Archstone Security
Archstone Security · Research
Every artifact here comes out of one question: does the system emit the evidence the assessment objective requires? The classification behind it was built against a single reference environment and is published in full, rows and reasoning included, so anyone who disagrees with a call can find it and say which one.
The analysis, as figures and interactive tools. Each is a single self-contained file with the data embedded — nothing is fetched, and nothing is tracked.
The published referents, the required state, the missing designation, and the three validation bands on one page.
How a published referent becomes a scored result, and where the chain breaks. Select any component to trace its connections.
All 320 objectives against five collection mechanisms. Add one collector class at a time and watch what the standard gives back.
The row-level classification. Every figure in every artifact above resolves to a row in here.
Objective, family, disposition, referent class, collection mechanism, rationale, and the change note for every revision a row has been through.
Collector ladder, family breakdown, mechanism coverage and referent split, as a workbook.
The emission test, what separates a full call from a partial one, the scope-predicate rule, and what the study does not establish.
What is published here, and what is not. The classification is published in full — every objective, its disposition, the reasoning, and the change note for each revision it has been through. The implementation mapping is not: which specific configuration item, registry path or benchmark rule satisfies a given objective on a given platform is engineering work, it is platform-specific and perishable, and it is retained. That line is the same one the submission argues the Department should draw. Designation is a policy act performed once and inherited by everyone; mapping is not.
On the numbers. The classification is an existence proof, not a population estimate — it shows that a substantial share of the standard is machine-answerable using only components the standard already requires, in one environment. A different architecture would move rows. Where a figure changed between dataset versions, the change note in the row-level data says what changed and why.
Citing this work. Every artifact carries its dataset version in the filename, so a citation to a specific file resolves to the content as it stood. This index is maintained and will change as material is added; cite the artifact, not the index.