Gemini Agent MCP Server

# Documentation for Commit 2f5f8f6 **Commit Hash:** 2f5f8f672ccb5f5be7dc03a9eccd0180504d227f **Commit Message:** docs: add auto-generated documentation for historical commits **Generated:** Thu Oct 16 20:31:47 EDT 2025 **Repository:** geminiAgentMCP --- Of course. Here is the professional documentation generated from the provided git diff. *** ## Project Documentation: Gemini Agent MCP Server This document provides a comprehensive overview of the key architectural changes, feature enhancements, and refactoring efforts for the Gemini Agent MCP Server. ### 1. Summary The Gemini Agent MCP Server is a Python-based service that exposes a suite of developer tools via a JSON-RPC 2.0 API, enabling AI agents to perform tasks like code generation, linting, and GitHub operations. The project has undergone several significant evolutions: * **Initial Implementation:** A Flask-based server providing core developer tools. * **Architectural Migration:** A shift from the synchronous Flask framework to the high-performance, asynchronous **FastMCP** framework. * **API Enhancement:** A major refactor to use Pydantic models for strongly-typed API responses. * **Code Consolidation:** A structural simplification, merging all tool modules into a single application file. --- ### 2. Key Changes and Evolution #### a. Initial Implementation (Commit b19a169) The project was created as a Flask-based server with a set of modular tools designed for AI agent interaction. * **Core Functionality:** Provided a JSON-RPC endpoint (`/jsonrpc`) and an MCP manifest (`/mcp`). * **Initial Toolset:** Included tools for Gemini model interaction, GitHub issue/PR creation, code linting (`pylint`, `eslint`), test generation, and dependency analysis. * **Security:** Implemented a `is_safe_path` utility to prevent directory traversal attacks. #### b. Architectural Migration to FastMCP (Commit c74c5d0) The web service layer was migrated from Flask (WSGI) to the asynchronous FastMCP framework to improve performance and scalability. * **Dependencies:** `Flask` and `flask-jsonrpc` were replaced with `fastmcp` and `uvicorn`. * **Impact:** This was a **major breaking change**, moving the application to a fully asynchronous (ASGI) architecture. All Flask-specific code became obsolete. * **Usage Change:** The server is now started with an ASGI server. ```bash # Previous method of running with Flask is no longer valid. uvicorn main:app --reload ``` #### c. Enhanced Tool Definitions with Pydantic (Commit 2431280) To improve type safety and API clarity, all tool outputs were refactored to use Pydantic models instead of generic dictionaries. * **API Outputs:** Tools now return Pydantic `BaseModel` instances (e.g., `GitHubPrResponse`, `LintingReport`). * **Type Hinting:** Function parameters were enhanced with `typing.Annotated` for better self-documentation. * **Breaking Change:** The return type of every tool has changed. Client code must be updated to access data via model attributes instead of dictionary keys. * **Migration:** * **Before:** `pr_url = result['pr_url']` * **After:** `pr_url = result.pr_url` #### d. Code Consolidation and API Refinements (Commits 45536ee, daadf1b) The project structure was simplified by consolidating all tool modules from the `src/tools/` directory into a single `src/app.py` file. * **Structure:** The `src/tools/` directory was removed. `src/app.py` is now the single source for all tool logic and the application entry point. * **Usage Change:** The server startup command was simplified. ```bash python src/app.py ``` * **Breaking Changes:** 1. **`call_gemini` API:** The `context` parameter was renamed to `system_instruction`. Clients calling this tool must update the parameter name. 2. **Health Check Endpoint:** The `greeting://{name}` endpoint was removed and replaced with a `hello://{name}` health check endpoint. The response format also changed from `"Hello, {name}!"` to `"Server OK. Hello, {name}!"`. --- ### 3. Current Usage **1. Install Dependencies:** ```bash pip install -r requirements.txt ``` **2. Configure Environment:** ```bash export GEMINI_API_KEY="YOUR_API_KEY" export GITHUB_TOKEN="YOUR_GITHUB_TOKEN" ``` **3. Run the Server:** ```bash python src/app.py ``` The server will start on `http://0.0.0.0:8000` by default.