Detections
Explore public detection logic contributed by the community across SIEM and rule languages.
1,064 detections
Filters
Last updated
All Time
Detection languages
597
173
103
71
51
Contributors
184
155
139
90
73
Categories
20,439
11,829
5,873
5,151
4,864
1,064
Platforms
575
266
228
223
118
Products / Services
498
220
130
118
86
MITRE Techniques
488
451
363
349
341
CVEs
10
5
4
3
2
IDS Classtypes
34
3
2
2
1
IDS Protocols
29
9
1
Detects an Active Directory password spray attack originating from a single source IP, followed closely by a user account creation event (Event ID 4720) associated with a logon from that same source IP.
Detects Microsoft 365 token replay from the Huntress token spray and replay campaign. Alerts on sign-ins from an IP new to the account (not seen on 3 or more distinct days in the 14d baseline) that match one of three behaviors: a scripted client (node or bare Mozilla/5.0 user agent with Office or Authentication Broker client IDs), Device Registration Service access (Medium-High only when corroborated by a campaign IP, scripted user agent or session IP hop, otherwise Low), or one session moving across IPs within 60 minutes. Rows are raised to High when the same new IP hits 3 or more baseline accounts in the window (token spray pattern) or matches a campaign IOC. Low Signal rows are dropped when AlertOnly = true. Does not cover MFA changes, device registration events or inbox rules.
Detects a suspicious sequence where a client IP performs a brute-force attack against WordPress login or XML-RPC endpoints followed by enumeration of the wp-json REST API.
Detects CVE-2026-76504 Cisco SD-WAN vManage authentication bypass attempts: POST logins using internal viptela-reserved-* accounts via the percent-encoded /%6a_security_check path. Summarized per host and account.
Detects CVE-2026-76504 Cisco SD-WAN vManage authentication bypass attempts: POST logins using internal viptela-reserved-* accounts via the percent-encoded /%6a_security_check path. Summarized per host and account.
Detects authentication to the MFT platform where multi-factor was not initiated.
The extractor records MFA_INITIATED separately from LOGIN, so a login with no
corresponding MFA event indicates either an account exempt from the policy or a
path around it — on an internet-facing transfer platform, both matter.
Onboarding: GOANYWHERE-MFT is ingested via Syslog (Bloo extractor 850).
Integration guide: https://bloo.io/docs/hypercloud/syslog-connector
BQL (Bloo Query Language):
stream=authentication where sourcename='GOANYWHERE-MFT' and action='LOGIN'
| groupby user, srcip, authproto
The extractor records MFA_INITIATED separately from LOGIN, so a login with no
corresponding MFA event indicates either an account exempt from the policy or a
path around it — on an internet-facing transfer platform, both matter.
Onboarding: GOANYWHERE-MFT is ingested via Syslog (Bloo extractor 850).
Integration guide: https://bloo.io/docs/hypercloud/syslog-connector
BQL (Bloo Query Language):
stream=authentication where sourcename='GOANYWHERE-MFT' and action='LOGIN'
| groupby user, srcip, authproto
IOC sweep for ARTEX campaign indicators: C2 host, nine proxy IPs, xcai.pro and its subdomains, and the CLAUDE.md URL across Defender, CEF/firewall, DNS and Entra sign-in telemetry. Domains match on exact host or subdomain only, connection types are limited to real connection events, and sign-ins are limited to successes and credential failures.
Detects a Slack login from a source address without prior history for that account.
A compromised Slack account gives an attacker the organisation's informal channel —
where credentials, links and internal detail are shared far more freely than in
email — and is an effective base for internal phishing.
Onboarding: SLACK is ingested via Slack native connector (Bloo extractor 833).
Integration guide: https://bloo.io/docs/integrations/slack
BQL (Bloo Query Language):
stream=authentication where sourcename='SLACK' and action='LOGIN'
| groupby user, srcip, authproto, reason
A compromised Slack account gives an attacker the organisation's informal channel —
where credentials, links and internal detail are shared far more freely than in
email — and is an effective base for internal phishing.
Onboarding: SLACK is ingested via Slack native connector (Bloo extractor 833).
Integration guide: https://bloo.io/docs/integrations/slack
BQL (Bloo Query Language):
stream=authentication where sourcename='SLACK' and action='LOGIN'
| groupby user, srcip, authproto, reason
Detects a change in authentication protocol for privileged database access brokered
by ARCOS. A downgrade to a weaker protocol for an account that previously used a
stronger one suggests either a misconfiguration that weakens the estate or an
attacker steering authentication toward a method they can intercept or replay.
Baseline the protocol per account before deployment; some legacy database clients
genuinely only support older methods.
Onboarding: ARCOS-DBEVT is ingested via RDBMS Log Shipper (Bloo extractor 225).
Integration guide: https://bloo.io/docs/integrations
BQL (Bloo Query Language):
stream=authentication where sourcename='ARCOS-DBEVT' and action='LOGIN'
| groupby authproto, user, srcip
by ARCOS. A downgrade to a weaker protocol for an account that previously used a
stronger one suggests either a misconfiguration that weakens the estate or an
attacker steering authentication toward a method they can intercept or replay.
Baseline the protocol per account before deployment; some legacy database clients
genuinely only support older methods.
Onboarding: ARCOS-DBEVT is ingested via RDBMS Log Shipper (Bloo extractor 225).
