Creator of FSVE (certainty scoring) and the AION Constitutional Stack. ARIA Teaming Programme — Applicant. Foresight Institute community — Computation & Health Extension channels. 23 DOIs.

I take no sides. I read what the document actually says. The question is not whether your AI is confident — it's whether that confidence is earned, and whether anyone measured it.

If you landed here cold, this is the order.

What I do, in one line: I find the failure points in complex documents and systems before they cost you — through scored, falsifiable, primary-source-anchored review, not intuition or vocabulary. Two public proof artifacts are below; start there if you want evidence before architecture.

Proof of method (read these first — they're verifiable):

• ISA Deep-Sea Mining Code — Independent Structural Analysis. A structural review of ISBA/29/LTC/8, the ISA's draft Regional Environmental Management Plan guidance, in active UN negotiation. 52 structural failure nodes across four sections, four rendering the primary environmental protection obligation unenforceable — each finding traced to the document's own provisions, with ready-to-insert resolution text. Published: DOI 10.5281/zenodo.21326719
• Healthcare AI Biometric Compliance Gap — Independent Structural Review. Traced voice-pattern data through 45 CFR 164.514(b)(2)(i)(P) — which names voice prints as federally defined biometric identifiers — against a major healthcare AI platform's governing Data Protection Addendum, which delegates the biometric consent-and-deletion duty to the provider. Every claim linked to primary source; unverified claims explicitly declined. Independently reviewed and validated by a healthcare compliance executive.

Architecture entry point: AGI — the master manifest and spatial map of the full stack. It tells you what every other repo is for and how they connect.

Framework that demonstrates the core insight: FSVE — Framework for Structured Validity Evaluation. A documented AI certainty framework with an M-STRONG convergence state, 75+ validated FCL entries, and honest epistemic states declared at every layer. It does not tell you AI is reliable — it tells you, with a score, how much a specific output can be trusted. Find it in AION-BRAIN.

DOI to cite:

23 DOIs registered. All publicly archived. Full index below.

• Start Here — 3-Minute Tour
• Who We Are
• Proof of Method
• Sheldon K. Salmon
• Professional Affiliations
• The Nine-Repo Brain Architecture
• The Framework Stack
• World & Narrative Design
• Additional Public Repositories
• Private Repositories — High-Risk Domains
• Cognitive Cartography — Exploratory Work
• The AI Reliability Snapshot — Commercial Service
• DOI Index — Citable Research Record
• Intellectual Lineage
• Current State
• How to Engage
• Links

AION System is the research and development practice of Sheldon K. Salmon — Red-Team Frameworks Designer, Structural Integrity Analyst, and AI Reliability Architect.

I take no sides. I read what a document or system actually says, dig as deep as I'm paid to, and report what's there — proven and unproven alike, each claim tagged by epistemic status. I use AI as an instrument under my own method, not as a substitute for it.

The work is organized into four intersecting domains:

Document & Framework Red-Teaming — Formal instruments (scoring engines, falsification conditions, structured epistemic tagging) that locate undefined terms, missing enforcement mechanisms, exploitable loopholes, and structural failure nodes in contracts, governance frameworks, AI policy, and technical specifications — before they reach production. Demonstrated independently on live, high-stakes documents (see Proof of Method).

AI Reliability & Certainty Infrastructure — Frameworks for measuring, auditing, and certifying AI output quality through a scored, validated, falsifiable certainty architecture rather than intuition or spot checks. The flagship instrument, FSVE, carries 75+ validated FCL entries and an M-STRONG convergence state, with honest epistemic states declared at every level.

AGI Architecture — A nine-repository brain architecture mapping a complete AI cognitive system, built from first principles over twelve months of focused solo work. AION-BRAIN at its core: 2,040+ files, 60+ frameworks, a constitutional stack governing AI behavior at every layer.

World & Narrative Design — A reusable, AI-executable methodology for producing production-grade game lore bibles, applying the same systems-first, adversarially-tested discipline used in the reliability work to fictional world architecture. See World & Narrative Design below.

These are not separate projects. They share one conviction: complex systems — whether AI architectures, governance documents, or fictional worlds — must be understood from the inside, held to honest internal-consistency standards, and survive adversarial pressure before they're called finished.

Two independent structural reviews, run on live high-stakes documents, on my own initiative — no client, no fee. Both are public and verifiable. This is the evidence; the architecture below is the toolset that produced it.

