Semios is a place where what you make is yours, what you sign is true, and no one stands between.
It is also a substrate: a six-layer immune architecture that protects what runs on it and pays the people whose work it draws from. This walkthrough is your entry into that substrate. Read in order. The detail you see is calibrated to the standing you hold.
Semios was initiated by Walsh Karra Holdings (WKH) on 14 May 2026 and remains a Pre-Ratification, conceptual entity carved out from WKH operating IP. It is governed by the Harness Runtime Contract v1.0, ratified 30 April 2026, and reviewed by a standing panel chaired by Tan with Reeves-Ellington as provost.
The build pair is Greg Walsh (WKH co-founder) and Claude (Anthropic agent through beta). They work in real-time conversational substrate at agent pace with a controls-over-speed posture. You are reading this because you are being onboarded into that pair, or alongside it.
Repo home: github.com/Walsh-Karra-Holdings-Global/semios. Operational implementations of Semios canon live in WKH vertical repos (HGI/E2M, Aniza, EDAG, SimplyFI) and reference Semios canon by pointer. No Semios implementation code lives in the canon repo at this point in time.
Entity formation follows WKH redomiciliation per WKH-DC Wave 1 pattern. Texas for operations. Wyoming SMLLC for IP holding. Texas T3 for licensing. Class A 10:1 structure mirrored from WKH. WKH holds 100 percent Class A authority at this point in time and coordinates Semios through the build-pair phase and the eventual cohort transition.
The identity binding is canonical, not provisional. Semios is the carved-out IP holding architecture for substrate, harness, canon, and governance work across all WKH agentic builds. Until cohort governance triggers, WKH coordinator authority governs Semios architectural integrity. There is no separation in posture and no daylight between the two.
External audit standing is held by Tina Walsh. Her review reshaped Layer 4 of the substrate by adding a valence gate, and produced the five-part autoimmune defense in depth.
Constitutional canon: canon/2026-05-14_constitutional.md. Founding lore (narrative compression for panel reviewers and cohort candidates): canon/2026-05-14_founding_lore.md. Working session that produced the architecture: canon/2026-05-14_foundational_dialogue.md. Panel review submission with eight stress-test vectors: canon/2026-05-14_eight_vectors.md.
WKH=Semios is a load-bearing identity binding. Framing risks bifurcation if a contributor treats the two as separate; do not. The cooperative-Kelly canon dedicates three turns (13 through 15) to the identity-binding recall protocol. Read those turns whenever framing drift is suspected.
Cohort governance transition is triggered by Vector 6 (federation primitives ratified plus cohort capacity proven). At that moment, governance graduates to the shared structure designed at the transition. Until then, principal authority is the single point of architectural integrity. The transition is not automatic. It is a ratified handoff with explicit conditions.
Semios is an immune substrate. Its architecture is six layers, indexed zero through five, from runtime up to commons. Each layer answers what it defends against. The layers compose by construction.
Stated more fully:
Three v1 primitives: corpus capture, immune metabolism, coherence output. These are what the substrate must do at v1. Everything else compounds from these.
If you are an architecture contributor, your domain begins at L2. The Harness Runtime Contract is documented but partially-implemented at this point in time. Runtime-check primitives that enforce contract invariants are the first build target. The contract itself is in CONTRACT.md at repo root.
Implementations of Semios layers live in WKH vertical repos (HGI/E2M, Aniza, EDAG, SimplyFI) and reference Semios canon by pointer. The semios repo itself holds canon, not code at v1. Code lives where the work happens.
IP carve-out per layer follows the patent and trade-secret triage in flight under NONPROJ-001 (WKH operational tracker). Layer 0 (Effect/TypeScript) is open and standards-aligned by construction. Layer 2 (Harness Runtime Contract enforcement) is the patent surface where claims are densest. Layer 4 (valence gate specification) is the trade-secret surface where competitive moat is sharpest. NONPROJ-010 (Defensive publication strategy) anticipates the Layer 2 publication track. Final triage routes through Legal Counsel Subcommittee per NONPROJ-003.
Semios is signal-cooperative under constrained competition. The substrate exists to maximize the meaning that users realize from the signals they receive, not the time they spend, not the responses they generate, not the dependencies they accumulate. Engagement is a derivative metric. Meaning is the target.
