New Agent Added to Pool by New User or Added to a New OS Type

Status: available
Severity: medium
Time window: 14d
Source: github.com/Azure/Azure-Sentinel

'As seen in attacks such as SolarWinds attackers can look to subvert a build process by controlling build servers. Azure DevOps uses agent pools to execute pipeline tasks. An attacker could insert compromised agents that they control into the pools in order to execute malicious code. This query looks for users adding agents to pools they have not added agents to before, or adding agents to a pool of an OS that has not been added to that pool before. This detection has potential for false positives so has a configurable allow list to allow for certain users to be excluded from the logic.'

MITRE ATT&CK coverage

Tactic	Techniques
Execution	`T1053` Scheduled Task/Job

Rule body kusto

id: 4ce177b3-56b1-4f0e-b83e-27eed4cb0b16
name: New Agent Added to Pool by New User or Added to a New OS Type
description: |
  'As seen in attacks such as SolarWinds attackers can look to subvert a build process by controlling build servers. Azure DevOps uses agent pools to execute pipeline tasks. 
  An attacker could insert compromised agents that they control into the pools in order to execute malicious code. This query looks for users adding agents to pools they have not added agents to before, or adding agents to a pool of an OS that has not been added to that pool before. This detection has potential for false positives so has a configurable allow list to allow for certain users to be excluded from the logic.'
severity: Medium
status: Available
requiredDataConnectors: []
queryFrequency: 1d
queryPeriod: 14d
triggerOperator: gt
triggerThreshold: 0
tactics:
  - Execution
relevantTechniques:
  - T1053
query: |
  let lookback = 14d;
  let timeframe = 1d;
  // exclude allowed users from query such as the ADO service
  let allowed_users = dynamic(["Azure DevOps Service"]);
  union
  // Look for agents being added to a pool of a OS type not seen with that pool before
  (ADOAuditLogs
  | where TimeGenerated > ago(lookback) and TimeGenerated < ago(timeframe)
  | where OperationName =~ "Library.AgentAdded"
  | where ActorUPN !in (allowed_users)
  | extend AgentPoolName = tostring(Data.AgentPoolName)
  | extend OsDescription = tostring(Data.OsDescription)
  | where isnotempty(OsDescription)
  | extend OsDescription = tostring(split(OsDescription, "#", 0)[0])
  | project AgentPoolName, OsDescription
  | join kind=rightanti (ADOAuditLogs
  | where TimeGenerated > ago(timeframe)
  | where OperationName == "Library.AgentAdded"
  | extend AgentPoolName = tostring(Data.AgentPoolName)
  | extend OsDescription = tostring(Data.OsDescription)
  | where isnotempty(OsDescription)
  | extend OsDescription = tostring(split(OsDescription, "#", 0)[0])) on AgentPoolName, OsDescription),
  // Look for users addeing agents to a pool that they have not added agents to before.
  (AzureDevOpsAuditing
  | where TimeGenerated > ago(lookback) and TimeGenerated < ago(timeframe)
  | extend AgentPoolName = tostring(Data.AgentPoolName)
  | where ActorUPN !in (allowed_users)
  | project AgentPoolName, ActorUPN
  | join kind=rightanti (AzureDevOpsAuditing
  | where TimeGenerated > ago(timeframe)
  | where OperationName == "Library.AgentAdded"
  | where ActorUPN !in (allowed_users)
  | extend AgentPoolName = tostring(Data.AgentPoolName)
  ) on AgentPoolName, ActorUPN)
  | extend AgentName = tostring(Data.AgentName)
  | extend OsDescription = tostring(Data.OsDescription)
  | extend SystemDetails = Data.SystemCapabilities
  | project-reorder TimeGenerated, OperationName, ScopeDisplayName, AgentPoolName, AgentName, ActorUPN, IpAddress, UserAgent, OsDescription, SystemDetails, Data
  | extend timestamp = TimeGenerated
  | extend AccountName = tostring(split(ActorUPN, "@")[0]), AccountUPNSuffix = tostring(split(ActorUPN, "@")[1])
entityMappings:
  - entityType: Account
    fieldMappings:
      - identifier: FullName
        columnName: ActorUPN
      - identifier: Name
        columnName: AccountName
      - identifier: UPNSuffix
        columnName: AccountUPNSuffix
  - entityType: IP
    fieldMappings:
      - identifier: Address
        columnName: IpAddress
version: 1.0.6
kind: Scheduled

