## Purpose This is a companion pointer to the primary Drupal Eval Commons umbrella filed on `drupal/ai`: see [#3586445](https://git.drupalcode.org/project/ai/-/work_items/3586445). The umbrella issue is the coordination space for shared eval data across: - `drupal/ai` - `ai_eval` - `ai_agents_test` - `ai_best_practices` This companion exists so people following `ai_best_practices` can track the cross-module work from this queue as well. ## Why this project is involved `ai_best_practices` is directly involved in at least three parts of the proposed structure: - Layer 3: the in-tree tooling layer (\[#3586440\]) is the orchestration sibling to the registry/storage question on the primary umbrella; it covers how skills, evals, and guidelines get aggregated into and placed inside a Drupal project - Layer 4: browser / community submission work in \[#3588426\] - Layer 5: the `drupal_builder` domain bundle The proposal in the primary umbrella is that Layers 3 (tooling), 4 (browser), and 5 (domain bundle) remain here, while the cross-module coordination happens on `drupal/ai`. ## Scope This issue is not meant to create a second design discussion. Substantive discussion should happen on the primary umbrella: [#3586445](https://git.drupalcode.org/project/ai/-/work_items/3586445). This companion is only for: - visibility in the `ai_best_practices` queue - cross-linking from \[#3588426\] - pointing contributors to the primary coordination thread ## Related local issues - \[#3586440\] (tooling vision; orchestration sibling to umbrella Layer 3) - \[#3588426\] (browser / community submission; umbrella Layer 4) - \[#3588906\] (evals distribution; Layer 4 infrastructure dependency)