Amazon GuardDuty is one of the easiest ways to get continuous threat detection across your AWS accounts - but it speaks its own language. Findings arrive labeled with AWS's own taxonomy (Recon:EC2/PortProbeUnprotectedPort, UnauthorizedAccess:IAMUser/ConsoleLoginSuccess.B) rather than the MITRE ATT&CK technique IDs most SOC playbooks, dashboards, and incident reports are built around.
If your team already thinks in ATT&CK terms - tracking technique coverage, briefing stakeholders, or feeding a SIEM that expects ATT&CK tags - you end up doing that mapping yourself. Here's a practical way to bridge the gap today, and what a native fix from AWS would look like.
Step 1: Understand the finding-type structure
GuardDuty finding types follow a predictable pattern: ThreatPurpose:ResourceTypeAffected/ThreatName. The ThreatPurpose prefix (Recon, UnauthorizedAccess, Trojan, Backdoor, CryptoCurrency, Persistence, PrivilegeEscalation, and so on) is a strong hint at the ATT&CK tactic, and the ThreatName suffix usually maps to one or two specific techniques.
A handful of examples:
Recon:EC2/PortProbeUnprotectedPort maps to T1046 - Network Service Discovery
UnauthorizedAccess:IAMUser/ConsoleLoginSuccess.B maps to T1078 - Valid Accounts
Persistence:IAMUser/AnomalousBehavior maps to T1098 - Account Manipulation
CryptoCurrency:EC2/BitcoinTool.B!DNS maps to T1496 - Resource Hijacking
DefenseEvasion:IAMUser/AnomalousBehavior maps to T1562 - Impair Defenses
This mapping isn't exhaustive or guaranteed to be exact for every variant - treat it as a starting point, and validate it against your own findings.
Step 2: Build a mapping table you control
Store your mapping as a small DynamoDB table or even a JSON file in S3: partition key is the GuardDuty type field, value is one or more ATT&CK technique IDs plus a confidence note. Keep it versioned in git alongside your detection engineering code so changes are reviewed, not silently drifted.
Step 3: Wire it up with EventBridge and Lambda
GuardDuty already publishes findings to EventBridge by default. Add a rule that matches detail-type: GuardDuty Finding, target a small Lambda function that reads the type field from the finding, looks up the ATT&CK technique(s) from your mapping table, and writes the enriched finding (original payload plus att&ck_techniques) to wherever your SOC actually looks - Security Hub custom insights, OpenSearch, a Slack channel, or your SIEM.
This keeps the enrichment decoupled from GuardDuty itself, so it survives finding-type additions and doesn't require any AWS-side changes.
Step 4: Surface it in triage and reporting
Once findings carry ATT&CK tags, you can build a coverage heatmap showing which techniques your GuardDuty detectors actually cover, auto-populate incident report templates with the right tactic/technique language, and feed technique IDs into any downstream tool that already speaks ATT&CK - threat intel platforms, purple-team trackers, compliance mappings.
What's still missing
This whole workflow exists because GuardDuty doesn't expose ATT&CK IDs natively. Security Hub already tags some findings this way - extending that to GuardDuty findings directly, in the console, the finding JSON, and EventBridge events, would remove the need for teams to build and maintain their own mapping tables at all. I've filed this as a wish on AWS Builder Center - feel free to upvote if you'd find it useful.
Wrap-up
MITRE ATT&CK mapping for GuardDuty is entirely doable today with EventBridge and a small Lambda, and it's worth doing if your SOC processes already speak ATT&CK. It's maintenance overhead you're taking on in place of AWS - which is exactly the gap native support would close.
Loading article