feat: improve JSON parsing and error handling by introducing isLikelyTruncatedJson function to detect incomplete JSON responses. Update client logic to handle truncation retries more effectively and enhance error messaging for truncated responses. Add tests to validate new functionality and ensure robust error reporting.

feat: enhance question repair and parsing logic by introducing strict explanation checks. Update related functions to support lenient parsing options, improving flexibility in question validation. Adjust tests to verify behavior under different explanation strictness settings.

feat: refactor AI response handling by introducing parseAiJsonResponse and related utilities to improve JSON parsing and error handling. Streamline response processing in client and pipeline modules, enhancing robustness against HTML error pages and reasoning blocks. Add tests for new parsing functions to ensure reliability.

feat: revamp README.md for enhanced project clarity and user engagement. Introduce a new project description, features list, and supported exams section. Update environment variable examples in .env.example for better developer onboarding. Refactor GeneratePage and job state management to improve user experience with job dismissal functionality. Enhance AI question generation with new language rules and regeneration capabilities, ensuring compliance with exam standards.

feat: enhance testing capabilities and introduce new components for question management. Add testing section in AGENTS.md with instructions for running tests and configuring the test environment. Update package.json files to include test scripts and integrate @electric-sql/pglite for in-memory database testing. Introduce CalloutCard and update FilterBar and SoalBrowser components to support visibility filters for private and public questions. Implement new test cases for various functionalities across the application.