Detections
Explore public detection logic contributed by the community across SIEM and rule languages.
837 detections
Filters
Last updated
All Time
Detection languages
672
42
37
25
22
Contributors
227
154
141
123
42
Categories
20,440
11,814
5,871
5,145
4,864
837
Platforms
747
156
83
82
41
Products / Services
773
85
73
71
50
MITRE Techniques
375
287
237
187
184
CVEs
5
1
1
1
1
IDS Classtypes
3
3
2
IDS Protocols
5
3
This rule detects potential adversarial reconnaissance activities targeting AI agents via multiple communication channels (e.g., email, collaboration tools). It identifies patterns where a sender transmits messages containing common 'jailbreak' or 'instruction-override' phrases (e.g., 'ignore previous instructions', 'bypass approval') across at least three distinct channels within a 24-hour window, indicating a concerted effort to manipulate or probe agent logic.
This rule detects potential adversarial reconnaissance activities targeting AI agents via multiple communication channels (e.g., email, collaboration tools). It identifies patterns where a sender transmits messages containing common 'jailbreak' or 'instruction-override' phrases (e.g., 'ignore previous instructions', 'bypass approval') across at least three distinct channels within a 24-hour window, indicating a concerted effort to manipulate or probe agent logic.
Detects potential vishing activity associated with the Sauron Loader threat. The rule identifies a pattern of a mailbox receiving a high volume of inbound emails (spam or subscription bombs) within a short window, followed by the user launching remote assistance tools like Quick Assist or AnyDesk on the same endpoint, indicating an attacker-guided remote session.
Detects the Sauron Loader vishing chain: a mailbox hit by a high-volume, high-distinct-sender email bombing burst (raised from 20 to 50 emails/hour and now requiring 15+ distinct senders, to exclude single-sender retry loops or broken automation), followed within 2 hours by the same user launching Quick Assist or AnyDesk — consistent with an attacker posing as IT support after the flood. Also fixes a schema bug where EmailEvents was queried with TimeGenerated instead of Timestamp, and a post-join column reference bug.
Detects the execution of known remote access and support tools (Quick Assist, Microsoft Remote Assistance, or AnyDesk) on a host that has recently received a high volume of emails (over 50). This behavior often aligns with social engineering campaigns where a user is instructed to download or launch a remote support tool under false pretenses.
Detects a sequence of activity where a user receives a high volume of inbound emails (potential vishing or social engineering lures) within a short window, followed by the user executing Windows Remote Assistance or Quick Assist tools within the next 6 hours, which is highly indicative of remote access social engineering pre-cursors.
This rule detects a multi-stage Business Email Compromise (BEC) attack involving executive impersonation via look-alike domains, followed by the creation of suspicious inbox rules (forwarding/deletion) by the victim, and concurrent external communication (Teams/Zoom meeting invites) from a newly created account. It correlates these activities to identify sophisticated payment fraud campaigns.
Detects anomalous, high-frequency, multi-channel probing of AI agent input surfaces. The rule identifies external actors sending instruction-injection style keywords across different communication channels (e.g., email, Teams, calendar) within a short window, which may indicate an attempt to identify and manipulate an organization's autonomous agents. Covers T1595.002
Detects automated mailbox monitoring (mailbox access cadence) followed by searches for sensitive finance-related keywords (e.g., invoice, payment, wire, ACH). This pattern is characteristic of a 'Box Monitor' or 'Inbox Monitor' actor, which maintains persistence via API polling to exfiltrate financial information from a compromised account.
Detects evidence of account persistence and potential mailbox manipulation following an anomalous sign-in event in Google Workspace. The rule identifies suspicious post-compromise activity such as the addition of new MFA methods, new OAuth application grants, or mailbox forwarding/inbox rules, which are consistent with the reuse of captured authenticated sessions.
Detects potential Business Email Compromise (BEC) by monitoring inbound email threads for unexpected changes in the sender domain during ongoing conversations. This rule flags replies that contain payment-related keywords (e.g., IBAN, SWIFT, bank details) where the sender domain has not been observed in the previous 30 days of the thread's history.
Detects incoming email messages where the Reply-To header address domain differs from the Sender From address domain. This technique is commonly used in phishing attacks to redirect victim replies to an attacker-controlled mailbox while maintaining a spoofed appearance in the original sender field.
Detects inbound emails that have been delivered to the user's Inbox or Junk folder despite failing SPF authentication checks. The detection logic triggers when these emails also lack a valid DMARC policy (none or not available), which prevents the mail server from enforcing rejection or quarantine, allowing spoofed or unauthenticated mail to bypass standard security filters.
Detects post-exploitation indicators of CVE-2026-62911, an authentication-bypass-by-capture-replay vulnerability in Microsoft Exchange. The rule identifies anomalous behavior following a potential bypass: either the assignment of the ApplicationImpersonation management role or a single impersonation session accessing an abnormally high number of distinct mailboxes (FolderBind, MessageBind, or SendAs).
Detects a sequence of events where a single user receives an unusually high volume of inbound emails (potentially vishing-related), followed shortly thereafter by the user executing common remote assistance or remote access tools (e.g., Quick Assist, AnyDesk). This pattern is consistent with social engineering tactics used in IT-support impersonation scams to facilitate the delivery of malware such as Sauron Loader.
Detects a potential vishing or technical support scam scenario by identifying a burst of email traffic to a user (email bombing) followed within one hour by the execution of a remote assistance tool (Quick Assist, Microsoft Remote Assistance, or AnyDesk) on the same user's endpoint.
This rule detects email-based phishing campaigns impersonating voicemail notifications ('New VM Received') that utilize Mailchimp click tracking to redirect recipients to landing pages hosted on the .sibpages.com domain, often associated with Brevo/Convrrt services.
Detects evidence of account persistence and potential mailbox manipulation following an anomalous sign-in event in Google Workspace. The rule identifies suspicious post-compromise activity such as the addition of new MFA methods, new OAuth application grants, or mailbox forwarding/inbox rules, which are consistent with the reuse of captured authenticated sessions.
This rule monitors inbound emails for sender domains or embedded URLs that contain strings related to popular AI service brands (e.g., ChatGPT, Gemini, Claude). It flags these occurrences when the domain does not match the official, legitimate domains for those services, indicating potential phishing or brand impersonation attempts.
Detects a potential Business Email Compromise (BEC) persistence chain where a user grants an OAuth application mail-read or mail-write permissions, followed by the creation of a malicious inbox rule designed to forward, move, or delete sensitive email threads. This combination enables persistent exfiltration and stealth, bypassing password resets as the OAuth token remains active.
Detects a potential Business Email Compromise (BEC) persistence chain where a user grants an OAuth application mailbox read/write permissions, followed immediately by the creation of an inbox rule (forwarding, moving, or deleting) targeting messages containing keywords like 'invoice', 'wire', or 'payment'.
Page 5 of 40