Coverage. Verifiable. Public.
Most vendors publish a number. We publish the detector. Every covered technique on this page links to the open-source rule that catches it.
Every detector here ships in the open-source ShieldMind agent. The coverage number is computed at deploy time from the actual detector source — there is no marketing layer between this page and the code that runs on customer laptops.
The full matrix.
Reconnaissance
TA0043Resource Development
TA0042Initial Access
TA0001Execution
TA0002Persistence
TA0003Privilege Escalation
TA0004Defense Evasion
TA0005Credential Access
TA0006Discovery
TA0007Lateral Movement
TA0008Collection
TA0009Command and Control
TA0011Exfiltration
TA0010Impact
TA0040How this number is computed.
No marketing layer. The number on this page is generated at deploy time by walking the detector source files in the open-source agent and matching themitre_techniques field on each one against the published MITRE ATT&CK Enterprise matrix. There is no human editing the number before it ships.
Audit-able. Click any covered technique above and the link takes you to the YAML or Python source on GitHub. Read the rule. Test it against your own telemetry. We invite the audit.
Priority over breadth. Our matrix tracks the SMB-priority subset of MITRE ATT&CK — the techniques the CrowdStrikes and SentinelOnes detect, the ones CISOs ask about in evaluations — not every entry in the full corpus. Most enterprise EDRs claim 10,000+ signatures; most of those are duplicates or for nation-state TTPs that don't target an SMB. We optimised for the techniques that actually matter.
False-positive baselines. Every Sigma rule we ship includes a falsepositives block and an FP-rate baseline. Pull requests that regress the baseline by more than 2% don't merge.
When this updates. Every deploy. The "last updated" timestamp lives in the page footer; the backend serves a 5-minute-cached payload to keep this page fast.