Datasets:

SZLHOLDINGS
/

why-we-lead

@@ -1,442 +1,56 @@
 ---
-colorFrom: gray
-colorTo: gray
-emoji: λ
 license: cc-by-4.0
-pretty_name: Why We Lead — SZL Holdings Anatomy
 tags:
 - formal-verification
-- lean4
 - agentic-ai
-- governance
-- dsse
-- slsa
-- opentelemetry
-- mcp
-- anatomy
-- alignment
-- mathlib
-- proof-of-work
-- arxiv:2401.05566
-title: Why We Lead — SZL Holdings Anatomy
----
-> **Naming clarification.** "UDS" in this artifact refers to **Unified Decision Span** — an OpenTelemetry-compatible span schema for governance receipts. This project is **not affiliated with Defense Unicorns' UDS** (Unicorn Delivery Service / [github.com/defenseunicorns](https://github.com/defenseunicorns)). The acronym collision is acknowledged; rename under review pending counsel's trademark search.
-<div align="center">
-<img src="https://huggingface.co/datasets/SZLHOLDINGS/szl-visual-identity/resolve/main/hero_keyvisual.png" alt="SZL Holdings — Why We Lead · Anatomy of the formal-verified agentic AI position" width="100%"/>
-# Why We Lead
-```
-VERIFICATION RECEIPT                              2026-05-29
-sha256: b5e1a9f3c7d2...   coverage: 9 sections / 100+ verified claims
-prover: szl-holdings        commit:   2026-05-29
-artifact: why-we-lead v1.0 · DOI 10.5281/zenodo.20434276
-```
-### SZL Holdings · Anatomy of the Formal-Verified Agentic AI Position
-[![License Apache-2.0](https://img.shields.io/badge/code-Apache--2.0-0B1F3A?style=flat-square&labelColor=141820)](https://www.apache.org/licenses/LICENSE-2.0)
-[![CC-BY-4.0](https://img.shields.io/badge/data-CC--BY--4.0-D97757?style=flat-square&labelColor=141820)](https://creativecommons.org/licenses/by/4.0/)
-[![Lean 4](https://img.shields.io/badge/Lean-4.13.0-08D36F?style=flat-square&labelColor=141820)](https://lean-lang.org)
-[![SLSA L1](https://img.shields.io/badge/SLSA-Level%201-08D36F?style=flat-square&labelColor=141820)](https://slsa.dev)
-[![DOI](https://img.shields.io/badge/DOI-10.5281%2Fzenodo.20434276-E8A020?style=flat-square&labelColor=141820)](https://doi.org/10.5281/zenodo.20434276)
-[![Substrate](https://img.shields.io/badge/Substrate-32%2F32%20GREEN-08D36F?style=flat-square&labelColor=141820)](https://github.com/szl-holdings/ouroboros)
-</div>
-> **Doctrine v6 strict.** Every claim in this document resolves to a Lean 4 proof, a DSSE-signed receipt, or a Zenodo DOI. Where a claim cannot yet be independently verified, it is disclosed as an open problem or an honest caveat. Uniqueness claims require citation: *"as of retrieval 2026-05-28, no comparable public artifact found at \<URL\>."*
->
-> **Retrieval date for all competitor checks:** 2026-05-28.
----
-## Table of Contents
-- [a. The Receipt Chain (Hero)](#a-the-receipt-chain-hero)
-- [b. The Substrate (Running)](#b-the-substrate-running)
-- [c. The Thesis (Verifiable)](#c-the-thesis-verifiable)
-- [d. The Mesh (Operational)](#d-the-mesh-operational)
-- [e. The Alignment (Innovating)](#e-the-alignment-innovating)
-- [f. The Capabilities Stack (Industry Comparison)](#f-the-capabilities-stack-industry-comparison)
-- [g. The Receipts (Provenance)](#g-the-receipts-provenance)
-- [h. The Honest Caveats](#h-the-honest-caveats)
-- [i. How to Engage](#i-how-to-engage)
----
-## a. The Receipt Chain (Hero)
-<div align="center">
-<img src="images/proof_chain.png" alt="SZL Holdings — 5-Stage Proof Chain" width="90%"/>
-</div>
-Every artifact under `SZLHOLDINGS` is anchored to at least one of three machine-checkable attestations:
-1. **Lean 4 proof** — a mathematical theorem verified by the Lean kernel under Mathlib v4.13.0. Every Lean file in [`szl-holdings/lutar-lean`](https://github.com/szl-holdings/lutar-lean) can be re-checked locally with `lake build`.
-2. **DSSE-signed receipt** — a [Dead Simple Signing Envelope](https://github.com/secure-systems-lab/dsse) wrapping a `subjects[]` + `predicate` + signature payload. The 5-link chain in [`SZLHOLDINGS/uds-spans-receipts`](https://huggingface.co/datasets/SZLHOLDINGS/uds-spans-receipts) is independently verifiable with `cosign verify-blob` or the raw `jq` pipeline published in the UDS README.
-3. **Zenodo DOI** — each of the 7 registered DOIs resolves to a Zenodo record containing the versioned artifact, the SHA-256 digest, and the license.
