🚀 feat: Agent Cache Tokens & Anthropic Reasoning Support

[danny-avila]

Summary:

• Updated @librechat/agents dependency to version 2.1.7 for improved parallel tool calling with Google models and version 2.1.5 for Anthropic Reasoning support.
• Refactored agent title handling to aggregate text from multiple content blocks reliably and add maxTokens
• Revised the ContentParts component to display the correct "thinking" status when a message is actively being submitted.
• Added cache creation and cache read details into token spend calculations for Agents
• Updated reasoning handling for OpenRouter by configuring reasoning effort parameters according to new format
  • Closes [Bug]: only one of 'reasoning' or 'reasoning_effort' may be provided, when using with OpenRouter #6083
• Fixed top_k and top_p parameter handling for Claude-3.7 models to ensure consistent request options.
• Validated changes through updated unit tests and local testing on both backend services and frontend components.

Change Type:

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)

Testing:
I conducted local testing on both server and client sides. I verified that dependency upgrades, token spending calculations, and UI status displays work as expected. Updated unit tests for AnthropicClient, OpenRouter configurations, and the thinking status in ContentParts confirmed that the fixes and enhancements are effective.

Checklist:

• My code adheres to this project's style guidelines
• I have performed a self-review of my own code
• I have commented in any complex areas of my code
• My changes do not introduce new warnings
• I have written tests demonstrating that my changes are effective
• Local unit tests pass with my changes
• Any changes dependent on mine have been merged and published in downstream modules