A.X.E.L

A.X.E.L. (Automated X-electronic Engineering Link) is a CLI technical assistant specialized in debugging and optimizing Python code and SQL queries. It interprets natural language descriptions, classifies request intent, and responds with a structured diagnosis without inventing causes when context is missing.

The project originated as a DIO bootcamp challenge and evolved into an authorial implementation focused on prompt engineering and decoupled architecture. It uses a decoupled pipeline: intent classification, layered prompt assembly, an LLM call that is provider-agnostic with Ollama by default, and a structured response. Technologies include Python 3.11+, Ollama, requests, rich, and pytest.

Security measures include .env never versioned, .env.example documenting required variables without exposing values, the system prompt never exposed even under direct request or prompt injection, and no execution of arbitrary user-provided code. Evaluation uses a test case dataset covering correct intent classification, ambiguous requests, and prompt injection attempts.

Installation can be via pip (recommended, once published) with pip install axel-tech-assistant, or from source with Ollama as a prerequisite. Usage is python -m app.main. CLI commands include sair, /historico, and /limpar. Optional memory persistence via AXEL_MEMORY_BACKEND=sqlite writes messages to a local axel_history.db file. Deployment is not defined yet; the roadmap lists v1 as CLI, Python/SQL domain, and local Ollama, v2 as memory persistence and data analysis as a second domain, and v3 as REST API and a production provider as an alternative to Ollama.