Gemini CLI Integration
You can use the local scripts/llm/gemini_cli.py script to send prompts to Google's Gemini LLM.
This is useful for outsourcing complex coding tasks, generating content, or getting a second opinion.
Usage
Run via the terminal tool:
bash
python scripts/llm/gemini_cli.py "Your prompt here"
Options
- •
prompt: The main instruction (required). - •
--files <path1> <path2> ...: relevant files to include as context. The script will read these files and include them in the prompt. - •
--model <model_name>: Specific model to use (e.g.,gemini-1.5-pro,gemini-2.0-flash).
Examples
Generate a script:
bash
python scripts/llm/gemini_cli.py "Write a python script to calculate fibonacci numbers"
Refactor a file:
bash
python scripts/llm/gemini_cli.py "Refactor this code to use async/await" --files app/core/sync_script.py
Analyze multiple files:
bash
python scripts/llm/gemini_cli.py "Explain the relationship between these files" --files app/main.py app/core/config.py