SharePlane Next
SharePlane Prototype Lab
Open gallery
Interactive SharePlane artifact

Nine complete enterprise sites. One AI prototyping lab.

This artifact demonstrates what skilled human judgment, current-generation AI models, reusable design authority, deterministic tooling, and disciplined validation can produce without first constructing a traditional application platform. Each child is a complete self-contained website, preserved byte-for-byte inside this parent review environment.

Review the prototypes Understand the system
9complete site prototypes
3creative rounds
90customer route views
54working calculators
126lifecycle stage instances
934,474source HTML bytes
Why make it

Show the product before funding the platform

Stakeholders can inspect complete customer experiences, interactions, calculators, information architecture, agent surfaces, and responsive behavior before anyone commits to a production stack.

What it proves

AI speed becomes useful only when governed

The demonstration is not merely that models can emit HTML. The useful capability is preserving meaning, applying visual authority, testing behavior, separating human and agent surfaces, and retaining exact evidence.

What it avoids

Premature platform gravity

The prototypes do not require a framework, database, package manager, analytics layer, runtime AI service, remote assets, or enterprise integration to be useful for product and architecture decisions.

What remains human

Authority, judgment, and acceptance

Humans define the service, select the direction, approve public boundaries, and decide whether the prototype earns implementation. Automation accelerates execution; it does not inherit accountability.

Static nested-artifact architecture

How one HTML contains nine complete HTMLs.

The parent remains a normal SharePlane artifact. The child sites are exact byte streams encoded inside inert script payloads. Local JavaScript decodes only the selected candidate into an iframe. Nothing is fetched from a server.

01

Canonical artifact

artifact.json records identity, claims, provenance, relationships, runtime exception, and human gates.

02

Locked parent page

page.html contains the teaching narrative, gallery controls, inventory, and all child payloads.

03

Embedded prototypes

Nine original HTML files are Base64-encoded without rewriting their bytes.

04

Local review runtime

The selected payload is decoded into sandboxed iframe.srcdoc; compare mode loads two independently. A nonpersistent storage shim is injected only into the review rendering so local interactions work without same-origin authority. Original payloads and downloads remain exact.

05

Generated projections

SharePlane builds catalog, graph, context, agent package, receipt, sitemap, and discovery outputs.

Runtime boundary: the parent uses dependency-free progressive JavaScript because interactive inspection is the artifact’s purpose. The explanatory content and inventory remain readable without it. There is no runtime AI, network call, analytics, tracker, remote font, external script, or database.
Context as code

How the repository works.

One authoritative record

content/artifacts/<slug>/artifact.json governs structured identity, status, sources, claims, relationships, presentation, runtime, and validation.

One human source

content/artifacts/<slug>/page.html is the locked parent review experience. The embedded children retain their own byte identities.

Build-time compiler

Python tooling converts canonical records into the public site, catalog, graph, JSON-LD, contexts, agent packages, manifests, and receipts.

Validation before publication

Schemas, source hashes, runtime posture, copy identity, links, graph edges, agent boundaries, deterministic builds, and rendered UAT are checked before merge.

Static public runtime

The deployed artifact is ordinary static HTML and generated JSON. There is no service to patch, database to migrate, or application server to keep emotionally supported.

Receipts preserve evidence

Exact source, build, generated-tree, package, and review identities let later readers distinguish current authority from historical evidence.

Authority does not emerge from tool access

How the roles are divided.

Tony Malott

Owns product intent, semantic approval, public-safe boundaries, candidate selection, Creative Lock, acceptance, merge authority, and publication judgment.

ChatGPT

Co-develops thesis, architecture, information design, candidate systems, public framing, source packaging, and the bounded implementation contract.

Codex

Performs mechanical repository execution after authority is locked: exact placement, schema translation, hashing, build, validation, browser UAT, draft PR, and receipts. It does not rewrite or select.

SharePlane compiler and validators

Generate deterministic public and machine projections, enforce invariants, reject drift, and provide reproducible evidence. Passing validation is not human approval.

Why this is easier to manage

Manage the source once. Regenerate everything else.

Exact changes are visible

Git shows what changed, who changed it, and which source identity produced the output. A complete artifact can be reviewed as code, rendered experience, and machine record.

Rollback is ordinary

The source, metadata, schemas, and receipts are versioned together. Reverting a commit restores the complete known state without reconstructing a page from UI history.

Automation is deterministic

Catalogs, graphs, contexts, packages, and receipts are generated from reviewed authority rather than maintained manually in several places.

Agent context is explicit

Agents consume stable identifiers, statuses, provenance, relationships, and action boundaries instead of scraping visual components and inferring meaning from proximity.

Creative freedom remains local

The shared platform establishes trust and navigation without forcing every artifact into one template. Each child site keeps the visual system appropriate to its thesis.

Operational surface stays small

Static files reduce runtime dependencies, security exposure, platform administration, recovery complexity, and the number of places where truth can quietly diverge.

A bounded architectural choice

Why not build this review lab in SharePoint right now?

Use SharePoint when

  • Business teams need collaborative authoring, permissions, approvals, lists, forms, and Microsoft 365 integration.
  • The content is enterprise knowledge whose canonical ownership belongs in SharePoint.
  • Standard page components meet the experience requirement.
  • Operational support and governance already exist for the site.

Use this static prototype model when

  • The goal is rapid high-fidelity exploration before selecting a production experience.
  • Arbitrary HTML, calculators, local interaction, responsive systems, or agent surfaces must be preserved exactly.
  • Byte identity, deterministic rebuilds, diff review, rollback, and machine-readable projections matter.
  • The team needs a portable artifact that can be reviewed offline without tenant configuration or platform dependencies.
This is not “SharePoint bad.” SharePoint remains a strong knowledge, permission, collaboration, and workflow plane. The argument is sequencing: prototype and prove the experience in a reversible static substrate first, then decide what should be implemented in SharePoint, a custom web experience, or a hybrid. Do not force early product discovery through the constraints of the eventual enterprise platform.
Integrity inventory

Source identities.

PrototypeRoundBytesSHA-256
Lifecycle Command Center181,014423252e62c670524f0f76cfea99bff1d4edf2e6476235fa02b85594e51b303e7
Service Atlas179,743b1e3eb351f48958d62ec49f02bf9e9fe3f31aa2011a7bd173b7450a0f5672d5c
Engineering Confidence System179,344f26d6fee4cc8f347f31bd72495d1f6fe6074c5d5a5a6143c0346b2c246fc5b0d
Operational Broadsheet282,327e7882274c4bc59060778713c75c0d1b3438c2ca940a5312ba07616d292e9f890
Engineering Blueprint283,41971369c6efcd6a3e75ff5ff56b28aa4a0cb1b3b39aff9f4192ce7de36587f9c8e
Service Field Manual283,734e68005269aeaf6de01132123699e0c429ecd811f788a15984f021dee2c414e22
Context Ledger3146,39384ea9f829668df26b02020604895bb2ffa0c4bad3b14ad995df3bd820b0f7ab7
Authority Weave3148,1782979aac500c3167e91620ce327699969e222efe019f02f5247ff61357fab55ff
Briefing Compiler3150,322e096934ad38ac25cf09529af66f82047c7f6c31326070479fc4b1c733c3e12f1

Share or reuse

Send the page or take its public context with you.

What’s included

The agent package contains the canonical public HTML, public metadata, receipt, plain-text context, package guide, and member-hash manifest.

Agent context is a generated public-safe plain-text projection. It is not canonical Markdown and does not replace the artifact or its versioned authority.