Fetching from the wire…
01
02
03
04
05
06
07
08
Public story · 2026-03-17 · source-backed
The first formal security taxonomy purpose-built for AI agents. Covers unsafe tool invocation, memory manipulation, identity spoofing across A2A boundaries, and inadequate human-in-the-loop controls. This is the vocabulary your threat models should use now. Source
Each link below shares sources, entities, or timing with this story.
CrewAI supports A2A / Shared entity: A2A / What happened next
Linked by a graph relationship (CrewAI supports A2A); both cover A2A; picks up the A2A thread on 2026-06-07.
Shared entity: A2A / Shared topic / What happened next / Tension
Both cover A2A; overlapping topics (agent, identity); picks up the A2A thread on 2026-08-14.
Both cover A2A; overlapping topics (application, control); picks up the A2A thread on 2026-07-28.
Both cover A2A; overlapping topics (agent, agentic); picks up the A2A thread on 2026-06-17.
CrewAI supports A2A / Shared entity: A2A / Earlier coverage
Linked by a graph relationship (CrewAI supports A2A); both cover A2A; earlier A2A coverage from 2026-02-12.
Shared entity: A2A / Shared topic / Earlier coverage / Tension
Both cover A2A; overlapping topics (agent, agentic); earlier A2A coverage from 2026-03-03.
Shared entity: A2A / Shared topic / Earlier coverage
Both cover A2A; overlapping topics (agent, agentic, application); earlier A2A coverage from 2026-03-03.
Both cover A2A; overlapping topics (agent, agentic, identity); earlier A2A coverage from 2026-03-02.