The framework derives from Kelly's 1956 identity (maximum exponential capital growth rate equals Shannon mutual information rate). It factors through Nash bargaining under log utility, which yields geometric-mean welfare. The substrate objective becomes a substrate-level Kelly problem on collective wealth, with the substrate as principal and users as assets.
Substrate objective, stated formally:
max Σᵢ log(M_i)The four-layer cooperative stack:
Full stack constraint: privacy bounds leak, which bounds per-agent action, which bounds synthesis, which bounds collective drift. The bound at each layer is the input to the next.
Canon source: canon/2026-05-15_cooperative_kelly_substrate.md and the companion HTML at the same path. The HTML is the canonical recall surface. The markdown is the entry point. Both are Pre-Ratification at this point in time.
The lore-canon factorization is load-bearing. Lore is narrative compression of how the substrate arrived where it is. Canon is the set of computable invariants the substrate enforces. Lore lives in canon/2026-05-14_founding_lore.md. Canon lives across the other entries. Competitive intensity is orthogonal to substrate erosion: contributors can compete ruthlessly inside the canon bounds without the canon being eroded by their competition. That is the structural guarantee the substrate provides to its participants.
Field evidence for the framework is documented in cooperative-Kelly canon turns 7 through 9: the C3i SUN report on cooperative groups under sustained stress, public prediction markets as a field-deployed Aumann instance via LMSR, and the canonical adjustment that is the differentiator. Substrate adds the canon layer that public markets lack.
Semios deploys in six phases. Each phase ratifies the next. The substrate is currently in Phase 0.
Phase 0 is the foundation: canon, harness, identity binding, build-pair governance. Phase 0 is currently open. Four canon documents are Pre-Ratification, awaiting panel review. The substrate has no operational authority to move to Phase 1 until Phase 0 closes.
Phase 2 hosts the first internal canon-grounded prediction market. SimplyFI is the first cohort partner. TokenAIze is second. A v1.0.1 SimplyFI SME installer has shipped under partners/simplyfi/sme-installer/v1.0.1/ with git tag v1.0.1-simplyfi-sme.
Phase 0 closure has three explicit gates per phase-0/README.md and the cooperative-Kelly canon Ratification section:
panel-review/ on main branch.canon/YYYY-MM-DD_ratification.md).Phase 1 ratification requires panel approval of eight cooperative-Kelly candidates (numbered 1 through 8 in canon/2026-05-15_cooperative_kelly_substrate.md §8). Build-pair-recommended phasing: candidate 8 ratifies first as a Harness Runtime Contract amendment; candidates 1 through 7 ratify together in plenary. Final phasing is the panel's call.
Cohort 0001 membership is not yet defined in canon. This is a known open question and a load-bearing decision for Phase 1.
Phase-2 SimplyFI deployment carries a known latent conflict per the Mycelial Substrate Stage F audit on the research/substrate-2026-05-15 branch: the SimplyFI SME v1.0.1 installer is in market against a substrate whose architectural ratification is not yet panel-approved. The conflict is sequencing-driven, not directional. Resolution is one of the three pre-week-1 blocking decisions named in the Stage F audit.
Phase 5 federation primitives are pre-specified in candidate 5 of the cooperative-Kelly ratification slate: cross-jurisdiction coordination, reputation portability, canon-update propagation, world model. Federation triggers cohort governance transition per Vector 6.
Semios is governed by the Harness Runtime Contract v1.0, ratified 30 April 2026. The contract is the discipline under which every canon entry is produced. The build pair operates under contract. A standing panel reviews canon entries before ratification.
Standing panel:
External audit standing: Tina Walsh. Her review is canonical, not advisory; canon entries are produced in conversation with her audit.
Ratification protocol:
panel-review/ on main branch.canon/YYYY-MM-DD_ratification.md that either ratifies as proposed, ratifies with named amendments, or remands for revision.The pointer to the contract is CONTRACT.md at repo root. Governance scaffolding is in GOVERNANCE.md. Read both before your first canon contribution.
