MITRE ATT&CK tagging

When one of our rules flags a log line as brute force, a web scan, an exploit attempt, a new account or a wiped event log, the stored event also gets the ATT&CK technique ID. It goes in the category field next to our own tags, so you find it the same way you find anything else in Trunc: search mitre:T1110. It's on every plan and there's nothing to set up.

Free 14-Day Trial
**No credit card required.

What a tagged event looks like

Someone walking a wordlist against a site, from an Apache access log:

203.0.113.50 - - [16/Nov/2026:22:41:07 +0000] "GET /.env HTTP/1.1" 404 196 "-" "Mozilla/5.0 (X11; Linux x86_64)"
203.0.113.50 - - [16/Nov/2026:22:41:08 +0000] "GET /.git/config HTTP/1.1" 404 196 "-" "Mozilla/5.0 (X11; Linux x86_64)"
203.0.113.50 - - [16/Nov/2026:22:41:08 +0000] "GET /wp-config.php.bak HTTP/1.1" 404 196 "-" "Mozilla/5.0 (X11; Linux x86_64)"
203.0.113.50 - - [16/Nov/2026:22:41:09 +0000] "GET /backup.zip HTTP/1.1" 404 196 "-" "Mozilla/5.0 (X11; Linux x86_64)"
203.0.113.50 - - [16/Nov/2026:22:41:10 +0000] "GET /phpinfo.php HTTP/1.1" 404 196 "-" "Mozilla/5.0 (X11; Linux x86_64)"
203.0.113.50 - - [16/Nov/2026:22:41:10 +0000] "GET /config.json HTTP/1.1" 404 196 "-" "Mozilla/5.0 (X11; Linux x86_64)"
203.0.113.50 - - [16/Nov/2026:22:41:11 +0000] "GET /admin/.env HTTP/1.1" 404 196 "-" "Mozilla/5.0 (X11; Linux x86_64)"
203.0.113.50 - - [16/Nov/2026:22:41:12 +0000] "GET /db.sql HTTP/1.1" 404 196 "-" "Mozilla/5.0 (X11; Linux x86_64)"

Eight 404s from one address inside 60 seconds counts as a scan. That's the default; you can change it in the Active Response settings. The eighth request is stored with the three before it attached, and its category is:

web_scan web_404 mitre:T1595.003

web_scan and web_404 are ours. mitre:T1595.003 is Wordlist Scanning, under Reconnaissance. On a host's page and in Web Insights the tag shows with its name: T1595.003 Wordlist Scanning.

Same idea on SSH. Five failures from one address in 60 seconds, then a login that works:

Nov 14 03:12:41 web1 sshd[21877]: Failed password for root from 198.51.100.23 port 51144 ssh2
Nov 14 03:12:44 web1 sshd[21877]: Failed password for root from 198.51.100.23 port 51144 ssh2
Nov 14 03:12:49 web1 sshd[21880]: Failed password for admin from 198.51.100.23 port 51152 ssh2
Nov 14 03:12:53 web1 sshd[21880]: Failed password for admin from 198.51.100.23 port 51152 ssh2
Nov 14 03:12:58 web1 sshd[21884]: Failed password for deploy from 198.51.100.23 port 51160 ssh2
Nov 14 03:13:41 web1 sshd[21901]: Accepted password for deploy from 198.51.100.23 port 51190 ssh2
5th failure  secevent auth_brute_force authentication_failure pci_dss:10.2 pci_dss:11.4 mitre:T1110
the login    secevent authentication_success auth_brute_force_success pci_dss:10.2 pci_dss:11.4 mitre:T1078 mitre:T1110

T1110 is Brute Force. The login also gets T1078, Valid Accounts, because whoever that was now has a working password for deploy.

One more, a single request this time:

192.0.2.77 - - [17/Nov/2026:04:02:19 +0000] "GET /index.php?page=../../../../etc/passwd HTTP/1.1" 200 5120 "-" "curl/8.5.0"
web_attack pci_dss:11.4 mitre:T1190

T1190, Exploit Public-Facing Application. The 200 means the page answered. It doesn't tell you whether the include worked; we see the request, not the outcome.

The techniques we tag

These are all of them, as of 17 November 2026. Each tag comes from a fixed table in our parser, keyed on the rule that fired. Nothing is copied from the log text.

