M365 Identity OAuth Flow by First-Party Microsoft App from Multiple IPs
Identifies sign-ins on behalf of a principal user to the Microsoft Graph or legacy Azure AD API from multiple IPs using first-party Microsoft applications from the FOCI (Family of Client IDs) group. Developer tools like Azure CLI, VSCode, and Azure PowerShell accessing these resources from multiple IPs are flagged, along with any FOCI application accessing the deprecated Windows Azure Active Directory from multiple IPs. This behavior may indicate an adversary using a phished OAuth authorization code or refresh token, as seen in attacks like ConsentFix where attackers steal localhost OAuth codes and replay them from attacker infrastructure.
Elastic rule (View on GitHub)
1[metadata]
2creation_date = "2025/05/01"
3integration = ["o365"]
4maturity = "production"
5updated_date = "2025/12/17"
6
7[rule]
8author = ["Elastic"]
9description = """
10Identifies sign-ins on behalf of a principal user to the Microsoft Graph or legacy Azure AD API from multiple IPs using
11first-party Microsoft applications from the FOCI (Family of Client IDs) group. Developer tools like Azure CLI, VSCode,
12and Azure PowerShell accessing these resources from multiple IPs are flagged, along with any FOCI application accessing
13the deprecated Windows Azure Active Directory from multiple IPs. This behavior may indicate an adversary using a phished
14OAuth authorization code or refresh token, as seen in attacks like ConsentFix where attackers steal localhost OAuth
15codes and replay them from attacker infrastructure.
16"""
17from = "now-60m"
18interval = "59m"
19language = "esql"
20license = "Elastic License v2"
21name = "M365 Identity OAuth Flow by First-Party Microsoft App from Multiple IPs"
22note = """## Triage and analysis
23
24### Investigating M365 Identity OAuth Flow by First-Party Microsoft App from Multiple IPs
25
26This rule detects when the same user authenticates to Microsoft Graph or legacy Azure AD using FOCI applications from multiple IP addresses within a 30-minute window. This pattern is a strong indicator of OAuth code/token theft attacks like ConsentFix, where the victim completes the OAuth authorize flow on their device (first IP), and the attacker exchanges the stolen authorization code for tokens from their infrastructure (second IP).
27
28The rule aggregates events by user, application, and resource, requiring both `OAuth2:Authorize` and `OAuth2:Token` requests from at least 2 different IPs to fire - this indicates the code was generated on one IP and exchanged on another.
29
30### Possible investigation steps
31
32- Review `o365.audit.UserId` to identify the affected user and determine if they are a high-value target.
33- Analyze `Esql.source_ip_values` to see all unique IP addresses used within the 30-minute window. Determine whether these originate from different geographic regions, cloud providers (AWS, Azure, GCP), or anonymizing infrastructure (Tor, VPNs).
34- Use `Esql.time_window_date_trunc` to pivot into raw events and reconstruct the full sequence of resource access events with exact timestamps.
35- Check `Esql.source_as_organization_name_values` for unfamiliar ASN organizations that may indicate attacker infrastructure.
36- Review `Esql.o365_audit_ApplicationId_values` to confirm which first-party application was used.
37- Pivot to `azure.auditlogs` to check for device join or registration events around the same timeframe, which may indicate persistence attempts.
38- Correlate with `azure.identityprotection` to identify related risk detections such as anonymized IP access or token replay.
39- Search for additional sign-ins from the IPs involved across other users to determine if this is part of a broader campaign.
40
41### False positive analysis
42
43- Developers or IT administrators working across environments (office, home, cloud VMs) may produce similar behavior.
44- Users on VPN who switch servers or traveling between networks may show multiple IPs.
45- Mobile users moving between cellular and WiFi networks during the time window.
46- Consider correlating with device compliance status to distinguish managed vs. unmanaged access.
47
48### Response and remediation
49
50- If confirmed unauthorized, immediately revoke all refresh tokens for the affected user via Entra ID.
