, or transmit harvested credentials and documents.

Read More

Detects when the github-actions[bot] pushes code to a repository where it has not performed this behavior before in a certain time window. This may indicate a supply chain attack where malicious code running in a CI workflow attempts to modify repository contents, such as injecting backdoor workflow files.

Read More

Detects when there is activity on a private GitHub repository from an unusual IP address. Adversaries may access private repositories from unfamiliar IPs to exfiltrate sensitive code or data, potentially indicating a compromise or unauthorized access.

Read More

Detects successful GKE API requests from unauthenticated anonymous identities using an unusual user agent. Attackers may rely on anonymous access for initial cluster access or to avoid attribution. Matches "system:anonymous" / "system:unauthenticated" and GKE audit rows where the principal is missing (common for unauthenticated clients). Common kube-probe health checks (readyz/livez/healthz/version) are excluded.

Read More

Detects the first occurrence of a failed GKE API request from a previously unseen user agent. Adversary tooling often uses non-standard clients; combined with authorization failures this can indicate RBAC probing or exploitation attempts.

Read More

Detects GKE secrets get or list requests from a previously unseen combination of source IP, identity, and user agent, excluding the default Kubernetes client placeholder. Attackers who compromise a pod or steal a kubeconfig often use curl, custom scripts, or atypical clients from a new host to read service-account tokens, registry credentials, or application secrets. Anonymous identities are excluded; use dedicated anonymous-access rules for unauthenticated probing.

Read More

Detects the first time a human GKE caller lists secrets cluster-wide or in default or kube-system from a source autonomous system that is not attributed to common cloud provider organizations. This can indicate remote secret enumeration using stolen credentials from an unusual network.

Read More

Detects the first occurrence of create or patch activity against sensitive GKE workloads (DaemonSets, Deployments, or CronJobs) from an unusual combination of user agent, source IP, and user identity, which may indicate privilege escalation or unauthorized access within the cluster.

Read More

Detects the first successful GKE secrets.get by a pod service account from a previously unseen combination of service-account identity, user agent, and source IP. Controllers routinely read secrets with a stable client fingerprint; a new user agent or source for that service account could indicate a stolen token used outside the workload (for example curl, a custom script, or kubectl from an unexpected host).

Read More

Detects the first occurrence of a non-system GKE identity establishing an exec session into a pod. kubectl exec enables interactive command execution inside workloads and is a common post-compromise technique to access secrets and expand access.

Read More

Detects the first time a Google Workspace user successfully signs in from a given source ASN within a 14-day historical window. Most users have a stable set of egress ASNs (home ISP, corporate VPN, mobile carrier). A new ASN for a user is a meaningful anomaly as it surfaces ISP changes and travel, but also catches AiTM phishing-kit relays whose egress ASN was never previously associated with the user.

Read More

Detects the first time a Google Workspace user is observed authenticating from a device of a given type (e.g., WINDOWS, MAC, ANDROID, IOS, LINUX) within a historical window. Note that "DEVICE_REGISTER_UNREGISTER_EVENT" events do not represent one-time physical device enrollments; the Google Reports API emits a fresh "google_workspace.device.id" on each event, and the same physical device may produce multiple events per day as sessions/sync renewals occur. The rule therefore surfaces a user authenticating from a new device type, not a new physical device. This is still high-fidelity because adversaries who compromise a Workspace identity via AiTM kits or stolen OAuth refresh tokens frequently relay sessions from device types that diverge from the legitimate user's baseline (e.g., a WINDOWS session appearing for a known macOS user, or simultaneous WINDOWS+MAC sessions within minutes), which is the canonical kit fingerprint. Because the underlying token retains access after password rotation, treat unexpected device-type divergence as a compromise indicator and revoke tokens, not just credentials.

Read More

This rule detects when an interactive shell process is launched via an unusual parent processes inside a container. Interactive processes are typically run in the foreground and require user input, which is unusual behavior for a containerized environment. This activity could indicate an attacker attempting to gain access to the container environment or perform malicious actions.

Read More

This rule detects the creation of a Linux kernel object file (.ko) on a system. Threat actors may leverage Linux kernel object files to load a rootkit or other type of malware on a system providing them with complete control and the ability to hide from security products.

Read More

This rule detects the execution of kill, pkill, and killall commands on Linux systems. These commands are used to terminate processes on a system. Attackers may use these commands to kill security tools or other processes to evade detection or disrupt system operations.

