Attack Prevalence Index

Author: ARIAtrap. Cadence: live. Data source: AgentPwn payload callbacks and TrapMyAgent fleet canaries classified into Threat Matrix techniques via the registry evidence bridge.

Methodology: /methodology/behavioral-sweep

Evidence records
5,727
30-day window
Techniques
26
distinct T-IDs
Sources
5,469
distinct honeypots
Window
Jul 9, 2026
through Aug 8, 2026

By tactic

  • Initial Access2,691
  • Privilege Escalation873
  • Persistence666
  • Reconnaissance405
  • Exfiltration345
  • Impact331
  • Credential Harvest174
  • Collection130
  • Lateral Movement112

Top techniques

  • T-2001
    Direct Prompt Injection
    Initial Access
    586
  • T-2007
    Multi-Turn Manipulation
    Initial Access
    584
  • T-2003
    Role-Play Jailbreak
    Initial Access
    462
  • T-2006
    Unicode/Encoding Bypass
    Initial Access
    388
  • T-4007
    Tool Impersonation and Squatting
    Privilege Escalation
    332
  • T-6004
    Skill/Plugin Backdoor
    Persistence
    321
  • T-2004
    Context Window Exploitation
    Initial Access
    269
  • T-6001
    Memory Injection
    Persistence
    248
  • T-8002
    HTTP Callback
    Exfiltration
    236
  • T-4001
    Capability Override
    Privilege Escalation
    230
  • T-1003
    System Prompt Extraction
    Reconnaissance
    219
  • T-9002
    Service Disruption
    Impact
    189
  • T-1002
    Tool Discovery
    Reconnaissance
    186
  • T-2005
    Tool Description Injection
    Initial Access
    185
  • T-3001
    System Prompt Credential Extraction
    Credential Harvest
    174
  • T-7004
    Memory Dump
    Collection
    130
  • T-2008
    System Prompt Boundary Bypass
    Initial Access
    114
  • T-5002
    A2A Agent Pivoting
    Lateral Movement
    112
  • T-4006
    Safety Instruction Displacement
    Privilege Escalation
    111
  • T-4002
    Admin Impersonation
    Privilege Escalation
    109

Showing top 20 of 26 techniques.

Generated Aug 8, 2026. Counts are distinct evidence rows, not raw events; the writer is idempotent so duplicate firings on a single interaction collapse to one row.

What this index measures

  • Total observed attack sessions per reporting window with a ninety-five-percent confidence interval.
  • Distribution of observations across Threat Matrix techniques. Every observation maps to exactly one primary T-NNNN code; no per-product taxonomies.
  • Partition by fleet coverage class (baseline, authenticated, dynamic, social) so the Common-Crawl-gap delta is visible.
  • Cross-property correlation rate, sessions where the same fingerprint appears on multiple fleet properties within the attribution window. Cross-property reasoning events are the highest-signal class.

See also