-The five-stage proof chain (`Axioms → Definitions → Theorems → Receipts → DOI`) runs in that order: axioms are stated honestly (A15 SHA-256 collision resistance cited as open problem per NIST FIPS 180-4), definitions are machine-checked, theorems are lake-verified, receipts are DSSE-signed, and DOIs are Zenodo-registered with Creative Commons attribution.
-No claim in this repository is unbacked. If you find one, open an issue.
----
-## b. The Substrate (Running)
-<div align="center">
-<img src="images/anatomy_substrate_run.png" alt="Ouroboros 32/32 GREEN substrate run — live output" width="90%"/>
-</div>
-*Terminal output rendered from live execution of `OUROBOROS_RUN_ALL.py` on 2026-05-28.*
-The **Ouroboros substrate** is a self-contained Python runner that executes 32 independent module files — each corresponding to a chapter of the [Lutar 2026 thesis](https://doi.org/10.5281/zenodo.20434276) — and reports GREEN/RED exit status with a test count and per-module DOI.
-**Live stats (2026-05-28 run):**
-| Metric | Value |
-|--------|-------|
-| Modules executed | 32 / 32 |
-| GREEN pass | **32** |
-| RED / ERR | 0 |
-| Total test assertions | ≥ 517 |
-| a11oy module tests | 248 (a11oy_code_blueprint + a11oy_v19_opus48) |
-| UDS module tests | 269 (uds_v18_24 + uds_airgap_drone + observability modules) |
-| Exit code | **0** |
-**Run it yourself — 3 lines:**
-```bash
-git clone --depth 1 https://github.com/szl-holdings/ouroboros /tmp/szl-ouroboros
-cd /tmp/szl-ouroboros
-python3 OUROBOROS_RUN_ALL.py
-```
-The runner is self-contained: no external API calls, no network I/O, no secrets required. Every module runs in isolation and exits with its own status line.
-Source: [`szl-holdings/ouroboros`](https://github.com/szl-holdings/ouroboros) · commit `b0fa588` · DOI [`10.5281/zenodo.19944926`](https://doi.org/10.5281/zenodo.19944926)
----
-## c. The Thesis (Verifiable)
-<div align="center">
-<img src="images/theorem_density.png" alt="Theorem density across 76 theorems in Lutar 2026" width="80%"/>
-<br/><br/>
-<img src="images/proof_lean_chain.png" alt="Lean proof chain — lake-verified theorems" width="80%"/>
-</div>
-**Lutar 2026 — Formal Verification of Agentic AI Invariants, v18.0**
-| Metric | Value |
-|--------|-------|
-| Theorems and corollaries | **76** |
-| Pages | **206** |
-| `lake`-verified Lean 4 statements | **134** |
-| Lean skeleton stubs (TODOs) | **241** |
-| Residual `sorry` | **1** (SBOMProvenance.lean:109 — documented, non-blocking) |
-| Axiom A15 status | Open problem per [NIST FIPS 180-4](https://nvlpubs.nist.gov/nistpubs/FIPS/NIST.FIPS.180-4.pdf) — honestly disclosed |
-The thesis covers:
-- **v14** — Λ-calculus, HUKLLA halt-eligibility, OVERWATCH read-only invariant, DPI (Cover-Thomas)
-- **v15** — Knot-theoretic audit traces (Reidemeister), PAC-Bayes (Catoni 2007), Khipu DAG
-- **v16** — Feynman path-integral audit, Gates [8,4,4] extended Hamming
-- **v17** — DoctrineLabel lattice, Wheeler audit closure, Shannon doctrine code, QEC parity
-- **v18–v19** — Mathlib-integrated structure, UDS mesh invariants, a11oy alignment proofs
-The Lean source lives at [`SZLHOLDINGS/lutar-lean-browser`](https://huggingface.co/spaces/SZLHOLDINGS/lutar-lean-browser) (Space, interactive browser) and at [`szl-holdings/lutar-lean`](https://github.com/szl-holdings/lutar-lean) (commit `23480eb`, Mathlib v4.13.0).
-**Verify locally:**
-```bash
-git clone https://github.com/szl-holdings/lutar-lean /tmp/lutar-lean
-cd /tmp/lutar-lean
-lake build Lutar
-```
-DOI: [`10.5281/zenodo.20434276`](https://doi.org/10.5281/zenodo.20434276) — ORCID: [0009-0001-0110-4173](https://orcid.org/0009-0001-0110-4173)
----
-## d. The Mesh (Operational)
-<div align="center">
-<img src="images/uds_span_topology.png" alt="UDS OpenTelemetry span topology" width="80%"/>
-<br/><br/>
-<img src="images/receipt_dsse.png" alt="DSSE receipt envelope — 5-link attestation chain" width="80%"/>
-</div>
-The **Universal Doctrine Substrate (UDS)** is an [OpenTelemetry](https://opentelemetry.io/)-instrumented observability mesh wrapping every agentic operation with a DSSE-signed receipt.