Stages and Predicates

Parameters

let lookback = 14d;
let timeframe = 1d;
let allowed_users = dynamic(["Azure DevOps Service"]);

union (2 sources)

Each leg below queries one source; the rule matches if any leg does. Sources: ADOAuditLogs, AzureDevOpsAuditing

Leg 1: `ADOAuditLogs`

ADOAuditLogs
| where TimeGenerated > ago(lookback) and TimeGenerated < ago(timeframe)
| where OperationName =~ "Library.AgentAdded"
| where ActorUPN !in (allowed_users)
| extend AgentPoolName = tostring(Data.AgentPoolName)
| extend OsDescription = tostring(Data.OsDescription)
| where isnotempty(OsDescription)
| extend OsDescription = tostring(split(OsDescription, "#", 0)[0])
| project AgentPoolName, OsDescription
| join kind=rightanti (ADOAuditLogs
| where TimeGenerated > ago(timeframe)
| where OperationName == "Library.AgentAdded"
| extend AgentPoolName = tostring(Data.AgentPoolName)
| extend OsDescription = tostring(Data.OsDescription)
| where isnotempty(OsDescription)
| extend OsDescription = tostring(split(OsDescription, "#", 0)[0])) on AgentPoolName, OsDescription

Leg 2: `AzureDevOpsAuditing`

AzureDevOpsAuditing
| where TimeGenerated > ago(lookback) and TimeGenerated < ago(timeframe)
| extend AgentPoolName = tostring(Data.AgentPoolName)
| where ActorUPN !in (allowed_users)
| project AgentPoolName, ActorUPN
| join kind=rightanti (AzureDevOpsAuditing
| where TimeGenerated > ago(timeframe)
| where OperationName == "Library.AgentAdded"
| where ActorUPN !in (allowed_users)
| extend AgentPoolName = tostring(Data.AgentPoolName)
) on AgentPoolName, ActorUPN

Applied to the combined result

| extend AgentName = tostring(Data.AgentName)
| extend OsDescription = tostring(Data.OsDescription)
| extend SystemDetails = Data.SystemCapabilities
| project-reorder TimeGenerated, OperationName, ScopeDisplayName, AgentPoolName, AgentName, ActorUPN, IpAddress, UserAgent, OsDescription, SystemDetails, Data
| extend timestamp = TimeGenerated
| extend AccountName = tostring(split(ActorUPN, "@")[0]), AccountUPNSuffix = tostring(split(ActorUPN, "@")[1])

Exclusions

Top-level NOT(...) conjuncts: predicates this rule actively suppresses.

Field	Kind	Excluded values
`ActorUPN`	eq	`allowed_users`
`OperationName`	eq	`Library.AgentAdded`
`OsDescription`	is_not_null	(no value, null check)
`ActorUPN`	eq	`allowed_users`
`OperationName`	eq	`Library.AgentAdded`

Indicators

Each row is a field, operator, and value that the rule matches. The corpus column counts how many other rules in the catalog look for the same combination: high numbers point to widely-used, community-vetted indicators. Blank or 1 shows that the indicator is specific to this rule.

Field	Kind	Values
`ActorUPN`	in	`allowed_users` transforms: `cased`
`OperationName`	eq	`Library.AgentAdded`
`OsDescription`	is_not_null	(no value, null check)

Output fields

Fields the rule emits when it matches. Chronicle authors list these in the outcome block; they appear on the detection and $risk_score drives alerting. Sentinel / Defender XDR rules build them up through project / summarize / extend stages. Sentinel maps these into alert fields via entityMappings and customDetails; Defender XDR custom detections surface them as alert fields directly.

