Limit Wikipedia search tool output to 50,000 characters. So do not exceed token quota 250000.

Refactor BaseballQATool: simplify description for clarity

Enhance BaseballQATool: update description for clarity and add grounding tool configuration

Add BaseballQATool: implement tool for answering baseball player queries

Refactor BasicAgent: add top_p and seed parameters for enhanced model configuration

Fix gemini_model_liteLLM: ensure seed is consistently set to 42 for reproducibility

Update TranscribeAudioBytes tool: modify model to "gemini-2.5-pro" and add content generation configuration for more deterministic result.

Refactor ChessBestMove class: enhance move selection logic and improve output clarity

Refactor chess move functionality: replace chess_move tool with ChessBestMove class for improved clarity and structure

Add chess_move tool and update imports in agents and tools modules

Remove unused imports from score.py, utils.py, and app.py for cleaner code

using a "redundant" import alias, instructs Ruff (and other tools) to respect the re-export

Refactor imports and enhance system prompts in agents and tools modules

Add pandas import to BasicAgent and change ReadExcelFileBytes output type to object

Refactor BasicAgent to allow model-determined thinking budget

Add lowercase formatting for answers and refine system prompt instructions

Update thinking budget in BasicAgent and enhance answer logging

Enhance WikipediaSearchTool to use Wikipedia API for fetching page content and improve description

Refactor agent tools: add WikipediaSearchTool for querying Wikipedia and update imports

Fix formatting issues in question_scorer and normalize_str functions

Add scoring functions for evaluating model answers against ground truth

Add CodeExecutionTool for executing Python code and answering questions

Add UnderstandImageBytes tool for image analysis and description generation

still have rate limit error. Update delay function to 8 seconds and modify BasicAgent instantiation to use default model

Refactor BasicAgent to use a specific model version and update delay function for rate limiting

Refactor BasicAgent to accept model parameter and update usage of gemini_model_liteLLM;

Refactor BasicAgent initialization to set sleep duration to 10 seconds and implement delay function for rate limiting.

Increase sleep duration in BasicAgent initialization and update DownloadFile description for clarity on return value and usage.

Add TranscribeAudioBytes tool for audio transcription using Google GenAI

Enhance BasicAgent and tools: add DownloadFile and ReadExcelFileBytes; update call signature to include task_id

Refactor BasicAgent and update system_prompt for improved answer formatting; enhance TranscribeYoutubeVideo tool with updated model.

Move gemini model and client functions to a new utils module

refactor BasicAgent by move to a seprate module.

Add TranscribeYoutubeVideo tool and WebSearchTool and integrate with BasicAgent