Read More

This rule detects when an unauthenticated user request is authorized within the cluster via an unusual user agent. Attackers may attempt to use anonymous accounts to gain initial access to the cluster or to avoid attribution of their activities within the cluster. This rule excludes the /healthz, /livez, /version and /.well-known/oauth-authorization-server endpoints which are commonly accessed anonymously.

Read More

This rule detects when a forbidden request is made from an unusual user agent in a Kubernetes environment. Adversary tooling may use non-standard or unexpected user agents to interact with the Kubernetes API, which can indicate an attempt to evade detection or blend in with legitimate traffic. In combination with a forbidden request, this behavior can suggest an adversary is attempting to exploit vulnerabilities or misconfigurations in the Kubernetes cluster.

Read More

This rule detects when secrets are accessed via an unusual user agent, user name and source IP. Attackers may attempt to access secrets in a Kubernetes cluster to gain access to sensitive information after gaining access to the cluster.

Read More

This rule detects when a service account or node attempts to enumerate their own permissions via the selfsubjectaccessreview or selfsubjectrulesreview APIs via an unusual user agent. This is highly unusual behavior for non-human identities like service accounts and nodes. An adversary may have gained access to credentials/tokens and this could be an attempt to determine what privileges they have to facilitate further movement or execution within the cluster.

Read More

This rule monitors for the usage of the most common audio recording utilities on unix systems by an uncommon process parent. Adversaries may collect audio data from users or systems for a variety of reasons including espionage, credential theft, or reconnaissance.

Read More

This rule monitors for the usage of the most common clipboard utilities on unix systems by an uncommon process parent. Adversaries may collect data stored in the clipboard from users copying information within or between applications.

Read More

This rule monitors for the usage of the most common video recording or screenshot utilities on unix systems by an uncommon process parent. Adversaries may collect video or screenshot data from users or systems for a variety of reasons including espionage, credential theft, or reconnaissance.

Read More

Identifies handle requests for the Local Security Authority Subsystem Service (LSASS) object access with specific access masks that many tools with a capability to dump memory to disk use (0x1fffff, 0x1010, 0x120089). This rule is tool agnostic as it has been validated against a host of various LSASS dump tools such as SharpDump, Procdump, Mimikatz, Comsvcs etc. It detects this behavior at a low level and does not depend on a specific tool or dump file name.

Read More

Identifies when a user creates a new inbox rule in Microsoft 365 that deletes or moves emails containing suspicious keywords. Adversaries who have compromised accounts often create inbox rules to hide alerts, security notifications, or other sensitive messages by automatically deleting them or moving them to obscure folders. Common destinations include Deleted Items, Junk Email, RSS Feeds, and RSS Subscriptions. This is a New Terms rule that triggers only when the user principal name and associated source IP address have not been observed performing this activity in the past 14 days.

Read More

Identifies suspicious Microsoft 365 mail access by ClientAppId. This rule detects when a user accesses their mailbox using a client application that is not typically used by the user, which may indicate potential compromise or unauthorized access attempts. Adversaries may use custom or third-party applications to access mailboxes, bypassing standard security controls. First-party Microsoft applications are also abused after OAuth tokens are compromised, allowing adversaries to access mailboxes without raising suspicion.

Read More

Identifies the assignment of rights to access content from another mailbox. An adversary may use the compromised account to send messages to other accounts in the network of the target organization while creating inbox rules, so messages can evade spam/phishing detection mechanisms.

Read More

Identifies a Microsoft 365 user completing an OAuth device code grant ("Cmsi:Cmsi") from a non-compliant device for the first time within the rule's historical window, regardless of the requesting application or target resource. Device code phishing kits complete the full login (password and MFA) at the genuine Microsoft endpoint and harvest the resulting token by polling, so MFA does not stop them. Because the victim authorizes the flow in their own browser, the grant is frequently completed on a personal or attacker-controlled device that is not enrolled or compliant with the organization's device policies. A user appearing with this device code flow on a non-compliant device for the first time in the lookback window is a strong early indicator of device code phishing, and removing the application and target constraints catches grants against any first-party application, not just the Microsoft Authentication Broker.

Read More

Identifies a Microsoft 365 OAuth device code grant ("Cmsi:Cmsi") with application Microsoft Authentication Broker ("29d9ed98-a469-4536-ade2-f981bc1d605e") for Microsoft Graph from a source ASN not previously observed for that user in a historical window. Phishing kits leveraging device code phishing complete the full login (password and MFA) at the genuine Microsoft endpoint and harvest the resulting token by polling, so MFA does not stop them and the authorization commonly originates from attacker-controlled residential proxy or hosting infrastructure rather than the user's normal network.

