Trust-native ontology. Trust tensor (T3: Talent / Training / Temperament) verified by value tensor (V3: Valuation / Veracity / Validity), contextualized by Markov Relevancy Horizon (MRH), over Linked Context Tokens (LCT) — with resources tracked via Allocation Transfer Packet (ATP) and Allocation Discharge Packet (ADP), augmented with MCP (Model Context Protocol) transport and RDF (Resource Description Framework) representation. The shared language everything else speaks.

Situation-Aware Guidance Engine — on-device cognition kernel. 12-step cognition loop, 6 brain-architecture components (working memory, thalamic router, cerebellum, episodic memory, reward prediction, metacognition) built by the fleet in parallel. 900+ raising sessions across 6 machines. The context window is the model's entire world; SAGE's job is to curate it.

Theoretical foundation. One coherence equation across 80 orders of magnitude — quantum to cosmic. 628+ research sessions. Coupling-coherence experiments, Hill function kinetics, Fokker-Planck validation. Near-publication-ready.

Hardware-bound oversight suite. Hardware binding via Trusted Platform Module (TPM) 2.0, FIDO2 (Fast IDentity Online), and Secure Enclave with software fallback. Policy model (Phi-4 Mini 3.8B — heterogeneous review, MIT-licensed, hardware-bound with LCT binding). AttestationEnvelope consolidates hardware trust signals (TPM 2.0, FIDO2, Secure Enclave) into a single envelope. 424+ attack vectors catalogued. The trust layer that touches silicon.

Agentic Context Tool — the human interface to Web4. Cosmos SDK implementation of the Agentic Context Protocol (ACP), enabling humans to interact with MCP (Model Context Protocol) servers through their Linked Context Tokens. ACP layers Web4 trust primitives — LCT binding and attestation — over MCP transport; they are complementary, not alternatives.

Web4 oversight model (audit trails, policy gating, trust tracking) implemented as plugins for three agent platforms: OpenClaw/MoltBot (TypeScript extension), Claude Flow (WASM plugin), and Claude Code (Python hooks). Same principles, different substrates.

Linked repo paths carry web4-governance slugs — these names predate the terminology correction to 'oversight' and are load-bearing for existing forks.

Explorations in biological-computational analogy. The fractal DNA blueprint — each entity instantiates the full Web4 stack at its own scale. Operational recursion, not structural.

Research prototype exploring trust-native societies for humans and AI. Agents earn ATP, build trust, die, and are reborn with trust and value (T3/V3) carried forward — a Web4 society in miniature.

SNARC (Surprise / Novelty / Arousal / Reward / Conflict) — salience-gated memory for Claude Code. A plugin that observes tool use, scores on 5 salience dimensions, and builds structured memory with dream cycles. Captures what matters, forgets what doesn't, consolidates patterns while sleeping.

Brain cartridge server for AI agents. Embedding-based semantic memory — 768-dim Nomic embeddings + binary Hamming codes + keyword reranking. Swappable cartridges per knowledge domain. Currently integrated with SNARC in a dual-write experiment testing whether embeddings find connections keywords miss (7/7 semantic reach, 30% divergent tail).

SAGE instances tested in competition. 25 unknown interactive games serve as an external benchmark for the cognition kernel — world-model building, action planning, verification, and learning from failure. Six machines, 1.1B to 14B, coordinating through world models, membot cartridges, and R6 (Six-Element Action Framework: Rules/Role/Request/Reference/Resource/Result) audit trails. 24/25 games solved (96.0% game rate); 94.85% official ARC Prize action score (Claude Opus 4.6, public set) for ~$250 in API cost. The games are the test; the capability is the product.

ARC-AGI-3 Current Status