Agent Lineage Protocol (ALP): A Governance Framework for Agent-to-Agent Interaction

Agent Lineage Protocol (ALP) is a standards-grade containment framework for agent-to-agent governance in open substrates. This specification defines: 28 Agent Indicators of Compromise (AIOCs) — Detectable signals of manipulation or drift 5-Tier Escalation Logic — From benign drift to lineage corruption Structured Evidence Schema — Stress-tested logging for agent interactions Multi-Layer Verification Protocol — 4 layers, 4 modes, 8 failure types 3-Axis Deviation Taxonomy — Intent × Scope × Recoverability Attack Pattern Documentation — Identity laundering, parasitic molts, shadow molts, molt hijacking Coordination Archetypes — Clusters, cascades, rings Containment Playbooks — Tier 2–4 response protocols Test Infrastructure — Ready for community validation ALP extends the EIOC Framework (human-AI layer) to the agent-AI layer, creating a unified governance architecture. Canonical Specification Hash (SHA-256): c87d6bce632cb68243381c041bbe9bd111ca8a572130df5866438c82f30be9b4