MWLABA

Review checklist

Version 1.0. Each review records the tested build and platform, then publishes a finding, evidence, and a result for every criterion.

  1. 1. AI disclosure

    What we check
    Record where generative AI was used, which outputs ship in the game, and whether any content is generated during play.
    Pass condition
    The disclosure names the actual uses and is consistent with the submitted materials and the reviewed build.
  2. 2. Human review and revision

    What we check
    Inspect a representative example for each major AI-assisted area. Record what the team selected, changed, or tested before release.
    Pass condition
    The evidence connects the team’s decisions to what players see, hear, or use in the reviewed build.
  3. 3. Quality and integration

    What we check
    Inspect a typical gameplay scene, including its art, interface, text, and sound. Record inconsistencies and what they change for the player.
    Pass condition
    The sampled content is readable, coherent, and fit for its purpose. Issues with a material effect on understanding or play have been addressed.
  4. 4. Game design and progression

    What we check
    Start a fresh game and complete its core loop. Test one or two advertised features and show one player decision that changes the outcome or workflow.
    Pass condition
    The loop works as described, the player decision has an observable effect, and the sampled features work in the tested build.
  5. 5. Quality assurance

    What we check
    Try a typical failure case, retry it, and test saving and loading when supported. Record how to reproduce any material issue.
    Pass condition
    The core loop recovers from the sampled failure and supported save/load preserves the tested progress.
  6. 6. AI during play

    What we check
    For live generation, test ordinary use, an unexpected input, and a repeated run. Record the outcome and what happens if the feature fails.
    Pass condition
    Live generation serves its gameplay purpose and the game remains usable when output is unexpected. Mark this item Not applicable with a reason when generation is limited to development.

When releases look very similar

Strong overlap prompts a comparison of the core loop, level structure, and progression. The review records any reused elements and the differences players can observe. Release frequency helps select a comparison.

How the badge is awarded

Passed
The evidence meets the pass condition for this point.
Needs changes
A reproducible issue materially affects understanding, play, progress, or an advertised feature. The review records the steps and the impact.
Pending evidence
A test or supporting material is still needed before a decision on this point.
Not applicable
The review records why a conditional check does not apply.

Under this checklist, all applicable points must pass for approval. The review records minor issues as observations. A completed review without approval describes what would be checked on resubmission.