Introduce MCP tools

[YuchengZhou821]

This pull request introduces integration with MCP servers, enabling external tool execution and orchestration within the runtime system. The changes add support for configuring MCP servers, managing their connections, injecting tool schemas into the LLM, and handling tool calls and results in the agent's workflow. The implementation is modular, with new classes for client management and orchestration, and updates to configuration and runtime initialization to support MCP functionality.

MCP Server Integration and Orchestration

• Added MCPClientManager and transport support for stdio and HTTP, including tool discovery, schema conversion, and tool call execution (src/mcp_servers/client.py).
• Introduced MCPOrchestrator to manage MCP tool execution, intercept tool calls from LLM output, execute tools concurrently, and re-invoke LLM with tool results (src/mcp_servers/orchestrator.py).
• Provided a loader function for MCP servers (load_mcp) to initialize connections based on configuration (src/mcp_servers/__init__.py).

Configuration and Runtime Changes

• Extended ModeConfig and RuntimeConfig to support MCP server configuration, including parsing and loading from config files (src/runtime/config.py, src/runtime/converter.py). [1] [2] [3] [4] [5] [6] [7]
• Updated runtime initialization and orchestrator lifecycle to create and close MCP orchestrator, ensuring proper startup and shutdown of MCP connections (src/runtime/cortex.py). [1] [2] [3] [4]

LLM and Action Handling

• Injected MCP tool schemas into LLM function schemas, and updated action conversion logic to handle MCP tool calls distinctly (src/llm/function_schemas.py).
• Modified the agent tick loop to process MCP tool actions and merge results before further action execution (src/runtime/cortex.py).

Prompt and Tool Description Improvements

• Enhanced agent prompt generation to include descriptions of available MCP tools for LLM guidance (src/fuser/__init__.py).

These changes collectively enable seamless integration of external MCP tools into the agent's workflow, allowing the LLM to call tools, receive results, and generate informed responses.

[Copilot]

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.

[codecov]

Codecov Report

❌ Patch coverage is 81.78295% with 47 lines in your changes missing coverage. Please review.
✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
src/mcp_servers/client.py	73.22%	34 Missing ⚠️
src/mcp_servers/orchestrator.py	93.00%	7 Missing ⚠️
src/fuser/__init__.py	25.00%	3 Missing ⚠️
src/runtime/cortex.py	85.71%	2 Missing ⚠️
src/llm/function_schemas.py	66.66%	1 Missing ⚠️

📢 Thoughts on this report? Let us know!

[Copilot]

Pull request overview

Copilot reviewed 8 out of 8 changed files in this pull request and generated 12 comments.

[Copilot]

Pull request overview

Copilot reviewed 15 out of 17 changed files in this pull request and generated 6 comments.

[Copilot]

Pull request overview

Copilot reviewed 15 out of 17 changed files in this pull request and generated 3 comments.