Field	Source
`ActorUPN`	`project`
`AgentPoolName`	`project`
`AgentName`	`extend`
`OsDescription`	`extend`
`SystemDetails`	`extend`
`timestamp`	`extend`
`AccountName`	`extend`
`AccountUPNSuffix`	`extend`

`j` / `k`	Scroll down / up
`d` / `u`	Half-page down / up
`gg` / `G`	Top / bottom
`h` / `l`	History back / forward
`f`	Follow link (`Shift` = new tab)
`/`	Focus search
`?`	Toggle this help
`↑` / `↓`	Navigate search results
`Enter`	Open highlighted result
`Esc`	Close results / dialog

`type:`	`events` / `rules` / `providers`
`vendor:`	`sigma` / `elastic` / `splunk` / `kusto` / `chronicle` (vendor name alone also works: `sigma:`, `kql:`, `secops:`…)
`tactic:`	TA-id, slug, or name: `credential_access`, `TA0006`
`technique:`	technique or sub-technique ID: `T1003`, `T1003.001` (alias `tech:`)
`severity:`	`critical` / `high` / `medium` / `low` / `informational` (alias `sev:`)
`risk_score`	Numeric comparison on the Elastic risk score (0 to 100): `risk_score>50`, `risk_score<=20`, `risk_score=99` (alias `risk`; Elastic rules only)
`stages:`	Rules with exactly N pipeline stages
`correlation:`	`single_event` / `sequence` / `alternatives` / `alternatives_cross_log` / `all_required` / `correlated`
`with:`	Co-occurrence event-id; stacks (`with:4624 with:4769`) to require all, while a comma list in one occurrence (`with:4624,4769`) is an either-or group. Implies multi-event
`like:`	Structural neighbors of a rule slug (equivalents + subsumption stricter / broader): `like:comsvcs_lsass_memory_dump-splunk-sysmon`
`groupby:`	Entity-grouping substring match against `group_by_keys`: `groupby:user`, `groupby:host`
`uses:`	Rules whose predicate tree touches the field (any kind, any value): `uses:CommandLine`
`excludes:`	Rules with top-level `not()` clauses on the field (FP whitelists): `excludes:ParentImage`
`field:` / `value:`	Predicate search; narrows rule cards to those with a matching leaf and drives the indicator tier. Unquoted = substring, wildcards allowed (`value:mimikatz`)
`indicator:`	Shorthand for `field:F value:V`: `indicator:Image=*\powershell.exe`
`kind:`	Filter by predicate kind. Narrows rule cards to those carrying a matching predicate leaf (`vendor:elastic kind:cidr_match`) and drives the indicator tier: `contains` / `starts_with` / `ends_with` / `regex` / `cidr` / `eq` / `in` … (operator aliases `op:`/`match:`)
`has:` / `no:`	`sample`, `field`, `notes`, `refs`, `trace`, `thirdparty`, `rule`, `pattern`, `timewindow`, `threshold`, `newterms`, `sigma`/`elastic`/`splunk`/`kusto`/`chronicle`
`-op:val`	Exclude matches; works on most operators but not `type:`/`like:`/`has:`/`no:` (use `no:<flag>` to exclude a rule flag): `tactic:execution -vendor:splunk`. Standalone `-kind:`/`-field:`/`-value:` drop every rule carrying a matching predicate leaf (`type:rules -kind:is_null`)
`field:"…"` / `value:"…"`	Quoted value = anchored exact match (also allows spaces): `value:"net user"`
`a,b`	Comma = OR inside one operator (`vendor:sigma,elastic`, `severity:high,critical`); repeating a facet merges the same way. `field:`/`value:` never split (literal commas)
`vendors:` / `stage:`	Singular and plural spellings fold to the canonical operator and value: `tactics:` = `tactic:`, `type:event` = `type:events`, `correlation:sequences` = `correlation:sequence`, `has:thresholds` = `has:threshold`
`"quoted phrase"`	Exact-match a multi-word phrase (free text)

New Agent Added to Pool by New User or Added to a New OS Type

MITRE ATT&CK coverage

Rule body kusto

Stages and Predicates

Parameters

union (2 sources)

Leg 1: `ADOAuditLogs`

Leg 2: `AzureDevOpsAuditing`

Applied to the combined result

Exclusions

Indicators

Output fields

Keyboard shortcuts

Search operators

New Agent Added to Pool by New User or Added to a New OS Type

MITRE ATT&CK coverage

Rule body kusto

Stages and Predicates

Parameters

union (2 sources)

Leg 1: ADOAuditLogs

Leg 2: AzureDevOpsAuditing

Applied to the combined result

Exclusions

Indicators

Output fields

Leg 1: `ADOAuditLogs`

Leg 2: `AzureDevOpsAuditing`