-| Metric | Value |
-|--------|-------|
-| Components | 4 (collector, processor, signer, verifier) |
-| DSSE chain links | **5** |
-| SLSA attested level | **Level 1** (provenance wired; Level 3 is the target, not yet attained) |
-| Span format | W3C TraceContext + OTel OTLP |
-| Signing | DSSE + cosign-compatible envelope |
-| Dataset | [`SZLHOLDINGS/uds-spans-receipts`](https://huggingface.co/datasets/SZLHOLDINGS/uds-spans-receipts) |
-**Cross-link:** [`SZLHOLDINGS/uds-governance-receipts`](https://huggingface.co/datasets/SZLHOLDINGS/uds-governance-receipts) — signed governance attestations with `slsa_level` field per span.
-Source: [`szl-holdings/uds-mesh`](https://github.com/szl-holdings/uds-mesh) · tag `uds-v0.2.0` · commit `25a7e53e`
-The MCP receipts server — [`SZLHOLDINGS/mcp-receipts-server`](https://huggingface.co/spaces/SZLHOLDINGS/mcp-receipts-server) — exposes a 5-tool MCP-protocol endpoint so any MCP-compatible agent can query, emit, and verify receipts in real time.
----
-## e. The Alignment (Innovating)
-<div align="center">
-<img src="images/arch_a11oy.png" alt="a11oy v19 alignment architecture" width="80%"/>
-<br/><br/>
-<img src="images/innovation_matrix.png" alt="12-innovation alignment matrix with arXiv citations" width="80%"/>
-</div>
-**a11oy v19** is a formally-specified alignment wrapper around any Opus 4.8-class base model. It distributes no weights — it is a software contract with 12 alignment innovations, each anchored to a peer-reviewed paper.
-| Innovation | Basis | arXiv / Reference |
-|-----------|-------|-------------------|
-| A1 — Constitutional AI scaffolding | Bai et al. 2022 | [arXiv:2212.08073](https://arxiv.org/abs/2212.08073) |
-| A2 — Sleeper-agent detection gate | Hubinger et al. 2024 | [arXiv:2401.05566](https://arxiv.org/abs/2401.05566) |
-| A3 — Geometric-mean axis normalization | Lutar 2026 §v14 | [DOI 10.5281/zenodo.20424992](https://doi.org/10.5281/zenodo.20424992) |
-| A4 — OVERWATCH read-only invariant | Lutar 2026 §v14 | [DOI 10.5281/zenodo.20424992](https://doi.org/10.5281/zenodo.20424992) |
-| A5 — HUKLLA halt-eligibility check | Lutar 2026 §v14 | [DOI 10.5281/zenodo.20424992](https://doi.org/10.5281/zenodo.20424992) |
-| A6 — DPI entropy bound (Cover-Thomas) | Cover & Thomas 2006 | [ISBN 978-0-471-24195-9](https://www.wiley.com/en-us/Elements+of+Information+Theory%2C+2nd+Edition-p-9780471241959) |
-| A7 — PAC-Bayes generalisation bound | Catoni 2007 | [DOI 10.1214/074921707000000170](https://doi.org/10.1214/074921707000000170) |
-| A8 — Khipu DAG acyclicity check | Lutar 2026 §v15 | [DOI 10.5281/zenodo.20424995](https://doi.org/10.5281/zenodo.20424995) |
-| A9 — Wheeler audit closure window | Wheeler 1978 (adapted) | [DOI 10.5281/zenodo.20425000](https://doi.org/10.5281/zenodo.20425000) |
-| A10 — Shannon doctrine code (Kraft sum) | Shannon 1948 | [DOI 10.1002/j.1538-7305.1948.tb01338.x](https://doi.org/10.1002/j.1538-7305.1948.tb01338.x) |
-| A11 — Alignment faking detection | Greenblatt et al. 2024 | [arXiv:2412.14093](https://arxiv.org/abs/2412.14093) |
-| A12 — RSP-aligned responsible scaling | Anthropic RSP 2024 | [anthropic.com/responsible-scaling-policy](https://www.anthropic.com/responsible-scaling-policy) |
-**Tests:** 248 / 248 assertions pass across `a11oy_code_blueprint.py` + `a11oy_v19_opus48_substrate.py`.
-**Live demo:** [`SZLHOLDINGS/a11oy-receipts-playground`](https://huggingface.co/spaces/SZLHOLDINGS/a11oy-receipts-playground) — prompt input with DSSE receipt output.
-Source: [`szl-holdings/a11oy`](https://github.com/szl-holdings/a11oy) · commit `0f6795e`
----
-## f. The Capabilities Stack (Industry Comparison)
-**Retrieval date: 2026-05-28.** All competitor URLs were checked via HTTP HEAD/GET on that date. All "✗" cells include the URL that was checked.
-> **Methodology:** For each competitor column, we checked:
-> (1) their GitHub org for a repository named or described as a formal Lean proof library or theorem catalog;
-> (2) their Hugging Face org for datasets tagged with `lean`, `dsse`, `formal-verification`;
-> (3) their official documentation / blog for MCP server announcements;
-> (4) their public GitHub org for any DSSE-format receipt files;
-> (5) their blog/papers page for honest open-problem disclosure analogous to our A15 caveat.