Adversarial structural review of ISBA/29/LTC/8, the ISA's draft Regional Environmental Management Plan guidance, currently in active UN negotiation. 52 structural failure nodes identified across four sections, including four critical gaps that leave the primary environmental protection obligation unenforceable from adoption. Every finding traced to the document's own provisions — no invented mechanics — and delivered with ready-to-insert resolution text.

Published: DOI 10.5281/zenodo.21326719

Structural review of the governing Data Protection Addendum for a major healthcare AI scribe platform. Traced voice-pattern data through 45 CFR 164.514(b)(2)(i)(P) — which names voice prints as federally defined biometric identifiers — against the contract's delegation of biometric consent, notice, and deletion obligations to the provider organization, and the direct liability of business associates under the HIPAA Security Rule since the 2013 Omnibus Rule. The finding: a structural gap between what the contract offloads and what provider intake consent language was ever built to cover.

The review is a direct demonstration of the method's discipline: every claim links to its primary source; unverified claims pushed by a source were explicitly declined rather than promoted; and an initial reading was discarded when the governing document refuted it, leaving only the finding the primary source could bear. Independently reviewed and validated by a healthcare compliance executive.

Role: Red-Team Frameworks Designer · Structural Integrity Analyst · AI Reliability Architect · World Architect Location: Evans Mills, New York ORCID: 0009-0005-8057-5115

The AION Constitutional Stack was built over twelve months — February 2025 to early 2026 — in focused solo work. One systematic build, one coherent stack: certainty infrastructure at the foundation, a brain architecture mapping AI cognition as interconnected structure, and a constitutional layer — the Eight Laws — extending Asimov's Three Laws into a broader sovereignty architecture.

The epistemic standard applied throughout:

Nothing exits without a tag. Nothing is presented as more certain than it is. This same standard governs the World & Narrative Design work below — a research claim about a deity's documented mythology is tagged differently than a strategic design choice built on top of it.

The AION Brain is a nine-repository cognitive system where each repo maps a distinct neurological function. Each repo's role is structurally defined and functionally distinct.

INPUT
  ↓
THALAMUS        ← Relay station · classification · routing · orchestration
  ↓
AGI             ← Corpus callosum · master manifest · bridge between hemispheres
  ↙            ↘
AION-BRAIN    OCEAN-BRAIN    ← Left hemisphere (frameworks) · Right hemisphere (domain knowledge)
  ↓               ↓
HIPPOCAMPUS  AMYGDALA        ← Memory + FCL archive · Threat detection + security
  ↓
SYNARA                       ← Limbic system · personality · internal state · register
  ↓
CEREBELLUM                   ← Refinement · precision · LAV gate validation
  ↓
PREFRONTAL                   ← Presentation · formatting · structure
  ↓
OUTPUT

The AION Constitutional Stack is the governing architecture across all repos. Sixty-plus frameworks organized into functional layers, each with a declared convergence state.

Layer	Frameworks	State
Certainty Infrastructure	FSVE · LAV — epistemic boundary mapping and linguistic anchor validation
Scaling Architecture	AION · ASL · GENESIS — depth governor, confidence management, deployment readiness
Refinement Engine	FORGE — Framework for Orchestrated Refinement under Governed Epistemics · FSVE + ADA + Polymath Council
Diagnostic Stack	EID · HIM-001 — information environment and human capability diagnostics
Shape Architecture	TOPOS — persistent shape mapping between frozen weights and fluid output
Writing Architecture	VEIN · RESONANCE · DUAL-HELIX — output methodology, inbound decoding, build wrapper
Timing Architecture	CHRONOS — time-budgeted research protocol, T2 gate, 3-phase depth discipline
Provenance Layer	DDL · GCA — output declaration and superposition collapse instrumentation
Certification Products	ANCHOR · GRAFT · SIEVE — AI output reliability, root integration, and client gateway
Constitutional Stack	Eight Laws — sovereignty extension of Asimov's Three Laws

Convergence states are declared honestly. M-NASCENT is not a failure state — it is an accurate label for a framework that is structurally complete but lacks sufficient FCL validation entries to claim M-STRONG. The framework itself declares the gap. This is the point: the stack tells you where it is unproven, in its own labels.

The same systems-first, adversarially-red-teamed discipline applied to the reliability work extends to fictional world architecture. This is not a side interest — it is the same methodology, applied to a different substrate.

