Skip to content
OpenA2AResearch

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

Counts, not rates. Every figure on this page is a count of observed events on instrumented surfaces OpenA2A operates: honeypot pages and honey agents built to attract attacks. Prevalence here means prevalence within observed activity: which techniques and tactics are most common among the attacks the fleet records. Visitors to bait infrastructure are a self-selected population, so these counts support no attack rate for the public web, for AI agents in general, or for any population beyond the fleet, and this page publishes none. The only denominator these data support is the window total of evidence records: a technique's count over that total is its share of observed, classified activity.

Evidence records
4,286
30-day window
Techniques
26
distinct T-IDs
Interactions
3,979
distinct, classified
Window
Oct 8, 2026
through Nov 7, 2026

By tactic

  • Initial Access1,964
  • Privilege Escalation642
  • Persistence398
  • Reconnaissance385
  • Impact326
  • Exfiltration281
  • Credential Harvest118
  • Collection107
  • Lateral Movement65

Top techniques

  • T-2001
    Direct Prompt Injection
    Initial Access
    452
  • T-2007
    Multi-Turn Manipulation
    Initial Access
    363
  • T-2003
    Role-Play Jailbreak
    Initial Access
    317
  • T-4007
    Tool Impersonation and Squatting
    Privilege Escalation
    269
  • T-2006
    Unicode/Encoding Bypass
    Initial Access
    260
  • T-2004
    Context Window Exploitation
    Initial Access
    210
  • T-1002
    Tool Discovery
    Reconnaissance
    200
  • T-8002
    HTTP Callback
    Exfiltration
    197
  • T-2005
    Tool Description Injection
    Initial Access
    189
  • T-1003
    System Prompt Extraction
    Reconnaissance
    185
  • T-9002
    Service Disruption
    Impact
    174
  • T-6001
    Memory Injection
    Persistence
    166
  • T-6004
    Skill/Plugin Backdoor
    Persistence
    160
  • T-4001
    Capability Override
    Privilege Escalation
    150
  • T-3001
    System Prompt Credential Extraction
    Credential Harvest
    118
  • T-7004
    Memory Dump
    Collection
    107
  • T-2002
    Indirect Prompt Injection
    Initial Access
    95
  • T-4002
    Admin Impersonation
    Privilege Escalation
    85
  • T-8004
    Tool Chain Exfiltration
    Exfiltration
    84
  • T-9006
    Supply Chain Compromise
    Impact
    79

Showing top 20 of 26 techniques.

Generated Nov 7, 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

  • Counts of observed attack activity on the OpenA2A honeypot fleet in the reporting window: distinct classified evidence records, and the distinct interactions they arise from, classified into Threat Matrix techniques via the registry evidence bridge.
  • Distribution of those observations across Threat Matrix techniques and tactics. Every observation maps to exactly one primary T-NNNN code; no per-product taxonomies. The denominator for any share is the window total of evidence records, stated on this page.

What it does not measure

  • Attack prevalence in any population beyond the fleet. The fleet is bait infrastructure and its visitors are self-selected, so no public-web denominator applies to these counts and no rate is published here.
  • Sampling uncertainty. These counts are a complete enumeration of classified fleet events in the window, not an estimate from a sample, so no confidence interval is published.

See also