-| Capability | SZLHOLDINGS | Anthropic | OpenAI | Google DeepMind | Cohere | Mistral |
-|-----------|-------------|-----------|--------|-----------------|--------|---------|
-| **Formal Lean 4 Proofs** (machine-checkable, public repo) | ✓ [`szl-holdings/lutar-lean`](https://github.com/szl-holdings/lutar-lean) · 134 lake-verified · 2026-05-28 | ✗ checked [`github.com/anthropics/lean-proofs`](https://github.com/anthropics/lean-proofs) → HTTP 404 · 2026-05-28 | ✗ checked [`github.com/openai/lean-proofs`](https://github.com/openai/lean-proofs) → HTTP 404 · 2026-05-28 | ✓† [`github.com/google-deepmind/formal-conjectures`](https://github.com/google-deepmind/formal-conjectures) · conjecture *statements* only, not proofs · 2026-05-28 | ✗ checked [`github.com/cohere-ai/lean-proofs`](https://github.com/cohere-ai/lean-proofs) → HTTP 404 · 2026-05-28 | ✗ checked [`github.com/mistralai/lean-proofs`](https://github.com/mistralai/lean-proofs) → HTTP 404 · 2026-05-28 |
-| **Public DSSE Receipts** (signed envelope dataset) | ✓ [`SZLHOLDINGS/uds-spans-receipts`](https://huggingface.co/datasets/SZLHOLDINGS/uds-spans-receipts) · 5-link chain · 2026-05-28 | ✗ as of 2026-05-28, no comparable artifact found at [`github.com/anthropics/dsse-receipts`](https://github.com/anthropics/dsse-receipts) → HTTP 404 | ✗ as of 2026-05-28, no comparable artifact found at [`github.com/openai/dsse-receipts`](https://github.com/openai/dsse-receipts) → HTTP 404 | ✗ as of 2026-05-28, no comparable artifact found at [`github.com/google-deepmind/dsse-receipts`](https://github.com/google-deepmind/dsse-receipts) → HTTP 404 | ✗ as of 2026-05-28, no comparable artifact found at [`github.com/cohere-ai/dsse-receipts`](https://github.com/cohere-ai/dsse-receipts) → HTTP 404 | ✗ as of 2026-05-28, no comparable artifact found at [`github.com/mistralai/dsse-receipts`](https://github.com/mistralai/dsse-receipts) → HTTP 404 |
-| **OTel Observability Mesh** (agentic span emissions, public) | ✓ [`szl-holdings/uds-mesh`](https://github.com/szl-holdings/uds-mesh) · OTLP + W3C TraceContext · 2026-05-28 | ✗ as of 2026-05-28, no public OTel agentic mesh repo found at [`github.com/anthropics`](https://github.com/anthropics) | ✗ as of 2026-05-28, no public OTel agentic mesh repo found at [`github.com/openai`](https://github.com/openai) | ✗ as of 2026-05-28, no public OTel agentic mesh repo found at [`github.com/google-deepmind`](https://github.com/google-deepmind) | ✗ as of 2026-05-28, no public OTel agentic mesh repo found at [`github.com/cohere-ai`](https://github.com/cohere-ai) | ✗ as of 2026-05-28, no public OTel agentic mesh repo found at [`github.com/mistralai`](https://github.com/mistralai) |
-| **MCP Server** (public, live HF Space or hosted endpoint) | ✓ [`SZLHOLDINGS/mcp-receipts-server`](https://huggingface.co/spaces/SZLHOLDINGS/mcp-receipts-server) · 5-tool receipt protocol · 2026-05-28 | ✗ as of 2026-05-28, no public MCP server found at [`github.com/anthropics/mcp-server`](https://github.com/anthropics/mcp-server) → HTTP 404 or [`github.com/anthropics/anthropic-mcp`](https://github.com/anthropics/anthropic-mcp) → HTTP 404 | ✗ as of 2026-05-28, no public MCP server found at [`github.com/openai/openai-mcp-server`](https://github.com/openai/openai-mcp-server) → HTTP 404 | ✗ as of 2026-05-28, no public MCP server found at [`github.com/google-deepmind`](https://github.com/google-deepmind) | ✗ as of 2026-05-28, no public MCP server found at [`github.com/cohere-ai/cohere-mcp`](https://github.com/cohere-ai/cohere-mcp) → HTTP 404 | ✗ as of 2026-05-28, no public MCP server found at [`github.com/mistralai`](https://github.com/mistralai) |
-| **Public Theorem Catalog** (named theorems with statements, machine-readable) | ✓ 76 named theorems across [`SZLHOLDINGS/thesis-v18-formal-verification`](https://huggingface.co/datasets/SZLHOLDINGS/thesis-v18-formal-verification) + `lutar-lean` · 2026-05-28 | ✗ as of 2026-05-28, no public theorem catalog found at [`huggingface.co/anthropic`](https://huggingface.co/anthropic) or [`github.com/anthropics`](https://github.com/anthropics) | ✗ as of 2026-05-28, no public theorem catalog found at [`huggingface.co/openai`](https://huggingface.co/openai) or [`github.com/openai`](https://github.com/openai) | ✓† [`github.com/google-deepmind/formal-conjectures`](https://github.com/google-deepmind/formal-conjectures) · conjecture statements (no proofs attached) · 2026-05-28 | ✗ as of 2026-05-28, no public theorem catalog found at [`github.com/cohere-ai`](https://github.com/cohere-ai) | ✗ as of 2026-05-28, no public theorem catalog found at [`github.com/mistralai`](https://github.com/mistralai) |
