Atlas Evidence Boundary Review

A public benchmark-style review of whether Atlas-related pages keep observed facts, derived labels, provider gaps, and recommendations separate.

Methodology

Inspect public route metadata, static fallback text, JSON-LD, robots policy, llms.txt, and artifact pages for claim-boundary language and crawlable evidence structure.

Version: 2026-06-26

Inputs

Outputs

Caveats

Results

MetricValueMethod
Route taxonomyProject, competition, dataset, video, benchmark, and tag classes are represented by crawlable routes.Canonical route contract in npm test.
Structured dataProject and dataset pages expose JSON-LD appropriate to public evidence.Schema contract and static build output.
Inference safetyPages include explicit what-to-infer and not-to-infer language where claims could be overstated.Visible route copy and static fallback content.

Related pages