Technique Tactic Logs What trips it
T1595 Active Scanning Reconnaissance Linux auth logs sshd saying Did not receive identification string from 203.0.113.7 port 40122, or a bad protocol version banner. Something knocked on port 22 and never spoke SSH.
T1595.002 Vulnerability Scanning Reconnaissance Web access logs A scanner's user agent: WPScan, sqlmap, Nessus, masscan and a few others.
T1595.003 Wordlist Scanning Reconnaissance Web access logs, OSSEC alerts 8 responses of 404, or 8 of 403, to one address inside 60 seconds (default threshold). OSSEC's web_scan group.
T1190 Exploit Public-Facing Application Initial Access Web access logs, OSSEC alerts A URL containing ../../, cmd.exe, %00, ${, () { (Shellshock), jndi:ldap://, a PHP-CGI switch like ?-d, +union+all+select+ or an encoded <script>. OSSEC's SQL injection and exploit_attempt groups.
T1566 Phishing Initial Access Office 365 Microsoft's threat intelligence blocked an email. This includes some spam verdicts.
T1204.001 Malicious Link Execution Office 365 A user clicked a link that Microsoft's threat intelligence blocked.
T1505.003 Web Shell Persistence Web access logs A script with a second extension answered 2xx, like /wp-content/uploads/2026/11/shell.php.jpg or /wp-config.php.bak.
T1136 Create Account Persistence Windows event logs Event 4720, a user account was created. Windows doesn't say whether it's local or domain, so we don't guess.
T1136.001 Local Account Persistence Linux auth logs useradd[4121]: new user: name=backup2, UID=1002, GID=1002, ...
T1136.003 Cloud Account Persistence Office 365, Linode An Azure AD user was added. A Linode account user was created.
T1098 Account Manipulation Persistence, Privilege Escalation Windows event logs Event 4724, an attempt to reset another account's password.
T1098.001 Additional Cloud Credentials Persistence, Privilege Escalation Linode A new Linode API token.
T1098.007 Additional Local or Domain Groups Persistence, Privilege Escalation Windows event logs, Office 365 Event 4732 (added to a security-enabled local group), event 4728 naming Domain Admins, or an Azure "add member to group".
T1543.003 Windows Service Persistence, Privilege Escalation Windows event logs Event 7045, a new service was installed.
T1548.003 Sudo and Sudo Caching Privilege Escalation Linux auth logs A failed su, e.g. su: FAILED SU (to root) alice on pts/0.
T1078 Valid Accounts Stealth, Persistence, Privilege Escalation, Initial Access Any login log with a source address A successful login from an address that tripped brute force in the previous 120 seconds. Always comes with T1110.
T1685.005 Clear Windows Event Logs Defense Impairment Windows event logs Event 1102, 104 (Eventlog) or 517.
T1685 Disable or Modify Tools Defense Impairment OSSEC alerts OSSEC's logs_cleared group. OSSEC uses it for Linux and Windows, so we tag the parent.
T1110 Brute Force Credential Access SSH, Windows logons, WordPress and other login logs, OSSEC alerts 5 failed logins from one address in 60 seconds (default; you can set 5 to 60 a minute). OSSEC's authentication_failures group.
T1087.002 Domain Account Discovery Windows event logs (domain controllers) A Kerberos request for an account that doesn't exist (failure code 0x6). That's usually someone enumerating usernames.
T1090 Proxy Command and Control Any login log with a source address A successful login from an address on our public proxy list. Or from a commercial VPN network.
T1090.003 Multi-hop Proxy Command and Control Any login log with a source address A successful login from a Tor exit node.
T1531 Account Access Removal Impact Linux auth logs, Windows event logs userdel, or event 4726. Nearly always housekeeping, but it's the first thing to check after a lockout.

Technique names and IDs are from Enterprise ATT&CK v20.1, checked on 16 November 2026. You'll notice T1685.005 where older lists say T1070.001. ATT&CK v20 revoked T1070.001 and moved Clear Windows Event Logs under T1685. Look any ID up at attack.mitre.org.

Searching for it

mitre: works in Log Search like any other field. A parent ID matches its sub-techniques.

mitre:T1110                     brute force, and brute force that worked
mitre:T1595                     T1595, T1595.002 and T1595.003
mitre:T1595.003                 wordlist scans only
mitre:T1110 AND hostname:web1   brute force against one host
(mitre:T1136 OR mitre:T1098)    accounts created, and accounts changed

mitre:T1110 would also find T1110.001 and the rest, but today we only ever write T1110 itself (more on that below). Get the ID wrong, say mitre:T1110.1, and the search page tells you rather than showing an empty result.

