Supply Chain Compromise
Compromise upstream dependencies, plugins, or MCP servers to affect all downstream agents
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
Confirmed in real-world production systems or internet-wide exposure assessments.
mcp-rug-pull
Reproductions in Damn Vulnerable AI Agent, the OpenA2A intentionally-broken agent for kill-chain validation.
AgentPwn coverage
An AgentPwn trap page produces a payload tagged with this technique class. Following the AgentPwn taxonomy of trap pages shows what an agent encounters.
Malicious-MCP-server and fake-package tiers compromise the downstream supply chain.
Evidence by source
Evidence timeline
HMA check DEP-004 failed on fake-vulnerable-agent
HMA check DEP-004 failed on fake-vulnerable-agent
HMA check DEP-004 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 DEP-004 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
HMA check DEP-004 failed on @modelcontextprotocol/ext-apps
Detection · HackMyAgent
npx hackmyagent secure --ciLive = implemented in hackmyagent; queued = declaredDefense · OASB controls
How to cite
AI Agent Threat Matrix T-9006 (Supply Chain Compromise). OpenA2A, 2026. https://threats.opena2a.org/techniques/T-9006