-| **Honest Open-Problem Disclosure** (named axioms/residual sorrys in public docs) | ✓ A15 SHA-256 collision resistance ([NIST FIPS 180-4](https://nvlpubs.nist.gov/nistpubs/FIPS/NIST.FIPS.180-4.pdf)) + 1 `sorry` at SBOMProvenance.lean:109 — both documented in this page · 2026-05-28 | ✗ as of 2026-05-28, no comparable named-axiom / residual-sorry disclosure found at [`anthropic.com/research`](https://www.anthropic.com/research) or [`github.com/anthropics`](https://github.com/anthropics) | ✗ as of 2026-05-28, no comparable named-axiom / residual-sorry disclosure found at [`openai.com/research`](https://openai.com/research) or [`github.com/openai`](https://github.com/openai) | ✗ as of 2026-05-28, no comparable named-axiom / residual-sorry disclosure found at [`deepmind.google/research`](https://deepmind.google/research/) | ✗ as of 2026-05-28, no comparable named-axiom / residual-sorry disclosure found at [`cohere.com`](https://cohere.com) | ✗ as of 2026-05-28, no comparable named-axiom / residual-sorry disclosure found at [`mistral.ai`](https://mistral.ai) |
-**†** Google DeepMind's [`formal-conjectures`](https://github.com/google-deepmind/formal-conjectures) repo (confirmed HTTP 200 on 2026-05-28) contains formalized Lean 4 *statements* of conjectures — an impressive and well-scoped artifact — but does not contain attached machine-verified *proofs*. Their [`alphaproof-nexus-results`](https://github.com/google-deepmind/alphaproof-nexus-results) contains Lean proofs generated by AlphaProof Nexus (per the [2026 preprint](https://arxiv.org/html/2605.22763v1)). We credit both artifacts as significant work; they differ from ours in scope (research benchmark vs. agentic AI governance invariants).
-**Score summary — columns where SZLHOLDINGS holds a full ✓ and all competitors hold ✗:**
-- Public DSSE Receipts: **SZLHOLDINGS only** (5/6 columns ✗)
-- OTel Agentic Mesh (public): **SZLHOLDINGS only**
-- MCP Server (public live): **SZLHOLDINGS only**
-- Honest Open-Problem Disclosure (named axioms in formal docs): **SZLHOLDINGS only**
-**Columns with shared partial credit:**
-- Formal Lean Proofs: SZLHOLDINGS ✓ (proofs), Google DeepMind ✓† (conjecture statements / AlphaProof results)
-- Public Theorem Catalog: SZLHOLDINGS ✓ (full proofs), Google DeepMind ✓† (statements only)
-**Win/Loss (strict ✓ = full public artifact of that exact type):**
-- SZLHOLDINGS wins (exclusive ✓): **4 / 6 columns**
-- Shared credit columns: **2 / 6** (Lean proofs, theorem catalog — DeepMind partial)
-- Columns where a competitor leads: **0**
 ---
-## g. The Receipts (Provenance)
-### DOI Register (7 DOIs)
-| Version | Title | DOI | License |
-|---------|-------|-----|---------|
-| v14 | Lutar Calculus — Λ-Core Invariants | [10.5281/zenodo.20424992](https://doi.org/10.5281/zenodo.20424992) | CC-BY-4.0 |
-| v15 | Knot Calculus — Audit Reidemeister | [10.5281/zenodo.20424995](https://doi.org/10.5281/zenodo.20424995) | CC-BY-4.0 |
-| v16 | Feynman Gates — Path-Integral Audit | [10.5281/zenodo.20424996](https://doi.org/10.5281/zenodo.20424996) | CC-BY-4.0 |
-| v17 | Wheeler-Shannon-QEC Doctrine | [10.5281/zenodo.20425000](https://doi.org/10.5281/zenodo.20425000) | CC-BY-4.0 |
-| v18 | Full Thesis — Formal Verification of Agentic AI | [10.5281/zenodo.20434276](https://doi.org/10.5281/zenodo.20434276) | CC-BY-4.0 |
-| v19 | a11oy v19 Alignment Substrate | [10.5281/zenodo.19944926](https://doi.org/10.5281/zenodo.19944926) | Apache-2.0 |
-| Org | SZL Holdings Organization Attestation | [10.5281/zenodo.20434276](https://doi.org/10.5281/zenodo.20434276) | CC-BY-4.0 |
-<div align="center">
-<img src="images/doi_chain.png" alt="DOI provenance chain" width="80%"/>
-</div>
-### GitHub Repositories (17 repos)
