Can I vibecode Qodo?

Build a robust, personal AI code quality assistant inspired by Qodo, tailored for an individual developer's local workflow. Use exactly this stack: Python 3.12, Typer for the CLI interface, and SQLite for the local database; do not offer alternative stacks.
Your primary job is to create a repository-local CLI tool that indexes the current codebase, calls a local or user-provided AI model to propose code diffs, automatically generates and runs tests for those diffs, and records every change and test result in the local SQLite database.
Make the first run work locally with one documented command. Ensure the default mode is single-user and strictly private. Store all user data, including analysis history, test results, and file metadata, locally by default.
Put all secrets and external model credentials in .env, ship .env.example, and never commit credentials.
Implement the core loop: the user runs a command on a specific file or function, the tool reads the file context, prompts the AI for improvements or test cases, displays the suggested diff, allows the user to accept/reject/modify the suggestion, and then applies the change if accepted, immediately running the test suite to verify.
Support commands for generating unit tests, finding potential bugs in a specific file, and reviewing a local git diff before committing.
Include clear empty, loading, success, and recoverable error states in the CLI output. Add input validation, safe file handling, and graceful degradation if the AI model API is unavailable or rate-limited.
Write focused tests for the core CLI commands and database operations. Include one end-to-end happy path that analyzes a dummy python file.
Create a README with setup instructions, architecture overview, permissions needed, data location, and backup steps.
Do not add accounts, billing, telemetry, analytics, or a hosted control plane. Do not claim to reproduce proprietary models, enterprise policies, or large-context infrastructure.
Finish by running the tests and listing the exact commands used.