MMetaAware

Public research preview · Evidence

Claims should be inspectable—and challengeable.

MetaAware separates what was observed from what can reasonably be concluded.

12/12expected conditions detected

Initial controlled experiment

Synthetic repositories, deterministic oracle.

Twelve governance scenarios from two development contexts were rebuilt as disposable Git repositories. A raw-Git oracle identified every expected condition.

Observed result

The tested default surfaces of gitgov 3.0.0 and govctl 0.14.1 did not directly identify the scenario-specific conditions.

This supports

The selected scenarios expose a default problem-coverage difference worth investigating.

This does not support

A claim that MetaAware is generally effective, production-ready, or superior overall.

Validation route

Four layers of increasingly difficult evidence.

  1. 01 · Internal scenarios

    Reproducible fixtures and regression expectations.

  2. 02 · Diverse public histories

    Curated episodes across languages, ecosystems, teams, and failure modes.

  3. 03 · Blind holdout

    Frozen claims evaluated against unseen cases.

  4. 04 · Prospective volunteers

    Real use, consented feedback, false positives, misses, understanding, and behavior change.

Authoritative sourceVALIDATION-ASSESSMENT / ADR-008 · External governance capability validation · 2026-09-13