feat(LLM): Update configuration and remove hardcoded /v1 to support Copilot

[Kazuto]

Summary

This PR extends Weave's LLM capabilities beyond Ollama to support multiple providers including OpenAI-compatible APIs and GitHub Copilot. It refactors the configuration layer to centralize the maxDiff parameter in the LLM config and updates all documentation to reflect the new multi-provider support.

Changes

• Updated README documentation to highlight multi-provider LLM support (Ollama, OpenAI-compatible, GitHub Copilot)
• Refactored maxDiff parameter from commit, PR, and Ollama-specific configs to centralized LLM config
• Fixed config validation to skip Ollama-specific checks when a different provider is configured
• Updated commit and PR generators to use the new centralized LLM config structure
• Enhanced LLM factory pattern to support multiple provider implementations
• Added comprehensive test coverage for new multi-provider validation logic

Test Plan

• Verify config validation passes when using non-Ollama providers without Ollama settings
• Test commit message generation with both Ollama and OpenAI providers
• Test PR description generation with multiple LLM providers
• Validate that maxDiff configuration works correctly when centralized in LLM config
• Confirm existing Ollama workflows continue to work without changes
• Review README reflects accurate feature descriptions for all supported providers