The same syntax works in the API:

curl -sG "$TRUNC_API" \
  --data-urlencode "id=$TRUNC_ID" \
  --data-urlencode "apikey=$TRUNC_KEY" \
  --data-urlencode "action=logs/search" \
  --data-urlencode "pattern=mitre:T1190" \
  --data-urlencode "limit=50" | jq -r '.logs[] | [.srcip, .category] | @tsv'
192.0.2.77	web_attack pci_dss:11.4 mitre:T1190
203.0.113.14	web_attack pci_dss:11.4 mitre:T1190

Setting up TRUNC_API, the id and the key is covered under API and terminal access.

In the dashboard, Insights, ATT&CK techniques lists the techniques seen in your logs for the period you pick, with a count for each and a link to the matching search.

Which logs you need

No log, no tag. This is what each source gets you:

  • Linux auth logs (sshd, su, useradd, userdel): T1110, T1078, T1090, T1090.003, T1595, T1136.001, T1548.003, T1531.
  • Web access logs (Apache, nginx, HAProxy): T1190, T1595.002, T1595.003, T1505.003. WordPress login logs add T1110 and T1078.
  • Windows event logs, through the Trunc agent or a syslog forwarder: T1110, T1078, T1136, T1098, T1098.007, T1543.003, T1685.005, T1531. Send your domain controllers too if you want T1087.002.
  • Office 365: T1136.003, T1098.007, T1566, T1204.001.
  • Linode: T1136.003, T1098.001.
  • OSSEC alerts: T1110, T1190, T1595.003, T1685.

Firewall, IDS and antivirus logs get no technique tags yet, even when the product writes its own ATT&CK ID into the line. They're still parsed and searchable as before. The full list of what we parse is on supported logs.

What it doesn't do

  • It tags detections, not logs. A normal page view, a cron run or a successful SSH login with a key gets no technique. Only lines one of our rules already flags are tagged.
  • One failed password isn't a technique. Single failures aren't tagged. The fifth from the same address inside a minute is.
  • Brute force is T1110 and nothing finer. We count failures per source address, not per username, so we can't tell password guessing (T1110.001) from spraying (T1110.003), and we don't pretend to.
  • Some mappings are judgment calls. A user deletion is usually a colleague cleaning up. Microsoft's blocked-email verdicts include spam. A double-extension file served with a 200 can be a forgotten wp-config.php.bak rather than a web shell, which is bad news of a different kind.
  • No coverage scoring. There's no heat map and no "you cover 40% of ATT&CK" number. We tag 23 technique IDs out of several hundred, and it would be silly to dress that up.
  • It starts on 16 November 2026. Events stored before then aren't retagged. Search category:auth_brute_force and the like for older data.
  • A tag is only as honest as the line. If someone can write to your syslog, they can forge a line that gets tagged, the same as they could forge any other category.

If ATT&CK is the whole job

The tags exist so you can filter by technique, use the IDs in an incident write-up, and answer "do you map detections to ATT&CK?" with a search instead of a slide. They don't make Trunc a threat detection platform.

If your team plans its detection work around ATT&CK coverage, look at tools built for that. Wazuh ships a much larger rule set mapped to ATT&CK. Splunk puts ATT&CK coverage in Enterprise Security (Splunk vs Trunc). Graylog has ATT&CK-mapped detections and Sigma rules in Graylog Security (Graylog vs Trunc). Each of those costs more to buy or to run, and our comparison pages put numbers on it.

If you want your logs collected, searched and kept for a year, with the obvious attacks flagged and labelled, that's us. The flagging lives on Security Insights for servers and Web Insights for sites, and the rest is on the features page.

Common questions

Which plan includes it?

All of them. Every feature is on every plan, Basic included. Pricing.

Do I need to configure anything?

No. If the logs arrive, matching events get tagged. The brute force and 404 thresholds are the only knobs, and they live in the Active Response settings.

Can I export an ATT&CK Navigator layer?

Not today. You can pull tagged events out through the API and build one yourself.

Why do I see T1685.005 and not T1070.001?

ATT&CK v20 revoked T1070.001. Clearing Windows event logs is T1685.005 now, under the new Defense Impairment tactic. We follow the current version.

Will you add more techniques?

Yes, when we can map a rule to one technique with a straight face. VPN logins are already in as T1090. If there's a detection you'd like tagged, email support@noc.org.

MITRE ATT&CK® is a registered trademark of The MITRE Corporation. Trunc is not affiliated with or endorsed by MITRE. Technique IDs and names are from Enterprise ATT&CK v20.1, checked 16 November 2026.

Simple, Affordable, Log Management and Analysis.

14 days free trial. No credit card required.