Edit

Share via


Prioritize incidents and hunt for threats with triage collection (preview)

Important

This information relates to a prerelease product that may be substantially modified before it's released. Microsoft makes no warranties, expressed or implied, with respect to the information provided here.

The triage collection in the Microsoft Sentinel Model Context Protocol (MCP) server integrates your AI models with APIs that support incident triage and hunting. This integration lets you prioritize incidents quickly and hunt over your own data easily, reducing mean time to resolution, risk exposure, and dwell time.

Use the tool for the following scenarios:

  • Incident triage: Prioritize incidents quickly by using your own AI models, reducing mean time to resolution. Use the tools to fetch incidents, alerts, alerts evidence, entities, and other data.
  • Hunting: Hunt over your data easily by using your own AI models, reducing risk exposure and dwell time. Use the tools to run hunting queries and fetch the required data during your hunt.

Prerequisites

To access the triage tool collection, you must have the following prerequisites:

  • Microsoft Defender XDR, Microsoft Defender for Endpoint, or Microsoft Sentinel onboarded to the Defender portal
  • Any of the supported AI-powered code editors and agent-building platforms:

Add the triage collection

To add the data exploration collection, you must first set up add Microsoft Sentinel's unified MCP server interface. Follow the step-by-step instructions for compatible AI-powered code editors and agent-building platforms listed in the Prerequisites section.

The triage collection is hosted at the following URL:

https://sentinel.microsoft.com/mcp/triage

Tools in the triage collection

List security incidents (ListIncidents)

This tool lists security incidents and filters them by date range, severity, status, assigned analyst, and investigation state.

Parameters Required? Description
createdAfter No Time after which the incident was created
createdBefore No Time before the incident was created
Severity No Severity assigned to the incident (for example, Low or High)
Status No Current status of the incident (New, Active, or Closed)
AssignedTo No To which user the incident is assigned
Classification No Classification (for example, True Positive or False Positive)
Determination No Determination (for example, Malware or Phishing)
orderBy No Instructions to order the incidents returned
Search No Free-text search across the incident data
includeAlertsData No Option to include data from the underlying alerts
skip No Skips a specified number of items from the start of the result set
top No Limits the number of items returned in the response

Get a security incident (GetIncidentById)

This tool retrieves a security incident by ID, including its properties, correlated alerts, and metadata such as status, severity, classification, and timestamps.

Parameters Required? Description
incidentID Yes Identifier is associated with the incident
includeAlertsData No Option to include data from the underlying alerts

This tool lists security alerts, sorts them, and filters them by date range, severity, and status.

Parameters Required? Description
createdAfter No Time after which the alert was created
createdBefore No Time before the alert was created
Severity No Severity assigned to the alert (for example, Low or High)
status No Current status of the alert; possible values: Unknown, New, InProgress, and Resolved
skip No Skips a specified number of items from the start of the result set
top No Limits the number of items returned in the response

Get a security alert (GetAlertByID)

This tool retrieves a security alert by ID. It returns the complete alert details, including severity, status, classification, and related evidence entities.

Parameters Required? Description
AlertID Yes Unique identifier of the alert

List advanced hunting tables (FetchAdvancedHuntingTablesOverview)

This tool lists the names of available advanced hunting tables and their brief descriptions. It's essential for understanding data sources before writing Kusto Query Language (KQL) queries.

Parameters Required? Description
tableNames No Advanced hunting table names

Get advanced hunting table schema (FetchAdvancedHuntingTablesDetailedSchema)

This tool retrieves complete column schemas with descriptions for specified advanced hunting tables. The information it provides is crucial for constructing error-free KQL queries. Use this tool before calling RunAdvancedHuntingQuery.

Parameters Required? Description
tableNames Yes Advanced hunting table names

Run hunting query (RunAdvancedHuntingQuery)

Run an advanced hunting query by using KQL across supported Microsoft Defender tables to proactively search for threats. To understand data sources, first run FetchAdvancedHuntingTablesOverview. For error-free KQL, first run FetchAdvancedHuntingTablesDetailedSchema.

Parameters Required? Description
kqlQuery Yes KQL query to run over the selected table
timestamp No Timestamp to choose for the query

Get file information (GetDefenderFileInfo)

Get file details such as hashes, size, type, publisher, signer certificate info, and global prevalence with first and last seen timestamps.

Parameters Required? Description
fileHash Yes SHA-1, SHA-256, or MD5 hash of the file

Get file statistics (GetDefenderFileStatistics)

Get organizational file prevalence statistics, including the number of devices where the file was observed.

Parameters Required? Description
fileHash Yes SHA-1, SHA-256, or MD5 hash of the file

Get file alerts (GetDefenderFileAlerts)

List all security alerts generated by a specific file across your organization, including historical and active alerts.

Parameters Required? Description
fileHash Yes SHA-1, SHA-256, or MD5 hash of the file

List all devices that encountered a specific file to assess its spread across your environment.