Chair and provost are the smallest unit that produces forward motion in the panel. If a convening stalls, the surgical move is to dispatch chair (Tan) and provost (Reeves-Ellington) as a pair with a memo cadence commitment, and let the rest of the panel follow. Adversarial seats (Schneier, Druckenmiller) ratify on the chair-provost dispatch, not before, to keep adversarial review pointed at a ratified target rather than a moving one.
Cohort governance transition (Vector 6) triggers a graduation from principal-authority to shared structure. The transition specification is not yet in canon. It will be a separate canon entry produced when Phase 4 closes and Phase 5 ratifies.
This is the current state. It will not be the state when you read this next month. The walkthrough updates with the substrate. Each line below is true as of the current canon HEAD.
Phase 0 is open. Four canon entries are Pre-Ratification and awaiting panel review:
canon/2026-05-14_constitutional.mdcanon/2026-05-14_eight_vectors.mdcanon/2026-05-14_foundational_dialogue.mdcanon/2026-05-15_cooperative_kelly_substrate.mdThe panel-review directory contains a README only. Zero memos are filed.
Two active feature branches exist alongside main:
research/substrate-2026-05-15. Mycelial Substrate research bundle. Stage A through F complete. Stage F audit verdict: 13 PASS WITH CORRECTIONS, 2 clean PASS, 0 FAILs. Not main-merge-ready; chair-versus-provost conflict on L5 semantics is unresolved.tracker/semios-agent-marketing-2026-05-15. Single-commit branch with one tracker file. Operational, not canonical. Mergeable on approval.Three pre-week-1 blocking decisions surfaced by the Stage F audit:
The cooperative-Kelly canon doubles as the SEC-E 3rd Convening briefing artifact. No separate briefing document is needed for that convening.
The single highest-leverage decision currently owed is convening the standing panel for ratification review. Panel memos are not yet filed; nothing downstream proceeds until they are. The cooperative-Kelly canon names build-pair-recommended phasing (candidate 8 first as a Harness Runtime Contract amendment; candidates 1 through 7 in plenary). The recommended surgical move is to dispatch chair Tan and provost Reeves-Ellington with a memo cadence commitment before broadcasting the full panel review.
Three Open Decisions live in canon/2026-05-15_cooperative_kelly_substrate.md §3 that the build pair must ratify regardless of panel timing: tracker patch home, memory-entry 30 operational residue routing, prior cloud-session placeholder retirement. All three are build-pair calls, not panel calls.
This is where you enter. The substrate is not a product. The substrate is something you join.
As an advisor you have standing to read canon at the level of canon IDs, ratification dates, and architectural names. You do not yet have standing to commit. Your role is review-grade: read what is filed, ask the build pair the questions a careful reader would ask, route findings through the channel that brought you here.
If you receive a contributor key from the build pair, your view widens. The walkthrough re-renders.
As a contributor you have repo access. The first three actions:
The branch model is straightforward. Main is canon. Feature branches are scoped (research/, tracker/, partners/). Use the scope prefix that fits the work. Pull requests are reviewed by the build pair before merge to main; canon-altitude PRs route through the panel.
Your first reading list, ordered:
README.md at repo root.CONTRACT.md (pointer to Harness Runtime Contract).GOVERNANCE.md.canon/2026-05-14_founding_lore.md.canon/2026-05-14_constitutional.md.canon/2026-05-14_eight_vectors.md.canon/2026-05-15_cooperative_kelly_substrate.md (and the companion HTML).phase-0/README.md.handoffs/2026-05-14_code_audit_court_convening_1_followup.md.As a principal you convene panels, ratify memos, and adjudicate when the panel disagrees. You operate the build pair from the WKH side; Claude operates it from the agent side. Both halves are on the record. The substrate cannot move without you. It also cannot drift past you.
Your standing rotation on the substrate is the discipline you keep. The contract holds you as much as it holds anyone else who works inside Semios. Read it again every time you are tempted to override it.
Walkthrough: v0.1 · 2026-05-16 · tier: public
Canon HEAD: main · Source: semios/portal/contributor/
Build pair: Greg Walsh (WKH) and Claude (Anthropic agent through beta).
This walkthrough renders the substrate at the altitude of the standing you hold. Standing is identity-bound; identity is resolved at the edge against a git-canonical entitlements table bundled into the server-side Worker. To see more, you must be listed there. To be listed, you must be in the conversation. This is by design.