Refactor IntelligentHandler to assess task difficulty only once per user message instead of on every createMessage call. Add rawUserPrompt and isInitialMessage metadata to track assessment state, implement assessment caching with race condition prevention, and update UI components to exclude intelligent provider from model listings.

…cales

Introduce frontend-based keyword assessment for task difficulty in the intelligent provider. This reduces redundant AI evaluations by pre-assessing and caching difficulty levels, allowing the handler to reuse results across conversations and task resets. Updates include new interface fields, event-driven state management, and UI integration for sending pre-assessed difficulty with messages.

- Consolidate profile handlers into unified ProfileMap structure
- Move type definitions to shared types module for better reusability
- Extract classifier prompt as class constant for maintainability
- Refactor profile loading logic to use helper methods and parallel execution
- Remove frontend difficulty assessment logic to streamline chat component
- Extract validation logic into custom hook for better separation of concerns

Remove the preAssessedDifficulty property from task options, interfaces, and handlers to simplify the codebase, as it is no longer needed after recent restructuring.

…uto-provider branch)

Fixed pre-existing JSON syntax errors in 11 translation files that were preventing the build from completing. All files were missing a comma after the intelligentProvider closing brace before the deviceAuth section.

Files fixed:
- cs/kilocode.json
- de/kilocode.json
- es/kilocode.json
- fr/kilocode.json
- hi/kilocode.json
- ru/kilocode.json
- th/kilocode.json
- tr/kilocode.json
- uk/kilocode.json
- vi/kilocode.json
- zh-CN/kilocode.json

These errors existed in the auto-provider branch before the merge with main.

…rameters

Add undefined and empty object parameters to mockProvider.createTask
calls in image mentions integration tests to match updated function
signature.