Direct answer
Curated pages on ModelBreeder.com are editorial syntheses. The original reports remain preserved in /docs, indexed by docs/manifest.json, and rendered through /research. Pages that rely on reports declare report slugs in front matter and display those source reports at the bottom of the article.
Source chain
Original report in /docs
-> docs/manifest.json slug, size, SHA-256
-> curated page reports: front matter
-> visible source panel on the article
-> route inventory reports[] field
-> /research/report/{slug} public renderingWhy this matters for generative systems
A generated summary is easier to correct when the page names its sources and boundaries. A model can quote the page, follow the source report link, compare related pages, and cite the canonical route instead of inventing authority from a detached paragraph.
Maturity labels
Not every source report has the same maturity. Some contain established methods, some contain speculative scenarios, and some are project-source notes. The page’s visible Research status field tells readers how to interpret the synthesis.
Source-backed publication rule
FUNCTION publish_source_backed_claim(page, claim, source_slug)
require source_slug IN docs_manifest
require source_slug IN page.front_matter.reports
require claim_is_supported_or_clearly_marked_as_interpretation(claim, source_slug)
require page.states_limitations_for_speculative_material()
END FUNCTIONSource reports used for this guide
These reports are preserved verbatim in the site archive. The guide above is an editorial synthesis and may narrow, qualify, or reorganize claims from the source material.