51- Remove any devices registered during this session by checking `azure.auditlogs` for `Add device` events.
52- Notify the user and determine whether they may have shared an OAuth code via phishing.
53- Block the attacker IPs at the perimeter and add to threat intel feeds.
54- Implement Conditional Access policies to restrict OAuth flows for these applications to compliant devices and approved locations.
55- Monitor for follow-on activity like lateral movement, privilege escalation, or data exfiltration via Graph API.
56"""
57references = [
58 "https://www.volexity.com/blog/2025/04/22/phishing-for-codes-russian-threat-actors-target-microsoft-365-oauth-workflows/",
59 "https://github.com/dirkjanm/ROADtools",
60 "https://dirkjanm.io/phishing-for-microsoft-entra-primary-refresh-tokens/",
61 "https://pushsecurity.com/blog/consentfix",
62 "https://github.com/secureworks/family-of-client-ids-research",
63]
64risk_score = 73
65rule_id = "36188365-f88f-4f70-8c1d-0b9554186b9c"
66setup = """## Setup
67
68The Office 365 Logs Fleet integration, Filebeat module, or similarly structured data is required to be compatible with this rule.
69"""
70severity = "high"
71tags = [
72 "Domain: Cloud",
73 "Domain: Email",
74 "Domain: Identity",
75 "Data Source: Microsoft 365",
76 "Data Source: Microsoft 365 Audit Logs",
77 "Use Case: Identity and Access Audit",
78 "Use Case: Threat Detection",
79 "Resources: Investigation Guide",
80 "Tactic: Defense Evasion",
81]
82timestamp_override = "event.ingested"
83type = "esql"
84
85query = '''
86from logs-o365.audit-*
87| where
88 event.dataset == "o365.audit" and
89 event.action == "UserLoggedIn" and
90 source.ip is not null and
91 o365.audit.UserId is not null and
92 o365.audit.ApplicationId is not null and
93 o365.audit.UserType in ("0", "2", "3", "10") and
94 (
95 /* Developer tools accessing Graph OR Legacy AAD */
96 (
97 o365.audit.ApplicationId in (
98 "aebc6443-996d-45c2-90f0-388ff96faa56",
99 "29d9ed98-a469-4536-ade2-f981bc1d605e",
100 "04b07795-8ddb-461a-bbee-02f9e1bf7b46",
101 "1950a258-227b-4e31-a9cf-717495945fc2"
102 ) and
103 o365.audit.ObjectId in (
104 "00000003-0000-0000-c000-000000000000",
105 "00000002-0000-0000-c000-000000000000"
106 )
107 ) or
108 /* Any FOCI app accessing Legacy AAD only */
109 (
110 o365.audit.ApplicationId in (
111 "00b41c95-dab0-4487-9791-b9d2c32c80f2",
112 "1fec8e78-bce4-4aaf-ab1b-5451cc387264",
113 "26a7ee05-5602-4d76-a7ba-eae8b7b67941",
114 "27922004-5251-4030-b22d-91ecd9a37ea4",
115 "4813382a-8fa7-425e-ab75-3b753aab3abb",
116 "ab9b8c07-8f02-4f72-87fa-80105867a763",
117 "d3590ed6-52b3-4102-aeff-aad2292ab01c",
118 "872cd9fa-d31f-45e0-9eab-6e460a02d1f1",
119 "af124e86-4e96-495a-b70a-90f90ab96707",
120 "2d7f3606-b07d-41d1-b9d2-0d0c9296a6e8",
121 "844cca35-0656-46ce-b636-13f48b0eecbd",
122 "87749df4-7ccf-48f8-aa87-704bad0e0e16",
123 "cf36b471-5b44-428c-9ce7-313bf84528de",
124 "0ec893e0-5785-4de6-99da-4ed124e5296c",
125 "22098786-6e16-43cc-a27d-191a01a1e3b5",
126 "4e291c71-d680-4d0e-9640-0a3358e31177",
127 "57336123-6e14-4acc-8dcf-287b6088aa28",
128 "57fcbcfa-7cee-4eb1-8b25-12d2030b4ee0",
129 "66375f6b-983f-4c2c-9701-d680650f588f",
130 "9ba1a5c7-f17a-4de9-a1f1-6178c8d51223",
131 "a40d7d7d-59aa-447e-a655-679a4107e548",
132 "a569458c-7f2b-45cb-bab9-b7dee514d112",
133 "b26aadf8-566f-4478-926f-589f601d9c74",
