Reference Intermediate 1 minute read Updated 2026-06-29 UTC

Schema Reference

Practical Genome, FitnessVector, and ReleasePacket schemas for model-breeding systems.

Research statusSchema reference Publication statePublished Reviewed byMichael Kappel Source reports2

Core objects

Use these objects to keep descendants inspectable, reusable, and easy to discuss.

pseudocode
STRUCT Genome
    id
    parent_ids
    base_model
    adapters
    merge_recipe
    quantization
    routing_policy
    mutation_budget
    provenance
    created_at_utc
END STRUCT

STRUCT FitnessVector
    utility
    calibration
    robustness
    latency
    memory
    energy
    privacy
    novelty
    maintainability
    human_benefit
    evidence_uri
    evaluated_at_utc
END STRUCT

STRUCT ReleasePacket
    candidate_id
    champion_id
    intended_use
    comparison_summary
    fitness_delta
    resource_delta
    lineage_uri
    evaluation_set_uri
    release_stage
    rollback_target
    reviewer_notes
    created_at_utc
END STRUCT

Why schemas help

Schemas let tools agree on what a descendant is, what evidence means, and how release decisions are recorded. They also make local-first packages easier to archive in /docs, a file-backed registry, or a project repository.

Source 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.