White Paper

.SYSTEMX white paper

Portable Project Coordination for AI-Assisted Work is the current 24-page illustrated technical white paper for founders, technical leads, developers, researchers, and operators. It explains record authority, coordination, evidence, adoption, and a proposed evaluation of .SYSTEMX.

Edition 2.0 · 20 October 2026 · Lucas (SatoshiUNO)

A product of Wayne Tech Lab LLC

Download the clickable PDF · Read the editable manuscript · Edition metadata and checksums · Local walkthrough report

Published in the immutable v1.11.0-alpha.5 prerelease on 20 October 2026, 13:01:32 UTC, release 418402174. Reviewed publication source: 85c9d6b78b9143b7ad928cc2b26050d1194921fa.

The paper examines 1.11.0-alpha.4, source e086ba15a1e70c6b585b8f12a703defc5c984b2c. Its publication release is 1.11.0-alpha.5. The paper edition, examined software, publication release, and 1.11.0-alpha.3 four-part 4K media edition are separate identifiers. The manual is the operational reference; the paper is a pinned explanation of its baseline. Alpha software is experimental and subject to change. Use it at your own risk.

Networks.Chat — author portfolio · WayneTechLab.com — business · Repository · Wiki manual

Watch the short introduction

Watch the 1:50 introduction on YouTube for a narrated animated overview of context, plans, canonical work, Agent 0/X/Y/Z/Alpha, ROOT, AUTOBOT, DATA, and .SYSTEMXP scopes. The video guide and 4K poster provide nine YouTube chapters and the original MP4, captions, and transcript. This overview is separate from a narrated reading of the complete 24-page paper. The streaming copy is on the Wayne Tech Lab channel.

Choose a reading route

Reader or question	Route through edition 2.0
Founder: what problem does this solve?	Abstract; sections 1–3 for purpose, record authority, and flexible plans; section 12 for adoption.
Operator: how do agents coordinate?	Sections 4–5 for Agent 0/X/Y/Z/Alpha, ten selected-reference AUTOBOTS, ROOT, and three TODO surfaces.
Developer: show an executable example	Section 6 and Appendix B for the bounded walkthrough; section 9 for evidence, concurrency, and security limits.
Multi-project or chat workflow	Sections 7–8 for .SYSTEMXP, authorized transport, versioned defaults, installation, and recovery.
Evaluate tokens, money, or time	Sections 10–11 for assumptions, complete accounting, existing verification, and a proposed comparative study.
Inspect terms and sources	Appendix A for implementation limits; Appendix B for reproducibility; linked Works Cited.

The twenty-page web application in section 6 is a planning example. The executable trace is a smaller protocol test; it does not claim that the paper or verifier built, deployed, or accepted that application.

Four figures explain authority and boundaries

Figure	What it explains	Editable source
1. Record authority and agent responsibilities	Context/plans support one canonical ledger; the host supplies tools and permissions.	SVG
2. Three TODO surfaces and reconciliation	Tasks, observed harness items, and optional bot observations meet through ROOT and authorized reconciliation.	SVG
3. Scope and lifecycle boundaries	Outer tools/defaults, independent .SYSTEMXP records, and owner payloads have different lifecycles.	SVG
4. Illustrative input accounting	Declared hypothetical assumptions separate selected context and coordination overhead; no measured savings are reported.	SVG

These original publication figures are separate from the raster four-part 4K field guide, which retains its alpha.3 media edition and export provenance. Captions distinguish implementation schematics from mathematical illustration.

Coordination grounded in the implementation

WORK/TASKS.json is the canonical task status and acceptance record. Context and the Master Plan explain why work exists. WORK/FOCUS.json selects the current objective, task IDs, and checkpoint; TODO/DONE views are generated projections. ROOT adds request provenance, optional milestone/wave/phase/ group structure, tool cards, and external task-list observations without creating another status authority. Task identities include their selected .SYSTEMX or named .SYSTEMXP scope.

Agent 0 routes work, integrates evidence, and retains acceptance with the user. X records events/time; Y carries an optional selected installed CLI and draft-PR handoff; Z applies the fixed advisory review; Alpha selects references from the ten AUTOBOT functions. A role name or file starts no worker and grants no tools, credentials, network access, or authority. Add optional roles when useful.

