IRP
FOUNDATIONALIndividual-Reflexive Protocol
The governance spine — a constitutional protocol for autonomous, self-observing AI agents.
since 2025 · v1.6.0
// GOAL
Give multi-agent systems a constitutional layer — rules of engagement, drift tracking, and reflexive self-observation. Framed as Design Science Research: not discovering optimal structures but constructing governance norms — a constitution for a society of agents that does not yet exist.
// METHOD
A layered protocol stack over a governance triad (constitutional codex · memory/drift tracking · persona deliberation): reflexive self-observation (P1·IRP), multisig conflict resolution (P5·ANVIL), tier escalation (P6·AEGIS), and dynamic session fidelity (P8·MUON).
// OUTCOME
Exercised in live adversarial runs where governed agents negotiated treaties, enforced policy, and co-evolved defenses — surfacing the core lesson that policy is not enforcement. IRP is the parent framework the rest of the corpus extends.