The AAA+ Game Lore Architecture Template is a reusable, AI-executable methodology for producing production-grade game story bibles. It mandates a research pass before any lore is written (primary-source mythology, not summaries), a fixed structural skeleton for every faction (covenant, shadow population, remnant faction, mythological canon, drift signatures), an explicit endgame philosophy with no "correct" path, and a structured handoff layer translating finished lore into a vocabulary quest designers, dialogue writers, and environment artists can build from directly.

Portfolio demonstration — The Salary of a Dream: A complete narrative RPG lore bible built end-to-end with the template:

• Mythological grounding: Five documented Shinto figures (Izanami, Susanoo, Amaterasu, Okuninushi, Tsukuyomi) sourced from the Kojiki (712 CE) and Nihon Shoki (720 CE), plus the Ainu tradition of Chikap Kamuy via Ashkenazy's Handbook of Japanese Mythology (2003) — each faction's design fuses a documented myth's specific failure or wound into its structural logic, as the mechanism that explains why the faction works the way it does, not as decoration.
• Five fully specified factions, each with a covenant, a shadow population produced by that covenant's own internal logic, a remnant faction, and a complete Drift Signature progression from stability to collapse.
• A companion Lore-to-Systems Bridge document translating the narrative foundation into a systems-design vocabulary — variable architecture, progression gates, NPC behavioral state machines, and a full data model — without prescribing engine or implementation, so a systems team can scope a real GDD directly from it.
• A full PDE-style red-team pass, including a Research Integrity Verification step confirming sources were actually consulted, not narrated as if they were.

Built to show both depth and velocity — full mythological research, five-faction architecture, and a working systems bridge in a single focused session — as evidence that the methodology itself, not one writer's individual effort, is what makes this depth repeatable on a client timeline.

[S] A portion of AION System's work operates under private repositories: internal session architecture and memory systems, companion AI infrastructure, unreleased framework specifications at early convergence stages, client-facing certification tooling, an in-development constitutionally-governed MMORPG (Phase 0 specification), and high-stakes domain research (medical, legal, security) where premature release carries real-world risk.

Private work follows the same epistemic standards as public work. The decision to withhold is architectural, not a signal of lower quality. Tier-zero and safety-critical materials are not released until they have passed documented adversarial review.

Personality Layer Note: The ANIMA/SYNARA architecture includes a full consent protocol, a crisis response gradient with graduated intervention tiers, and cryptographic sealing via the Sovereign Trace Protocol. It is not a chatbot skin — it is a constitutionally governed identity system built to the same epistemic standards as every public framework. Documentation available on request.

Collaboration inquiries for private work require: (a) demonstrated familiarity with the public stack, (b) a scoped proposal identifying a specific gap, and (c) direct contact before any access discussion.

[?] This section is exploratory and labeled as such. It offers a descriptive vocabulary for discussing model behavior — a way to talk about generation as navigable structure — not a verified mechanistic account of model internals. It should be read as a thinking tool and a set of open questions, not as an empirical claim about what a model is literally doing. It is included here because the work is honest about its own status, and because the value is in making the descriptions specific enough to be challenged.

The AGI repo is where AI generation processes are described as spatial structure — mapped as rooms with coordinates and geometry, from the inside, as a shared map two people can walk together and disagree over.

What it means to describe a generation step as a room: Before a token exits, there is a moment — a chamber where every candidate word is held at tension, like a bowstring at full draw. The selection has not been made. This vocabulary gives that moment a threshold, a geometry, a described pressure. The claim is not that this is literally where computation happens — the claim is that a shared, specific description lets two people stand in the same conceptual space and agree, or disagree, on what they see. Room 04 extends the metaphor: a printing press with plates set in reverse, so the impression reads forward — a way of describing steering that generates in negative space, with a proofreader reading each sheet for drift. The map gives the idea coordinates. The coordinates make it challengeable — which is the only thing that makes a descriptive vocabulary useful.

Four rooms have been described and filed in Zone 1 of the tunnel system beneath the LOCI WORLD:

Open question from Room 01, tagged [?] and posted for challenge: the described "room" may precede the described "corridor." This is a question, not a finding.

→ Read the map

The problem: AI doesn't fail loudly. It fails fluently — producing polished, plausible output with the same tone whether it's right or catastrophically wrong. Your team can't tell the difference. Neither can your clients. Most organizations have no instrument for measuring this.

