Editorial method

Purpose

The source corpus combines technical surveys, conceptual frameworks, psychological analogies, governance proposals, and speculative scenarios. ModelBreeder.com preserves that breadth while preventing maturity levels from being blended into a single voice.

Editorial pipeline

1. Preserve source files verbatim. Every supplied Markdown report is copied into /docs with its original filename.
2. Compute integrity metadata. A manifest records byte size and SHA-256 digest for each source.
3. Normalize terminology. The curated guides distinguish established method names from proposed umbrella terms.
4. Separate evidence levels. Established techniques, emerging practice, conceptual synthesis, engineering blueprints, risk analyses, and speculation are labeled differently.
5. Extract implementation lessons. Repeated themes become contracts, control-plane boundaries, pseudocode, metrics, release rules, and failure modes.
6. Add safety framing. Anthropomorphic or adversarial scenarios are translated into governance, threat-model, and containment requirements rather than operational instructions for harmful behavior.
7. Cross-link sources and guides. Curated pages identify the archived reports from which their synthesis was developed.

PROCEDURE curate_report(report)
    preserved <- COPY_VERBATIM_TO_ARCHIVE(report)
    digest <- SHA256(preserved)
    claims <- EXTRACT_CLAIMS_METHODS_AND_SCENARIOS(report)

    FOR each item IN claims
        item.maturity <- CLASSIFY_MATURITY(item)
        item.scope <- DEFINE_ENGINEERING_SCOPE(item)
        item.cautions <- IDENTIFY_LIMITS_AND_RISKS(item)
        MAP_TO_CURRICULUM(item)
    END FOR

    PUBLISH_ONLY_WITH_SOURCE_LINKS_AND_STATUS_LABELS()
END PROCEDURE

What was not done

The source reports were not rewritten in place, and the archive should not be read as a set of equally verified publications. Individual citations embedded in the supplied reports may be incomplete, outdated, speculative, or based on secondary sources. Curated pages therefore emphasize durable methods and explicitly qualify uncertain claims.

Content style

Code examples are pseudocode so the architecture remains implementation-neutral. Durable records use UTC time. Examples emphasize immutable artifacts, least privilege, reproducibility, and reversible rollout. The site does not require a database or third-party service.

Updating the site

A future update should preserve the old digest manifest, add new source records rather than silently replacing them, and change the editorial version. Material changes to definitions or recommendations should be captured in the changelog.