134 "c0d2a505-13b8-4ae0-aa9e-cddd5eab0b12",
135 "d326c1ce-6cc6-4de2-bebc-4591e5e13ef0",
136 "e9c51622-460d-4d3d-952d-966a5b1da34c",
137 "eb539595-3fe1-474e-9c1d-feb3625d1be5",
138 "ecd6b820-32c2-49b6-98a6-444530e5a77a",
139 "f05ff7c9-f75a-4acd-a3b5-f4b6a870245d",
140 "f44b1140-bc5e-48c6-8dc0-5cf5a53c0e34",
141 "be1918be-3fe3-4be9-b32b-b542fc27f02e",
142 "cab96880-db5b-4e15-90a7-f3f1d62ffe39",
143 "d7b530a4-7680-4c23-a8bf-c52c121d2e87",
144 "dd47d17a-3194-4d86-bfd5-c6ae6f5651e3",
145 "e9b154d0-7658-433b-bb25-6b8e0a8a7c59"
146 ) and
147 o365.audit.ObjectId == "00000002-0000-0000-c000-000000000000"
148 )
149 )
150| eval
151 Esql.time_window_date_trunc = date_trunc(30 minutes, @timestamp),
152 Esql.oauth_authorize_user_id_case = case(
153 o365.audit.ExtendedProperties.RequestType == "OAuth2:Authorize" and o365.audit.ExtendedProperties.ResultStatusDetail == "Redirect",
154 o365.audit.UserId,
155 null
156 ),
157 Esql.oauth_token_user_id_case = case(
158 o365.audit.ExtendedProperties.RequestType == "OAuth2:Token",
159 o365.audit.UserId,
160 null
161 )
162| stats
163 Esql.source_ip_count_distinct = count_distinct(source.ip),
164 Esql.source_ip_values = values(source.ip),
165 Esql.o365_audit_ApplicationId_values = values(o365.audit.ApplicationId),
166 Esql.source_as_organization_name_values = values(source.`as`.organization.name),
167 Esql.oauth_token_count_distinct = count_distinct(Esql.oauth_token_user_id_case),
168 Esql.oauth_authorize_count_distinct = count_distinct(Esql.oauth_authorize_user_id_case)
169 by
170 o365.audit.UserId,
171 Esql.time_window_date_trunc,
172 o365.audit.ApplicationId,
173 o365.audit.ObjectId
174| keep
175 Esql.time_window_date_trunc,
176 Esql.source_ip_values,
177 Esql.source_ip_count_distinct,
178 Esql.o365_audit_ApplicationId_values,
179 Esql.source_as_organization_name_values,
180 Esql.oauth_token_count_distinct,
181 Esql.oauth_authorize_count_distinct
182| where
183 Esql.source_ip_count_distinct >= 2 and
184 Esql.oauth_token_count_distinct > 0 and
185 Esql.oauth_authorize_count_distinct > 0
186'''
187
188
189[[rule.threat]]
190framework = "MITRE ATT&CK"
191[[rule.threat.technique]]
192id = "T1550"
193name = "Use Alternate Authentication Material"
194reference = "https://attack.mitre.org/techniques/T1550/"
195[[rule.threat.technique.subtechnique]]
196id = "T1550.001"
197name = "Application Access Token"
198reference = "https://attack.mitre.org/techniques/T1550/001/"
199
200
201
202[rule.threat.tactic]
203id = "TA0005"
204name = "Defense Evasion"
205reference = "https://attack.mitre.org/tactics/TA0005/"
206[[rule.threat]]
207framework = "MITRE ATT&CK"
208[[rule.threat.technique]]
209id = "T1528"
210name = "Steal Application Access Token"
211reference = "https://attack.mitre.org/techniques/T1528/"
212
213
214[rule.threat.tactic]
215id = "TA0006"
216name = "Credential Access"
217reference = "https://attack.mitre.org/tactics/TA0006/"
218[[rule.threat]]
219framework = "MITRE ATT&CK"
220[[rule.threat.technique]]
221id = "T1566"
222name = "Phishing"
223reference = "https://attack.mitre.org/techniques/T1566/"
224[[rule.threat.technique.subtechnique]]
225id = "T1566.002"
226name = "Spearphishing Link"
227reference = "https://attack.mitre.org/techniques/T1566/002/"
228
229
230
231[rule.threat.tactic]
232id = "TA0001"
233name = "Initial Access"
