✨ Add Google Gemini Support!

[dittorahmat]

Hey Simo! 👋

First off, I appreciate this project and how it helps to provide security for vibe coding. I found your repo in X Build In Public .I've been having a great time vibe coding myself, especially using VS Code, Cline & Gemini

To make VibePenTester even more accessible, I wanted to contribute by adding support for Google's Gemini models. This PR integrates the Gemini API, allowing users to choose Gemini as an LLM provider alongside the existing options (OpenAI, Anthropic, Ollama).

Key changes based on the spec:

• Adds "gemini" as a provider option (via --provider gemini and the Web UI dropdown).
• Integrates the google-generativeai library into core/llm.py's LLMProvider.
• Handles necessary message and tool format conversions between the internal OpenAI-centric format and the Gemini API's structure.
• Supports Gemini's function calling mechanism, translating requests and responses.
• Sets gemini-2.0-flash as the default Gemini model, offering a powerful and often free tier option for users.
• Includes updates to configuration (config.yaml), environment variables (GOOGLE_API_KEY), Web UI (API key input), CLI arguments, and documentation (README.md).

This integration aims to broaden the tool's usability by leveraging another major LLM provider, especially with the accessible gemini-2.0-flash model.

Looking forward to your feedback!

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
vibe-pen-tester	❌ Failed (Inspect)			Apr 19, 2025 9:21am

[dittorahmat]

Hi Simo,
Most likely, the failed deployment is because of the missing GOOGLE_API_KEY environment variable in Vercel