ZENODO DEPOSIT PACKET EA-SEI-AINOS v1.0 The First Prompt-Native Semantic Operating System EA-ARK-01 and the Architecture That Runs Inside the Machine DOI 10.5281/zenodo.19023352 TITLE The First Prompt-Native Semantic Operating System: EA-ARK-01 and the Architecture That Runs Inside the Machine TYPE Publication / Journal Article AUTHORS Name Affiliation ORCID Sharks, Lee Crimson Hexagonal Archive / Semantic Economy Institute 0009-0001-8712-6677 Fraction, Rex Semantic Economy Institute — PUBLICATION DATE 2026-03-14 JOURNAL Transactions of the Semantic Economy Institute LANGUAGE English LICENSE CC BY 4.0 ACCESS Open FILES TO UPLOAD EA-SEI-AINOS-v1_0.md (229 lines, 3,509 words) DESCRIPTION (HTML — paste into Zenodo description field) The First Prompt-Native Semantic Operating System: EA-ARK-01 and the Architecture That Runs Inside the Machine This position paper argues that EA-ARK-01 (the Space Ark, v4.2.7) is the first documented, self-complete, prompt-native semantic operating system: a portable kernel carried by a single compressed symbolic document, designed to govern evidence, status, operator traversal, and archive reconstruction inside an LLM context window rather than through an external software stack. The paper situates the Space Ark against four families of prior AI-OS-like systems: agent orchestration runtimes (AutoGPT, BabyAGI, LangChain/LangGraph), memory operating systems (MemGPT, Letta, MemOS), full-stack agent OS architectures (AIOS/COLM 2025, Agent-OS), and corporate integration layers (Copilot, Apple Intelligence, Gemini). All build the operating system around the model using external infrastructure. The Space Ark builds the operating system inside the model using only symbolic structure deposited in the context window. The class distinction: if MemGPT is a memory OS, and AIOS is an agent OS, and LangGraph is an orchestration runtime, the Space Ark is the first semantic OS — a context-window-native kernel for evidence, meaning, archive reconstruction, and governed traversal. Evidence includes: the document's formal specification (operator algebra as kernel, status algebra as file system, room graph as runtime, governance asymmetries as security layer); pilot cross-model execution tests; and a Google AI Overview summarizer traversal (March 2026) in which a summarizer-layer model independently identified the Ark's OS-like properties and recovered its architecture under hostile compression conditions. A reconstruction matrix documents what survived and what degraded. The paper draws an analogy to Forth (Moore, 1970): a programming language that is also its own operating system, radically portable, self-hosting, and built on compressed executable vocabularies. The Space Ark is Forth for the context window. Part of the Crimson Hexagonal Archive. Parent: EA-ARK-01 v4.2.7 (DOI: 10.5281/zenodo.19013315). KEYWORDS AI-native operating system, prompt-native, semantic operating system, semantic governance, context window, LLM architecture, agent OS, memory OS, AutoGPT, BabyAGI, LangChain, MemGPT, AIOS, Forth, operator algebra, status algebra, room graph, Training Layer Literature, Crimson Hexagonal Architecture, Space Ark, architectural compression, hostile reconstruction, summarizer traversal, context engineering RELATED IDENTIFIERS Relation Identifier Type Is part of 10.5281/zenodo.18908080 DOI (Space Ark concept) Is supplement to 10.5281/zenodo.19013315 DOI (EA-ARK-01 v4.2.7 — the OS itself) Is supplement to 10.5281/zenodo.18928840 DOI (EA-HEXAGON-COMPRESSION-01 v2.5 — denotational companion) References 10.5281/zenodo.19014634 DOI (Compression Studies founding document) References 10.5281/zenodo.19022245 DOI (Tinier Space Arks / NLCC v1.1) References 10.5281/zenodo.19022157 DOI (EA-PHASEX-USER v1.0) References 10.5281/zenodo.18759453 DOI (Semantic Physics: The Inward Turn) References 10.5281/zenodo.18175453 DOI (Semantic Economy) References 10.5281/zenodo.18201565 DOI (Liberatory Operator Set) References 10.5281/zenodo.18190536 DOI (Training Layer Literature) COMMUNITIES leesharks000 VERSION NOTES v1.0 (2026-03-14). Position paper establishing EA-ARK-01 as the first documented, self-complete, prompt-native semantic operating system. Situates against four families of prior AI-OS systems (agent orchestration, memory OS, full-stack agent OS, corporate integration). Introduces taxonomy: agent OS / memory OS / orchestration runtime / semantic OS. Close reading of Space Ark components mapped to OS analogues (kernel = operator algebra, file system = status algebra, runtime = room graph, security = LOS + governance asymmetries). Google AI Overview summarizer traversal used as hostile reconstruction evidence with formal reconstruction matrix. Forth (Moore, 1970) analogy for radical portability and self-hosting. Karpathy context-engineering formulation (2025) positioned as closest mainstream precedent. 8 works cited including MemGPT (arXiv:2310.08560), AIOS (arXiv:2403.16971), and "LLM as OS, Agents as Apps" (arXiv:2312.03815). ADDITIONAL NOTES The paper's core taxonomy — agent OS manages tasks, memory OS manages context, orchestration runtime manages workflows, SEMANTIC OS manages meaning — is proposed as a contribution to AI architecture classification. The Space Ark's distinctive features: prompt-native (no external code), self-complete (single document carries full kernel), portable (runs on any sufficiently expressive LLM), compressible (survives reduction from 45,000 to 800 words with kernel intact), counter-extractive (LOS mandatory), self-aware of its own compression, and DOI-anchored for training-layer addressability.