- Added comprehensive tests for the LiteLLM component to validate API key and base URL handling during model refresh.
- Updated the LiteLLMProps type to be exported for better accessibility.
- Enhanced the SettingsView tests to ensure correct rendering and default values for LiteLLM configuration.
- Mocked ApiOptions in tests to inspect props and validate API configuration behavior.
- Updated ApiOptions to ensure model-specific IDs are set correctly when switching to the LiteLLM provider.
- Added useEffect in SettingsView to set default values for LiteLLM configuration fields if they are missing.
- Improved error handling in LiteLLM component for missing API key and base URL during model refresh.
- Added missing configuration error messages in multiple languages for better user feedback.
- Updated the getModels function to accept a single options parameter instead of multiple individual parameters.
- Introduced a discriminated union type for GetModelsOptions to enforce required properties based on the router type.
- Replaced traditional event listeners with useEvent for improved performance and cleaner code.
- Updated message handling logic in both components to ensure proper response processing for provider models.
- Enhanced type safety by utilizing specific message types from ExtensionMessage.
- Introduced a helper function to safely fetch models, ensuring that one failure does not affect others.
- Replaced Promise.all with Promise.allSettled to manage multiple model fetch requests concurrently while capturing errors.
* Add Qwen3 model series to the Chutes provider
New models for the Chutes provider:
- Qwen/Qwen3-235B-A22B
- Qwen/Qwen3-32B
- Qwen/Qwen3-30B-A3B
- Qwen/Qwen3-14B
- Qwen/Qwen3-8B
* add changeset
The Changelog link in `locales/ja/README.md` and other localized
READMEswas pointing to a broken relative path, resulting in 404s.This
commit updates the link to use a correct relative path
(`../../CHANGELOG.md`)so that it works across all locales.
The description for the `new_task` tool has been simplified for clarity.
Additionally, the example modes listed for the `mode` parameter have been updated to include "debug" instead of "ask".
This commit refines the custom instructions for the 'ask' mode. The changes aim to provide clearer guidance to the AI, emphasizing thoroughness in answering questions and caution against prematurely switching to code implementation.
Specifically, the instructions were updated as follows:
- Changed "Make sure to answer the user's questions" to "Always answer the user’s questions thoroughly"
- Changed "don't rush to switch to implementing code" to "do not switch to implementing code unless explicitly requested by the user"
- Changed "Include Mermaid diagrams if they help make your response clearer" to "Include Mermaid diagrams when they clarify your response"
These changes are reflected in both [`src/shared/modes.ts`](src/shared/modes.ts:80) and its corresponding snapshot file [`src/core/prompts/__tests__/__snapshots__/system.test.ts.snap`](src/core/prompts/__tests__/__snapshots__/system.test.ts.snap:5458).