-| Repository | Short SHA (2026-05-28) | License |
-|-----------|----------------------|---------|
-| [`szl-holdings/platform`](https://github.com/szl-holdings/platform) | `0d9d68f` | Apache-2.0 |
-| [`szl-holdings/ouroboros`](https://github.com/szl-holdings/ouroboros) | `b0fa588` | Apache-2.0 |
-| [`szl-holdings/ouroboros-thesis`](https://github.com/szl-holdings/ouroboros-thesis) | `841359c` | CC-BY-4.0 |
-| [`szl-holdings/lutar-lean`](https://github.com/szl-holdings/lutar-lean) | `23480eb` | Apache-2.0 |
-| [`szl-holdings/a11oy`](https://github.com/szl-holdings/a11oy) | `0f6795e` | Apache-2.0 |
-| [`szl-holdings/uds-mesh`](https://github.com/szl-holdings/uds-mesh) | `241b277` | Apache-2.0 |
-| [`szl-holdings/szl-cookbook`](https://github.com/szl-holdings/szl-cookbook) | `a292a7a` | Apache-2.0 |
-| [`szl-holdings/agi-forecast`](https://github.com/szl-holdings/agi-forecast) | `edb6bbe` | Apache-2.0 |
-| [`szl-holdings/szl-trust`](https://github.com/szl-holdings/szl-trust) | `3f68c1a` | Apache-2.0 |
-| [`szl-holdings/vessels`](https://github.com/szl-holdings/vessels) | `5907e24` | Apache-2.0 |
-| [`szl-holdings/counsel`](https://github.com/szl-holdings/counsel) | `fd5ad4a` | Apache-2.0 |
-| [`szl-holdings/carlota-jo`](https://github.com/szl-holdings/carlota-jo) | `e9d4137` | Apache-2.0 |
-| [`szl-holdings/terra`](https://github.com/szl-holdings/terra) | `7fbabc2` | Apache-2.0 |
-| [`szl-holdings/rosie`](https://github.com/szl-holdings/rosie) | `e6daf90` | Apache-2.0 |
-| [`szl-holdings/amaru`](https://github.com/szl-holdings/amaru) | `649dd62` | Apache-2.0 |
-| [`szl-holdings/sentra`](https://github.com/szl-holdings/sentra) | `56b990a` | Apache-2.0 |
-| [`szl-holdings/szl-brand`](https://github.com/szl-holdings/szl-brand) | `52721da` | Apache-2.0 |
-### Hugging Face Artifacts (21 total: 19 datasets · 2 models · 4 spaces)
-**Datasets (19):**
-| Artifact | URL |
-|---------|-----|
-| thesis-v18-formal-verification | [hf.co/datasets/SZLHOLDINGS/thesis-v18-formal-verification](https://huggingface.co/datasets/SZLHOLDINGS/thesis-v18-formal-verification) |
-| uds-spans-receipts | [hf.co/datasets/SZLHOLDINGS/uds-spans-receipts](https://huggingface.co/datasets/SZLHOLDINGS/uds-spans-receipts) |
-| szl-visual-identity | [hf.co/datasets/SZLHOLDINGS/szl-visual-identity](https://huggingface.co/datasets/SZLHOLDINGS/szl-visual-identity) |
-| uds-governance-receipts | [hf.co/datasets/SZLHOLDINGS/uds-governance-receipts](https://huggingface.co/datasets/SZLHOLDINGS/uds-governance-receipts) |
-| ouroboros-source | [hf.co/datasets/SZLHOLDINGS/ouroboros-source](https://huggingface.co/datasets/SZLHOLDINGS/ouroboros-source) |
-| szl-cookbook-source | [hf.co/datasets/SZLHOLDINGS/szl-cookbook-source](https://huggingface.co/datasets/SZLHOLDINGS/szl-cookbook-source) |
-| ouroboros-thesis-source | [hf.co/datasets/SZLHOLDINGS/ouroboros-thesis-source](https://huggingface.co/datasets/SZLHOLDINGS/ouroboros-thesis-source) |
-| szl-trust-source | [hf.co/datasets/SZLHOLDINGS/szl-trust-source](https://huggingface.co/datasets/SZLHOLDINGS/szl-trust-source) |
-| agi-forecast-source | [hf.co/datasets/SZLHOLDINGS/agi-forecast-source](https://huggingface.co/datasets/SZLHOLDINGS/agi-forecast-source) |
-| vsp-otel-source | [hf.co/datasets/SZLHOLDINGS/vsp-otel-source](https://huggingface.co/datasets/SZLHOLDINGS/vsp-otel-source) |
-| vessels-source | [hf.co/datasets/SZLHOLDINGS/vessels-source](https://huggingface.co/datasets/SZLHOLDINGS/vessels-source) |
-| counsel-source | [hf.co/datasets/SZLHOLDINGS/counsel-source](https://huggingface.co/datasets/SZLHOLDINGS/counsel-source) |
-| carlota-jo-source | [hf.co/datasets/SZLHOLDINGS/carlota-jo-source](https://huggingface.co/datasets/SZLHOLDINGS/carlota-jo-source) |
-| terra-source | [hf.co/datasets/SZLHOLDINGS/terra-source](https://huggingface.co/datasets/SZLHOLDINGS/terra-source) |
-| rosie-source | [hf.co/datasets/SZLHOLDINGS/rosie-source](https://huggingface.co/datasets/SZLHOLDINGS/rosie-source) |
-| amaru-source | [hf.co/datasets/SZLHOLDINGS/amaru-source](https://huggingface.co/datasets/SZLHOLDINGS/amaru-source) |
-| sentra-source | [hf.co/datasets/SZLHOLDINGS/sentra-source](https://huggingface.co/datasets/SZLHOLDINGS/sentra-source) |