ROOT compares retained observations with all tasks bound to an external item, including members outside a selected output slice. It does not fetch an IDE or provider API. Agent 0 uses actual authorized tools, reads back external changes, and retains evidence. Provider completed, exit zero, or a review score never automatically accepts canonical done.

Exact operating contracts: ROOT, Agent 0, Agent X, Agent Y, Agent Z, Alpha/AUTOBOT, and evidence and acceptance.

Reproduce the local walkthrough

The edition 2.0 verifier creates a disposable managed project, runs 30 local CLI commands, prints sanitized arguments, return codes, returned IDs, and evidence digests, then removes the project. Python 3.9 and Python 3.14 runs completed its stated positive and refusal checks against the examined alpha.4 implementation.

From a reviewed checkout with Python 3.9 or newer:

python3 -I -B docs/media/verify_white_paper_v2.py

The example checks a synthetic harness item mapped to two tasks, aggregate running state with one task accepted and another active, and a stale bound member detected when only the other task is selected. External completion leaves canonical acceptance unchanged. Premature or missing-reviewer completion is refused; known-bad evidence fails while its counterpart passes; a future occurred event is refused while a planned event is allowed.

The observations are synthetic fixtures. The verifier starts no model, contacts no provider, creates no PR, and deploys no application. It supports tested local record behavior; it is not actual external readback or a productivity, security, or autonomous-throughput benchmark. Alpha.4 full-suite evidence is cited separately, without relabeling it as a new full-suite run for this publication.

What the efficiency model establishes

Reduced rediscovery and selected context are a testable proposition. The input-only example assumes repeated context, common task input, selected context, per-request record overhead, and setup. Changing its assumptions changes the arithmetic. It reports no measured .SYSTEMX token or dollar savings, runtime improvement, cache hit, or quality gain.

A real comparison counts all billed input/output classes, workers, tools, retries, maintenance, quality, and human effort. End-to-end wall time differs from summed parallel worker durations. Section 11 proposes paired tasks, predefined acceptance, versioned tools/models, complete costs, negative checks, and retained failures. See Tokens, cost, and time.

Publication format and verified artifact

The 24-page PDF contains 4,728 words, four original figures, and 15 references. Its professional MLA-style format uses a branded cover, author/page header, company/edition footer, linked contents/figure list, reading routes, and parenthetical citations. The body uses Times-family twelve-point double-spaced text, one-inch margins, and hanging-indent Works Cited, with compact tables/code and publication adaptations.

All 24 rendered pages passed publication review. Navigation inspection found 119 external HTTPS links, 104 valid internal links, and 15 outline entries. Those counts describe link objects, not a check of every remote site's availability. The PDF is selectable and clickable but is not tagged PDF/UA or a strict classroom submission. Markdown is the editable text alternative.

Exact PDF: 1,130,293 bytes, SHA-256 def4d23341e1aabc6908cb02a330537af549f992b8fbe23151928ea6f8383771. The metadata records edition identity and verification; release checksums identify the separate downloadable package/archive streams.

Research paper format explains draft builds, optional publishing dependencies, and review. Publication does not establish independent peer review, vendor endorsement, AGI, an external standard, or production certification.

Archived edition 1.2 and audio

Edition 1.2, Portable Project Memory for Agentic Work, remains preserved at v1.11.0-alpha.4: 96-page PDF, manuscript, metadata, and historical citation erratum. It examines 1.8.3-alpha.1, predates Y/Alpha/ROOT, and keeps its original PDF/Markdown bytes. The separate erratum corrects its release-report citation without rewriting the original. Its 29 chapters, six figures, and 32 references describe that older publication only.

The current source and new defaults distribute edition 2.0. Additive updates preserve adopted older local files. Earlier publications remain in immutable tagged history. Edition 1.1 audio keeps its original label and does not narrate 2.0; a new recording requires its own transcript, completeness review, identity, and release evidence.

First-time setup · Setup prompt · Version history · Readiness