The fabricFor adoptersCorpus healthAssurance Advisor
Corpus health

Where the Apeiris corpus is deep, and where we are deepening it.

This page is Apeiris’s own roadmap for the corpus. It shows where we add sources and mappings next — where our sourcing is already deep, first-class, and direct, and which controls we will source more deeply next. Every focus area here is an Apeiris sourcing priority, not an adopter task.

Corpus health · Apeiris roadmapWhere the Apeiris corpus already has deep, first-class sourcing — and which controls Apeiris will source more deeply next. Research priorities are our roadmap, not adopter tasks.

Depth guide & research priorities

Two founder questions, answered from the live corpus: where should we look for depth, and where should Apeiris deepen its own sourcing next. Apeiris ranks the 12 domains by how much of each is grounded by a direct, first-class source — then, for any domain, names the sources most worth reading for depth and the controls that are the clearest focus areas for deeper sourcing. These research priorities are Apeiris’s own roadmap — where we add sources next — not adopter tasks.

Direct coverage = the share of a domain's controls with at least one direct, first-class source. Pick a domain to see where its depth lives today and where new sourcing would add the most.

Where Apeiris deepens next · shallowest first

Loading depth guide…

Best places to look for depth

The source families that most strongly ground this domain — each with the control, requirement, and why it fits.

Where Apeiris will deepen sourcing next

Controls whose strongest source today is only partial or adjacent — the constructive focus areas on Apeiris’s sourcing roadmap.

Method: direct coverage counts controls with at least one direct, first-class source; best-for-depth ranks sources by how many of a domain's controls they satisfy at direct fit, with control-and-requirement evidence; research priorities are controls whose strongest source is still only partial or adjacent, shown with the closest current source. Apeiris defines these mappings; the platform layer evaluates action context against them at runtime.

Corpus health · Apeiris roadmapWhether Apeiris’s depth tracks capability risk. The ‘soft spots’ are high-risk controls Apeiris has not yet sourced to direct fit — items on our roadmap.

Risk-weighted coverage

Depth should track risk: the controls that gate the highest-capability actions deserve the deepest sourcing. This view groups every control by its capability tier — frontier, elevated, and down — and shows how deeply each tier is grounded, so you can read at a glance whether corroboration follows capability. The highest-risk controls with the lightest direct backing are named first, as Apeiris’s constructive priority to deepen — items on our sourcing roadmap.

Capability tier = the highest-capability action a control gates. Depth = assessed source fit (direct 1.0 → adjacent 0.3). % with direct = share of the tier's controls with at least one direct, first-class source.

Depth by capability tier · highest-risk first

Capability tierControlsAvg depth% with direct source

Frontier and elevated tiers gate the highest-capability actions. A full depth bar means nearly every control in the tier has a direct, first-class source; a shorter bar is the clearest invitation to deepen.

Highest-risk controls to deepen first

High-capability controls whose strongest source today is partial or adjacent — the constructive priority on Apeiris’s sourcing roadmap, not shortfalls.

Method: each control is placed in its capability tier and scored by best assessed source fit; a tier's depth is the mean, and % with direct counts controls with at least one direct source. Apeiris defines these mappings; the platform layer evaluates action context against them at runtime.

Corpus health · Opportunity mapReading the whole corpus as a map of where AI-assurance control strength is thin. Every figure is computed live from the data — these are opportunities to strengthen, stated plainly, not shortfalls hidden.

Where control strength is thin

Looked at across all twelve domains, the coverage manifests, and the threat and risk lenses, the data itself points to where deeper controls would matter most. These assertions update automatically as the corpus grows — each number is recomputed from opportunities.json, never hand-written.

Loading the opportunity map…

Method: cross-lens synthesis of framework partial-coverage, adversarial-technique (ATLAS) mapping density, thin threat categories, high-capability source gaps, and citation-validation debt — all recomputed on every rehash. A reading of the public data, not a market claim.