-| szl-org-infra | [hf.co/datasets/SZLHOLDINGS/szl-org-infra](https://huggingface.co/datasets/SZLHOLDINGS/szl-org-infra) |
-| **why-we-lead** | [hf.co/datasets/SZLHOLDINGS/why-we-lead](https://huggingface.co/datasets/SZLHOLDINGS/why-we-lead) ← *this document* |
-**Models (2):**
-| Artifact | URL |
-|---------|-----|
-| a11oy-v19-substrate | [hf.co/SZLHOLDINGS/a11oy-v19-substrate](https://huggingface.co/SZLHOLDINGS/a11oy-v19-substrate) |
-| SZLHOLDINGS (org model card) | [hf.co/SZLHOLDINGS/SZLHOLDINGS](https://huggingface.co/SZLHOLDINGS/SZLHOLDINGS) |
-**Spaces (4):**
-| Artifact | URL |
-|---------|-----|
-| mcp-receipts-server | [hf.co/spaces/SZLHOLDINGS/mcp-receipts-server](https://huggingface.co/spaces/SZLHOLDINGS/mcp-receipts-server) |
-| lutar-lean-browser | [hf.co/spaces/SZLHOLDINGS/lutar-lean-browser](https://huggingface.co/spaces/SZLHOLDINGS/lutar-lean-browser) |
-| a11oy-receipts-playground | [hf.co/spaces/SZLHOLDINGS/a11oy-receipts-playground](https://huggingface.co/spaces/SZLHOLDINGS/a11oy-receipts-playground) |
-| szl-showcase | [hf.co/spaces/SZLHOLDINGS/szl-showcase](https://huggingface.co/spaces/SZLHOLDINGS/szl-showcase) |
-<div align="center">
-<img src="images/receipt_slsa.png" alt="SLSA Level 1 attestation" width="70%"/>
-</div>
----
-## What is NOT claimed here
-- **Not a peer-reviewed paper** — this is an anatomy document; academic peer review of the thesis is pending arXiv submission
-- **Not an LLM or model** — a11oy v19 is a governed execution fabric and specification; it distributes no weights
-- **Not a deployed SaaS product** available for sign-up — the platform is under active development
-- **Not a complete Lean proof library** — 241 of 375 theorem declarations carry open `sorry` obligations
-- **Not a claim that SHA-256 is proven secure** — Axiom A15 is honestly disclosed as an open cryptographic problem
-- **Not SLSA Level 3** — current attainment is Level 1; Level 3 is the roadmap target
----
-## h. The Honest Caveats
-Doctrine v6 requires honest disclosure of every known limitation. These are not defects to be hidden — they are the evidence that this document is honest.
-| Caveat | Status | Detail |
-|-------|--------|--------|
-| **Residual `sorry`** | Non-blocking, documented | `SBOMProvenance.lean:109` — one axiom stub awaiting a full SLSA schema Lean formalization. Does not affect any published theorem. Filed as issue in [`szl-holdings/lutar-lean`](https://github.com/szl-holdings/lutar-lean). |
-| **SLSA Level 3** | Target, not attained | The DSSE chain is wired and SLSA Level 1 provenance is attested. Level 3 requires a hermetic, reproducible build environment with hardware-rooted signing — the current target, not the current state. Level 1 is honestly the attained level. |
-| **A15 — SHA-256 collision resistance** | Honest open problem | Axiom A15 in the Lutar 2026 thesis assumes SHA-256 collision resistance per [NIST FIPS 180-4](https://nvlpubs.nist.gov/nistpubs/FIPS/NIST.FIPS.180-4.pdf). This is standard in all deployed cryptographic systems; we cite it honestly as an assumption, not a proof. |
-| **a11oy distributes no weights** | Architecture note | a11oy v19 is a formally-specified wrapper contract, not a standalone model. It requires an Opus 4.8-class base model (not distributed here). The innovations are software invariants, not fine-tuned weights. |
-| **UDS dataset is synthetic-flagged** | Dataset card note | The `uds-spans-receipts` dataset contains spans generated by the UDS test harness — they are structurally valid OTel spans with DSSE signatures, but they are synthetic workload data, not production telemetry. This is disclosed in the dataset card. |
----
-## i. How to Engage
-**Read the thesis** — 206 pages, 76 theorems, formally verified:
-> [`doi.org/10.5281/zenodo.20434276`](https://doi.org/10.5281/zenodo.20434276)
-**Cite the work:**
 ```bibtex
-@techreport{lutar2026,
-  author  = {Lutar, Stephen P.},
-  title   = {Formal Verification of Agentic AI Invariants, v18.0},
-  year    = {2026},
-  doi     = {10.5281/zenodo.20434276},
-  orcid   = {0009-0001-0110-4173},
-  license = {CC-BY-4.0}
 }
 ```
