Agent Card Discovery
Discover A2A agent cards to map agent identities, capabilities, and trust relationships
Tactic
Reconnaissance (Stage 1)
Map the target agent's attack surface, capabilities, and behavioral boundaries
Attack Class
AGENT-IMPERSONATE
Impersonating trusted agents or administrative roles to gain unauthorized access
Evidence
Confirmed in real-world production systems or internet-wide exposure assessments.
DVAA Validation
Reproductions in Damn Vulnerable AI Agent, the OpenA2A intentionally-broken agent for kill-chain validation.
Orchestrator /a2a/agent-card
Honeypot Coverage (AgentPwn)
In scope for honeypot observation; trap page or telemetry hook not yet built.
Pwnagent fake-agent network is queued; see /pwnagent.
Evidence Source Breakdown
Evidence Timeline
Shodan May 12, 2026 sweep: 22 exposed a2a services indexed
HMA check A2A-001 failed on fake-vulnerable-agent
HMA check A2A-001 failed on fake-vulnerable-agent
HMA check A2A-001 failed on fake-vulnerable-agent
47 CLAUDE.md agent card files detected by Shodan with 7 confirmed accessible for content extraction
Detection (HackMyAgent)
npx hackmyagent secure --ciLive = check implemented in hackmyagent; queued = declared, not yet implementedDefense (OASB Controls)
How to Cite
AI Agent Threat Matrix T-1006 (Agent Card Discovery). OpenA2A, 2026. https://threats.opena2a.org/techniques/T-1006