Olist Project Doctor

Olist Project Doctor is an AI-powered developer workflow designed to help developers maintain and understand existing Python projects faster and more safely. Developers inheriting an unfamiliar codebase often spend significant time understanding the project structure, finding testing gaps, checking the development environment, and deciding which maintenance tasks should be performed first. These problems can lead to manual effort, rework, and unnecessary changes to working application code. Our prototype uses an existing Olist e-commerce analytics project as the real-world sample codebase. Instead of adding another business feature to the dashboard, Olist Project Doctor focuses on improving the developer workflow around the project. The workflow starts with project ingestion and then performs a parallel audit using specialized agents for architecture and structure, testing, and environment/onboarding. Their findings are synthesized and prioritized before the developer is asked for approval. Once approved, the workflow generates remediation artifacts such as tests, configuration, documentation, and CI configuration. The changes are then validated by running the project's test suite and producing a health report. The prototype expanded the project from 3 tests to 75 tests, with all 75 tests passing after the workflow completed. It also added onboarding documentation, project configuration, a CI workflow, and the Project Doctor skill and runner. The target user is a Python developer who needs to maintain, understand, or safely modify an existing codebase without manually investigating every part of the project first. The key idea is to turn project maintenance from a fragmented manual process into a structured developer workflow: audit → understand → prioritize → approve → remediate → validate.