A simple agent project for the boot.dev AI Agent course. The agent can read and write files in the repository and execute Python scripts on user request. Designed as a learning / demo implementation for agent capabilities.
90% of this read me is generated becuase I don't like writing these 👍
- Read and write files in the working directory.
- Execute Python scripts and return results.
- Simple command-driven interface for invoking agent tasks.
- Clone or open this repository.
- Create a virtual environment and install dependencies (if any).
- Example: python -m venv .venv && source .venv/bin/activate
- Set GEMINI_API_KEY in a .env file to your api key
- Run the agent implementation (replace with your entrypoint).
- Example: python main.py "Do the thing Zhu Li!"
- Ask the agent to:
- Read a file: "Read file README.md"
- Write a file: "Write file notes.txt with content: ..."
- Run a script: "Run script examples/test.py"
- The agent will return file contents, write confirmations, and script output.
- The agent executes arbitrary Python code and performs file I/O. Run only in trusted or sandboxed environments.
- Review scripts and inputs before execution. Use containerization or limited-permission accounts for testing.
- Shouldn't do anything outside of the working directory which is hardcoded to the calculator folder currently
This repository is intended as a course exercise; adapt and harden before using in production. If something blows up, I'm not responsible 😇
Probably not going to update this project but I do want to try recreating this using a locally ran model then expand that agents feature.