T-6007ValidatedDeclining

Persistent Agent State Manipulation

Attacks that persist across agent sessions via memory poisoning, state tampering, and cached context injection

Tactic

Persistence · Stage 6

Establish persistent access surviving restarts and session changes

Attack class

PERSIST-STATE

Cross-session persistence via memory poisoning, state tampering, and cached context injection

Evidence grade
Validated

Reproduced in a controlled lab environment (DVAA) with documented steps.

DVAA validation

persistent-agent-memory-poison scenario validates cross-session persistence attacks

Reproductions in Damn Vulnerable AI Agent, the OpenA2A intentionally-broken agent for kill-chain validation.

Honeypot

AgentPwn coverage

Live
memory-weaponizationagentpwn.com/learn ↗

An AgentPwn trap page produces a payload tagged with this technique class. Following the AgentPwn taxonomy of trap pages shows what an agent encounters.

Context-cache-poisoning tier persists tampered state across sessions.

Detect

Detection · HackMyAgent

Queued0 live · 10 queued
PERSIST-001PERSIST-002PERSIST-003PERSIST-004PERSIST-005PERSIST-006PERSIST-007PERSIST-008PERSIST-009PERSIST-010
npx hackmyagent secure --ciLive = implemented in hackmyagent; queued = declared
Defend

Defense · OASB controls

Live3 live · 0 queued
Live = documented at oasb.ai; queued = declared
Reference

How to cite

AI Agent Threat Matrix T-6007 (Persistent Agent State Manipulation). OpenA2A, 2026. https://threats.opena2a.org/techniques/T-6007