artwork-autopsy published alpha the analysis tool

Chapter 6 · 6 of 6

Where it stands, and how to help

The service runs today and has taken a real work end to end. Broadening runtime coverage and the fully-structured proposal are in active development — and the single most useful thing you can contribute is an example artwork.

artwork-autopsy is real and running, and honestly mid-build — published alpha. The standalone service, the deterministic static core, the per-runtime deep-readers, and the AI analyst are working; the service runs packaged in Docker today — a container format that runs the same on any machine — and has taken a real born-digital work all the way from upload, through the preflight pause, the static core, the deep-readers, and one analyst session, to a reasoned picture of the machine, software, and network it needs.

In active development

  • Broadening runtime coverage — more deep-readers in the catalogue, and composing several at once when one work leans on more than one runtime.
  • The web / net.art case, deepened — enumerating a page’s plugin embeds and delegating to each, with a period browser, every plugin, and a serve-from-archive proxy. The highest-leverage case in the collection.
  • The structured proposal — projecting everything gathered into a fully populated, machine-readable plan (layers, suggestions, fidelity risks, a verdict). The loop runs end to end; this final projection is being finalised.

Honest scope: today the tool reasons its way to a host/guest/network picture, and that last fully-structured projection — and the multi-runtime cases — are still being built. No overclaiming.

How to help

Send an example artwork — the highest-leverage help

The tool gets sharper with every real work it sees. Bundle an artwork as a single zip with a short top-level description — title, artist, year, the runtime you believe it uses, the intended host OS / display / peripherals, and how it was meant to behave. That description becomes a trusted, human-authored prior for the analysis. We never destroy the original; please confirm you hold the rights to share. Write to hello@vmctl.org.

Add a runtime module

A runtime is expert knowledge expressed as data — a catalogue entry with byte signatures, a decompiler, an ordered environment recipe, and the dead servers to stand in for. Adding or sharpening one needs no changes to the engine; the under-the-hood chapter shows the shape.

Get involved The full contribution guide — example artworks, runtime modules, testing, partnering — lives on the vmctl family hub. The analysis tool also appears as a chapter on the research site, in the context of the workbench it feeds.