Feat: Add GitHub Tool for Repository Interaction

[oomis]

Summary
This PR introduces Langchain::Tool::Github, enabling AI Agents to interact directly with GitHub repositories.
As the usage of Langchain.rb grows for developer-focused assistants (e.g., "Triage these issues" or "Explain this code"), having a native GitHub integration is essential. This implementation bridges that gap, allowing agents to retrieve context from issues, pull requests, and raw file content.
Features
Issue Listing: Fetch recent open issues for repository analysis.
Pull Request Retrieval: Get details, state, and body of specific PRs.
File Content Reading: Retrieve raw file content (Base64 decoded) for code analysis workflows.
Implementation Details
OpenAI Function Calling: Implemented using the define_function DSL, ensuring full compatibility with OpenAI's tool schema generation.
Lazy Loading: Uses depends_on "octokit" to ensure the gem is only loaded when the tool is initialized, keeping the core library lightweight.
Consistent Return Types: Returns Langchain::ToolResponse objects, maintaining consistency with the recent refactoring of other tools (e.g., Calculator).

Usage Example

require "langchain"

# Initialize the tool
github = Langchain::Tool::Github.new(access_token: ENV["GITHUB_ACCESS_TOKEN"])

# 1. List Issues
github.execute(
  action: "list_issues", 
  owner: "patterns-ai-core", 
  repo: "langchainrb", 
  limit: 3
)

# 2. Read File Content
github.execute(
  action: "file_content", 
  owner: "patterns-ai-core", 
  repo: "langchainrb", 
  path: "Gemfile"
)