Parameters Required? Description
fileHash Yes SHA-1, SHA-256, or MD5 hash of the file

List threat indicators (ListDefenderIndicators)

List tenant indicators of compromise (IOCs) in Microsoft Defender for Endpoint. Use filters for type, value, action, and severity.

Parameters Required? Description
indicatorType No Indicator type (for example, file hash, domain name, or IP address)
indicatorValue No Specific value of the indicator to filter results
Action No Action applied to the indicator (Alert, Block, or Allow)
ApplicationName No Application associated with the indicator
Title No Title or description of the indicator
Severity No Severity level (Informational, Low, Medium, or High)
createdAfter No Return indicators created after this timestamp
createdBefore No Return indicators created before this timestamp

List automated investigations (ListDefenderInvestigations)

List automated investigation cases in Defender for Endpoint. Use filters for state, target device, start time, or triggering alert ID.

Parameters Required? Description
startTime No Return investigations started after this timestamp
endTime No Return investigations started before this timestamp
Status No Investigation state (Running, Completed, or Failed)
skip No Skips a specified number of items from the start of the result set
top No Limits the number of items returned in the response

Get automated investigation (GetDefenderInvestigation)

Get details of a specific automated investigation, including state, timestamps, target device, and triggering alert.

Parameters Required? Description
ID Yes Unique identifier of the investigation

Get all security alerts for an IP address (GetDefenderIpAlerts)

List all security alerts in the organization that are related to a specified IP address.

Parameters Required? Description
ipAddress Yes IP address to retrieve related alerts for

Get statistics for an IP address (GetDefenderIpStatistics)

Get statistics for a given IP address, including the number of distinct devices that communicated with it.

Parameters Required? Description
ipAddress Yes IP address to retrieve statistics for

Get endpoint device (GetDefenderMachine)

Get detailed information about a specific Defender for Endpoint device, including operating system details, health status, risk score, and exposure level.

Parameters Required? Description
ID Yes Unique identifier of the device

List all security alerts associated with a specific device for a device-centric view of threats.

Parameters Required? Description
ID Yes Unique identifier of the device

Get users that signed into a device (GetDefenderMachineLoggedOnUsers)

List accounts that signed in to a device. For each user, the API provides context such as the account username and domain.

Parameters Required? Description
ID Yes Unique identifier of the device

Get device vulnerabilities (GetDefenderMachineVulnerabilities)

List discovered security vulnerabilities on a device with Common Vulnerabilities and Exposures (CVE) details and risk assessment scores.

Parameters Required? Description
ID Yes Unique identifier of the device

Find device by internal IP address (FindDefenderMachineByIp)

List all devices that communicated with a specific internal IP address in the time range of 15 minutes before and after the given timestamp, for network mapping and lateral movement analysis.

Parameters Required? Description
ipAddress Yes Internal IP address to search for
timestamp Yes The timestamp that defines the query window, checking 15 minutes before and 15 minutes after the specified time

List remediation tasks (ListDefenderRemediationActivities)

List remediation tasks and their execution status across devices. Each remediation activity corresponds to a security recommendation or task.

Parameters Required? Description
Type No Type of remediation activity
machineID No Identifier of the affected device
Status No Status of the remediation task (Pending or Completed)
createdTimeFrom No Return tasks created after this timestamp
createdTimeTo No Return tasks created before this timestamp
skip No Skips a specified number of items from the start of the result set
top No Limits the number of items returned in the response

Get detailed remediation task information (GetDefenderRemediationActivity)

Get detailed remediation task information, including execution status, results, and affected devices.

Parameters Required? Description
ID Yes Unique identifier of the remediation activity

List all security alerts associated with a specific user account. This information is essential for user-centric threat investigations and behavior analysis.

Parameters Required? Description
ID Yes Unique identifier of the user account

List all devices active for a user (ListUserRelatedMachines)

List all devices where a specific user has active or recent sign-in sessions. Use this tool to track user activity and analyze lateral movement.

Parameters Required? Description
ID Yes Unique identifier of the user account

List all devices affected by a vulnerability (ListDefenderMachinesByVulnerability)

List all devices affected by a specific CVE vulnerability. This tool is critical for patch management prioritization.

Parameters Required? Description
cveID Yes CVE identifier of the vulnerability

List vulnerabilities affecting software (ListDefenderVulnerabilitiesBySoftware)

List vulnerabilities that affect specific software on a specific device for targeted vulnerability assessment.

Parameters Required? Description
machineID Yes Unique identifier of the device
softwareID Yes Unique identifier of the software

Sample prompts

The following sample prompts demonstrate what you can do with the triage collection:

  • List the last five incidents from my tenant and assess which one is the most urgent to triage
  • Provide the alerts for <specific incident> and analyze the alert evidence for maliciousness
  • Run a hunting query to check which users interacted with <entity>

Limitations

  • You can't use this collection as a guest in another tenant or with delegated access. You can only use the MCP server on your own home tenant.
  • Microsoft Sentinel users can't choose which workspace to use.
  • You can't query data in Microsoft Sentinel lake. You can use the data exploration tools instead.