ssbushi

[CLI] Option to batch run on evaluation samples

1

**Is your feature request related to a problem? Please describe.** Current evals implementation is serial, and takes up a long time if using a complex evaluator / large dataset **Describe...

feat(js/plugins/mcp): Refactor to a client manager structure, with dynamic tool calling.

3

- Modified the current McpClient class to a Client manager class - GenkitMcpClient is now a dedicated single client-server connection. - Tools are now fetched dynamically to support adding/removing/disabling clients...

[Dev UI] Dev UI makes assumptions on schema for editors, breaks sometimes for Go/Py

1

**Describe the bug** The schema provided by the runtime looks different from what JS provides and in some places the Dev UI implicitly assumes the format to match JS. I...

## Overview Multimodal results look really bad today since they are rendered as text. Embed the right media in the `rendered-field` compnent, by using `app-data-renderer` on field contents. ### User...

feat(compat_oai): Refactor to match proposal.

1

TODOS: - `openAI.model` support - `listModels` support Checklist (if applicable): - [x] PR title is following https://www.conventionalcommits.org/en/v1.0.0/ - [x] Tested (manually, unit tested, etc.) - [ ] Docs updated (updated...

[Dev UI] 'Re-run evaluation' button on evaluation details page

3

## Overview Provide a button to re-run evaluation when viewing eval run details. ### User goal(s) Easy way to retry evaluations without filling the run-eval form again. ## Requirements ###...

**Is your feature request related to a problem? Please describe.** Support OpenAI compatible plugins in Genkit JS

## Overview Current eval results: We can move the `Preview/JSON` toggle to the top and make it affect the full table. ### User goal(s) Better UI

This PR will defer manager creation to the listFlows/ lookupTrace tools, which means that the .genkit folder is not created unless these tools are called (which in turn need the...