Integration guide: https://bloo.io/docs/integrations
BQL (Bloo Query Language):
stream=authentication where sourcename='ARCOS-DBEVT' and action='LOGIN'
| groupby authproto, user, srcip
Detects any authentication attempt against a Smokescreen decoy asset. Decoys are not
published in directories, hold no business function and are never referenced by
legitimate automation, so an authentication attempt is attacker activity by definition
rather than something requiring interpretation.
This is the highest-signal, lowest-noise detection available on any log source: there
is no benign path to a decoy login.
Onboarding: SMOKESCREEN is ingested via Syslog (CEF) (Bloo extractor 206).
Integration guide: https://bloo.io/docs/hypercloud/syslog-connector
BQL (Bloo Query Language):
stream=authentication where sourcename='SMOKESCREEN' and action='LOGIN'
| groupby user, srcip, authproto
published in directories, hold no business function and are never referenced by
legitimate automation, so an authentication attempt is attacker activity by definition
rather than something requiring interpretation.
This is the highest-signal, lowest-noise detection available on any log source: there
is no benign path to a decoy login.
Onboarding: SMOKESCREEN is ingested via Syslog (CEF) (Bloo extractor 206).
Integration guide: https://bloo.io/docs/hypercloud/syslog-connector
BQL (Bloo Query Language):
stream=authentication where sourcename='SMOKESCREEN' and action='LOGIN'
| groupby user, srcip, authproto
Detects and monitors sign-in events for specific high-privilege or sensitive accounts, such as break-glass accounts, within Microsoft Entra ID.
Detects RADIUS authentication activity observed in network traffic, including the
MAC address of the requesting device. Seeing RADIUS on the wire catches
authentication against network infrastructure that often logs nowhere else —
switches, wireless controllers and VPN concentrators.
Onboarding: NTA is ingested via Network Traffic Analysis (Bloo extractor 717).
Integration guide: https://bloo.io/docs/integrations
BQL (Bloo Query Language):
stream=nta-authentication-radius where sourcename='NTA'
| duration 1h
| groupby srcip, dstip, user, macaddress
MAC address of the requesting device. Seeing RADIUS on the wire catches
authentication against network infrastructure that often logs nowhere else —
switches, wireless controllers and VPN concentrators.
Onboarding: NTA is ingested via Network Traffic Analysis (Bloo extractor 717).
Integration guide: https://bloo.io/docs/integrations
BQL (Bloo Query Language):
stream=nta-authentication-radius where sourcename='NTA'
| duration 1h
| groupby srcip, dstip, user, macaddress
Detects VNC and remote framebuffer sessions, including the authentication method
negotiated. RFB is frequently deployed without strong authentication on industrial
and legacy systems, and an attacker finding one has interactive desktop access to a
machine that often cannot run an endpoint agent at all.
Onboarding: NTA is ingested via Network Traffic Analysis (Bloo extractor 717).
Integration guide: https://bloo.io/docs/integrations
BQL (Bloo Query Language):
stream=nta-authentication-rfb where sourcename='NTA'
| groupby srcip, dstip, dstport, authmethod
negotiated. RFB is frequently deployed without strong authentication on industrial
and legacy systems, and an attacker finding one has interactive desktop access to a
machine that often cannot run an endpoint agent at all.
Onboarding: NTA is ingested via Network Traffic Analysis (Bloo extractor 717).
Integration guide: https://bloo.io/docs/integrations
BQL (Bloo Query Language):
stream=nta-authentication-rfb where sourcename='NTA'
| groupby srcip, dstip, dstport, authmethod
A privileged SAP account attempted logons across multiple systems or clients from a single source IP address within a 6-hour window, potentially indicating credential compromise or lateral movement.
IOC sweep for ARTEX campaign indicators: C2 host, nine proxy IPs, xcai.pro and its subdomains, and the CLAUDE.md URL across Defender, CEF/firewall, DNS and Entra sign-in telemetry. Domains match on exact host or subdomain only, connection types are limited to real connection events, and sign-ins are limited to successes and credential failures.
Azure AD Device Code Flow Token Phishing
Cortex XDR
Detects potential device code phishing attacks by identifying OAuth token requests associated with the device code flow that originate from unmanaged or non-compliant devices. The rule specifically monitors for suspicious Cmsi (Client Managed Service Identity) activity or the use of unbound sign-in tokens within the device code flow for targeted Microsoft productivity and management applications.
Detects repeated OAuth device authorization ('device code') grant flow authentication events targeting the same user within a short window, which is indicative of an active device code phishing attempt.
Detects potential AiTM or credential harvesting campaigns where a device visits known lure or redirect paths (such as specific domains or workers.dev paths) and subsequently accesses the Microsoft device code authentication page within a 15-minute window.
IOC sweep for ARTEX campaign indicators: C2 host, nine proxy IPs, xcai.pro and its subdomains, and the CLAUDE.md URL across Defender, CEF/firewall, DNS and Entra sign-in telemetry. Domains match on exact host or subdomain only, connection types are limited to real connection events, and sign-ins are limited to successes and credential failures.
A SPNEGO authentication token was replayed against an SAP system
IOC sweep for ARTEX campaign indicators: C2 host, nine proxy IPs, xcai.pro and its subdomains, and the CLAUDE.md URL across Defender, CEF/firewall, DNS and Entra sign-in telemetry. Domains match on exact host or subdomain only, connection types are limited to real connection events, and sign-ins are limited to successes and credential failures.
Page 1 of 51