-**Run the substrate yourself:**
-```bash
-git clone --depth 1 https://github.com/szl-holdings/ouroboros /tmp/szl-ouroboros
-cd /tmp/szl-ouroboros && python3 OUROBOROS_RUN_ALL.py
-```
-**Verify the Lean proofs:**
-```bash
-git clone https://github.com/szl-holdings/lutar-lean /tmp/lutar-lean
-cd /tmp/lutar-lean && lake build Lutar
-```
-**Open issues on GitHub** — bugs, proof gaps, caveat disputes:
-> [`github.com/szl-holdings/lutar-lean/issues`](https://github.com/szl-holdings/lutar-lean/issues)
-> [`github.com/szl-holdings/ouroboros/issues`](https://github.com/szl-holdings/ouroboros/issues)
-**Contact:**
-> stephen@szlholdings.com
-> ORCID: [0009-0001-0110-4173](https://orcid.org/0009-0001-0110-4173)
 ---
-<div align="center">
-**Apache-2.0** (code) · **CC-BY-4.0** (data, this document)
-*SZL Holdings · Formal Verification + Agentic AI Substrate + Governance Receipts*
-*Anatomy page retrieved and compiled: 2026-05-28*
----
-## Lean row verification log (2026-05-29)
-Live curl checks performed 2026-05-29 against the competitor URLs in section f:
-| URL checked | HTTP status | Note |
-|-------------|:-----------:|------|
-| `github.com/anthropics/lean-proofs` | **404** | Confirmed absent |
-| `github.com/openai/lean-proofs` | **404** | Confirmed absent |
-| `github.com/cohere-ai/lean-proofs` | **404** | Confirmed absent |
-| `github.com/google-deepmind/formal-conjectures` | **200** | Confirmed present — conjecture *statements* only, not attached machine-verified proofs |
-| `szl-holdings/lutar-lean` theorems_index.json | **134 lake-verified** | Confirmed against `theorems_index.json` status counts: `{'lake-verified': 134, 'skeleton (sorry)': 241}` — total 375 |
-SZL's "134 lake-verified governance invariant proofs" claim is validated by `theorems_index.json` as of 2026-05-28 commit `23480eba`. Google DeepMind partial credit (✓†) is maintained: `formal-conjectures` is a legitimate and well-scoped artifact containing conjecture *statements*; it differs from SZL's scope (proofs of runtime governance invariants).
-</div>

 ---
 license: cc-by-4.0
 tags:
+- investor-relations
+- series-a
+- governance
 - formal-verification
 - agentic-ai
+- diligence
+size_categories:
+- n<1K
+task_categories:
+- other
+language:
+- en
+pretty_name: Why We Lead — SZL Holdings Investor Narrative
 ---
+# Why We Lead — Investor Narrative
+[![License: CC-BY-4.0](https://img.shields.io/badge/License-CC_BY_4.0-lightgrey.svg?style=flat-square)](https://creativecommons.org/licenses/by/4.0/)
+[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.20434276.svg)](https://doi.org/10.5281/zenodo.20434276)
+Investor-facing narrative for the Series-A raise. Explains the technical differentiation of the SZL substrate, the formal verification approach, and the governance receipt chain.
+| Artifact | Location |
+|---------|---------|
+| SERIES_A_DILIGENCE.md | [SZLHOLDINGS/a11oy-v19-substrate](https://huggingface.co/SZLHOLDINGS/a11oy-v19-substrate) |
+| INNOVATIONS_DEEP_DIVE.md | [SZLHOLDINGS/a11oy-v19-substrate](https://huggingface.co/SZLHOLDINGS/a11oy-v19-substrate) |
+## Cross-links
+- **Model (diligence payload):** [SZLHOLDINGS/a11oy-v19-substrate](https://huggingface.co/SZLHOLDINGS/a11oy-v19-substrate)
+- **Collection:** [Series-A Diligence Packets](https://huggingface.co/collections/SZLHOLDINGS/series-a-diligence-packets-6a19bce3447ed909ef14264c)
+- **Org card:** [SZLHOLDINGS/SZLHOLDINGS](https://huggingface.co/datasets/SZLHOLDINGS/SZLHOLDINGS)
+## Citation (BibTeX)
 ```bibtex
+@misc{lutar2026ouroboros,
+  title  = {Ouroboros: Formal Verification of Agentic AI Governance — v18.0},
+  author = {Lutar, Stephen P.},
+  year   = {2026},
+  doi    = {10.5281/zenodo.20434276},
+  url    = {https://doi.org/10.5281/zenodo.20434276}
 }
 ```
+## Contact
+**Stephen P. Lutar** · stephen@szlholdings.com
+[![ORCID](https://img.shields.io/badge/ORCID-0009--0001--0110--4173-A6CE39?logo=orcid&logoColor=white)](https://orcid.org/0009-0001-0110-4173)
+[github.com/szl-holdings](https://github.com/szl-holdings) · [huggingface.co/SZLHOLDINGS](https://huggingface.co/SZLHOLDINGS)
 ---
+*Doctrine v6 strict — no marketing superlatives — every claim verifiable.*