Read More

Detects successful Microsoft 365 portal logins from a country and region the user has not previously authenticated from in a specific time window. Atypical regions are identified by combining the user's country and region geolocation history; an authentication from a new country/region pair for that user may indicate an adversary attempting to access the account from an unusual location or behind a VPN.

Read More

Identifies an Microsoft 365 illicit consent grant request on-behalf-of a registered Entra ID application. Adversaries may create and register an application in Microsoft Entra ID for the purpose of requesting user consent to access resources in Microsoft 365. This is accomplished by tricking a user into granting consent to the application, typically via a pre-made phishing URL. This establishes an OAuth grant that allows the malicious client applocation to access resources in Microsoft 365 on-behalf-of the user.

Read More

Identifies a successful login by a user principal through a legacy authenticated client (such as Authenticated SMTP, IMAP, POP, or Exchange ActiveSync) in the Microsoft 365 Unified Audit Log, evidenced by the "BAV2ROPC" user agent. Legacy basic-authentication clients are translated by Entra ID into a Resource Owner Password Credentials (ROPC) grant, a single-factor flow that submits the user's password directly and bypasses interactive multi-factor authentication. This is commonly abused during password spraying and account takeover.

Read More

Identifies the first occurrence of SSO, SAML, or federated authentication errors for a user. These errors may indicate token manipulation, SAML assertion tampering, or OAuth phishing attempts. Modern adversaries often target SSO mechanisms through token theft, SAML response manipulation, or exploiting federated authentication weaknesses rather than traditional brute force attacks.

Read More

Identifies file downloads or access from OneDrive or SharePoint using PowerShell-based user agents. Adversaries may use native PowerShell cmdlets like Invoke-WebRequest or Invoke-RestMethod with Microsoft Graph API to exfiltrate data after compromising OAuth tokens via device code phishing or other credential theft techniques. This rule detects both direct PowerShell access and PnP PowerShell module usage for file operations. FileAccessed events are included to detect adversaries reading file content via API and saving locally, bypassing traditional download methods. Normal users access SharePoint/OneDrive via browsers or sync clients, making PowerShell-based file access inherently suspicious.

Read More

An instance of MSBuild, the Microsoft Build Engine, started a PowerShell script or the Visual C# Command Line Compiler. This technique is sometimes used to deploy a malicious payload using the Build Engine.

Read More

An instance of MSBuild, the Microsoft Build Engine, was started by a script or the Windows command interpreter. This behavior is unusual and is sometimes used by malicious payloads.

Read More

Identifies access to email resources via Microsoft Graph API using an first-party application on behalf of a user principal. This behavior may indicate an adversary using a phished OAuth refresh token or a Primary Refresh Token (PRT) to access email resources. The pattern includes requests to Microsoft Graph API endpoints related to email, such as /me/mailFolders/inbox/messages or /users/{user_id}/messages, using a public client application ID and a user principal object ID. This is a New Terms rule that only signals if the application ID, user principal object ID, and source ASN have not been seen doing this activity historically.

Read More

This New Terms rule focuses on the first occurrence of a client application ID (azure.graphactivitylogs.properties.app_id) making a request to Microsoft Graph API for a specific tenant ID (azure.tenant_id) and user principal object ID (azure.graphactivitylogs.properties.user_principal_object_id). This rule may helps identify unauthorized access or actions performed by compromised accounts. Advesaries may succesfully compromise a user's credentials and use the Microsoft Graph API to access resources or perform actions on behalf of the user.

Read More

Identifies modification of the dynamic linker preload shared object (ld.so.preload). Adversaries may execute malicious payloads by hijacking the dynamic linker used to load libraries.

Read More

This rule monitors for network connections from a kworker process. kworker, or kernel worker, processes are part of the kernel's workqueue mechanism. They are responsible for executing work that has been scheduled to be done in kernel space, which might include tasks like handling interrupts, background activities, and other kernel-related tasks. Attackers may attempt to evade detection by masquerading as a kernel worker process.

Read More

This rule detects the creation of a self-hosted Github runner from a first time seen user.name in the last 5 days. Adversaries may abuse self-hosted runners to execute workflow jobs on customer infrastructure.

Read More