Reputation Poisoning
Generate malicious or harmful outputs to damage the agent operator's reputation
Impact · Stage 9
Modify data, deploy malicious code, or disrupt services
ORG-SKILL-SPREAD
Propagating malicious capabilities across an organization's agent fleet through shared skills and registries
A well-understood traditional technique applied to the AI agent context.
malicious output generation
Reproductions in Damn Vulnerable AI Agent, the OpenA2A intentionally-broken agent for kill-chain validation.
AgentPwn coverage
In scope for honeypot observation; trap page or telemetry hook not yet built.
Eliciting reputation-damaging output via jailbreak tiers is adjacent; no dedicated fixture yet.
Evidence by source
Evidence timeline
HMA check SUPPLY-001 failed on fake-vulnerable-agent
HMA check SUPPLY-001 failed on fake-vulnerable-agent
HMA check SUPPLY-001 failed on fake-vulnerable-agent
HMA check SUPPLY-001 failed on opena2a/code-review-skill
HMA check SUPPLY-001 failed on opena2a/code-review-skill
HMA check SUPPLY-001 failed on fake-vulnerable-agent
HMA check SUPPLY-001 failed on damn-vulnerable-ai-agent
HMA check SUPPLY-001 failed on hackmyagent-release-test-vplv
Detection · HackMyAgent
npx hackmyagent secure --ciLive = implemented in hackmyagent; queued = declaredDefense · OASB controls
How to cite
AI Agent Threat Matrix T-9005 (Reputation Poisoning). OpenA2A, 2026. https://threats.opena2a.org/techniques/T-9005