234reference = "https://attack.mitre.org/tactics/TA0001/"
Triage and analysis
Investigating M365 Identity OAuth Flow by First-Party Microsoft App from Multiple IPs
This rule detects when the same user authenticates to Microsoft Graph or legacy Azure AD using FOCI applications from multiple IP addresses within a 30-minute window. This pattern is a strong indicator of OAuth code/token theft attacks like ConsentFix, where the victim completes the OAuth authorize flow on their device (first IP), and the attacker exchanges the stolen authorization code for tokens from their infrastructure (second IP).
The rule aggregates events by user, application, and resource, requiring both OAuth2:Authorize and OAuth2:Token requests from at least 2 different IPs to fire - this indicates the code was generated on one IP and exchanged on another.
Possible investigation steps
- Review
o365.audit.UserIdto identify the affected user and determine if they are a high-value target. - Analyze
Esql.source_ip_valuesto see all unique IP addresses used within the 30-minute window. Determine whether these originate from different geographic regions, cloud providers (AWS, Azure, GCP), or anonymizing infrastructure (Tor, VPNs). - Use
Esql.time_window_date_truncto pivot into raw events and reconstruct the full sequence of resource access events with exact timestamps. - Check
Esql.source_as_organization_name_valuesfor unfamiliar ASN organizations that may indicate attacker infrastructure. - Review
Esql.o365_audit_ApplicationId_valuesto confirm which first-party application was used. - Pivot to
azure.auditlogsto check for device join or registration events around the same timeframe, which may indicate persistence attempts. - Correlate with
azure.identityprotectionto identify related risk detections such as anonymized IP access or token replay. - Search for additional sign-ins from the IPs involved across other users to determine if this is part of a broader campaign.
False positive analysis
- Developers or IT administrators working across environments (office, home, cloud VMs) may produce similar behavior.
- Users on VPN who switch servers or traveling between networks may show multiple IPs.
- Mobile users moving between cellular and WiFi networks during the time window.
- Consider correlating with device compliance status to distinguish managed vs. unmanaged access.
Response and remediation
- If confirmed unauthorized, immediately revoke all refresh tokens for the affected user via Entra ID.
- Remove any devices registered during this session by checking
azure.auditlogsforAdd deviceevents. - Notify the user and determine whether they may have shared an OAuth code via phishing.
- Block the attacker IPs at the perimeter and add to threat intel feeds.
- Implement Conditional Access policies to restrict OAuth flows for these applications to compliant devices and approved locations.
- Monitor for follow-on activity like lateral movement, privilege escalation, or data exfiltration via Graph API.
References
Related rules
- M365 Exchange Inbox Phishing Evasion Rule Created
- M365 Identity Login from Atypical Travel Location
- M365 Identity Login from Impossible Travel Location
- Entra ID OAuth User Impersonation to Microsoft Graph
- Entra ID Sign-in BloodHound Suite User-Agent Detected