Add --provider flag to switch between LLM providers

[rajkumar42]

Summary

Add a --provider CLI flag that lets users switch between Claude, OpenAI, and Gemini without changing code or config files.

What needs to happen

• Add --provider flag (options: claude, openai, gemini) to relevant CLI commands
• Wire the flag to core-py's provider configuration
• Default to claude if no flag is provided
• Add validation and a helpful error if an unsupported provider is passed
• Update CLI help text and README

Example usage

opensymbolicai run my_agent.py --provider openai
opensymbolicai run my_agent.py --provider gemini

Acceptance criteria

• --provider flag is accepted on agent execution commands
• Passing a valid provider routes LLM calls to that provider
• Invalid providers produce a clear error message