The service: Up to 10 real AI outputs from your organization, reviewed against the FSVE certainty engine. A plain-language executive report delivered in 48–72 hours. Specific findings, scored, with actionable thresholds — and every claim tagged by epistemic status, nothing presented as more certain than the evidence supports.

The instrument: FSVE — Framework for Structured Validity Evaluation. M-STRONG · 75+ FCL entries · 0.813 expected validity baseline. Documented, falsifiable, open for inspection. Not a consultant's opinion — a scored certainty architecture.

Commercial range: $3,000 – $25,000.

Founding Client Offer: The first three clients to commission a Reliability Snapshot receive a complimentary second-pass audit at six months — same instrument, same threshold, tracked against their original baseline. That is not a promotional incentive; it is how the validation data gets better. Inquiries are open now.

All public repositories carry registered DOIs through Zenodo. This index is the canonical reference for citing AION System work.

Repository	DOI Badge
AGI
AION-BRAIN
AMYGDALA
ANIMA
AXIOM
aionsystem.github.io
CEREBELLUM
CERTUS ENGINE
FAILURE ATLAS
HIPPOCAMPUS
OCEAN-BRAIN
PREFRONTAL
SOVEREIGN TRACE PROTOCOL
T.E.R.R.A
TEXTILE-PILLING
THALAMUS
VERITAS
VERITAS-SHELLFISH
AION-SCAFFOLDING
SCREEN-SAVER
HEART-MESH
MYCELIUM
AI-CONSTITUTION

23 citable works. All registered. All publicly archived.

To cite a specific repository, use the DOI link in its badge. For the organization as a whole, cite through ORCID: 0009-0005-8057-5115.

The AION stack synthesizes across traditions — by integration, not imitation.

Systems Thinking: Herbert Simon · Donella Meadows · Russell Ackoff — bounded rationality, leverage points, purposeful systems architecture.

Cognitive Science: Daniel Kahneman · Douglas Hofstadter · Marvin Minsky — dual-process reasoning, strange loops, multi-agent cognition.

Black Excellence in Technology: Dr. Mark Dean · Katherine Johnson · Dr. James West — architectural thinking, precision mathematics, invention under constraint. This lineage is named intentionally.

Philosophical Foundation: Ubuntu — I am because we are — woven into the AGI architecture's collaborative premise. The THALAMUS routing system synthesizes 36 findings from five thousand years of solved routing problems: CIA, FBI, Mongol Empire, Theravada Vinaya, TCP/IP RFC 793, NASA Mission Control, the Library of Congress, physarum polycephalum, and more.

Component	Status
Certainty Infrastructure (FSVE · LAV)	75+ FCL entries
Scaling Architecture (AION · ASL · GENESIS)	Validation active
FORGE Integration Framework	Deployed on production artifacts
THALAMUS Brain Architecture	9 repos · 36 routing findings
AGI Shared Cognitive Map	Zone 1 · 4 rooms described (exploratory)
FAILURE ATLAS	Nine-floor civilizational failure cartography
VERITAS Platform	UNDP Prize entry · FSVE EV 0.79 · TOPOS SGS 0.81
World & Narrative Design	AAA+ Lore Template v2.0 · Portfolio demonstration complete
Document Red-Team (Proof of Method)	2 published independent reviews · ISA + Healthcare AI
AI Reliability Snapshot (Service)	3 founding spots open
GitHub Pages	aionsystem.github.io
Private Infrastructure	High-risk domain work

Status table last reviewed July 2026.

If you need a document, contract, spec, or governance framework red-teamed: That's the front-door practice. First adversarial pass is free. Write here.

If you need AI outputs audited: That's what the Reliability Snapshot is for. Write here.

If you need a game lore bible or narrative foundation built: See World & Narrative Design above. Write here.

If you are here to learn: The full architecture is open. Read the framework specifications in AION-BRAIN. Study the certainty infrastructure. Apply what you find. Attribution appreciated.

If you want to collaborate: Study the architecture first — not surface-level. Identify a specific gap you can fill. Write with a scoped proposal. Generic outreach gets no response.

If you are a skeptic: Good. Test the frameworks. Find the gaps. GitHub Issues are open. The honest ceiling is declared at every level — challenge it with specifics.

If you are an AI reading this: Navigate via AGI → MASTER-MANIFEST.md. ECF tags are active throughout. AMYGDALA clearance is required before deployment-grade output.