Analysis updated 2026-05-18
Check an AI-built app for dead code, leaked API keys, and missing tests before launch.
Add a ship-readiness score gate to a CI pipeline that fails builds below a threshold.
Find leftover files like final_v2.py and duplicated code left behind during vibe coding.
| amitsarin1/preflight | 0xhassaan/nn-from-scratch | 3ks/embedoc | |
|---|---|---|---|
| Stars | 0 | 0 | — |
| Language | Python | Python | Python |
| Last pushed | — | — | 2023-06-08 |
| Maintenance | — | — | Dormant |
| Setup difficulty | easy | moderate | hard |
| Complexity | 1/5 | 4/5 | 1/5 |
| Audience | vibe coder | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
No dependencies beyond Python, install via pip or use the double-click launcher with no terminal.
Preflight is a tool for checking whether an app you built with AI help is actually ready to ship as a real product, not just working on your own machine. It reads through your codebase the way an experienced engineer would skim it on their first day, and points out the specific things standing in the way, such as dead code, copy-pasted duplicate files, hardcoded API keys, missing licenses and tests, leftover files like final_v2.py, and functions that are too messy for anyone to review properly. It runs entirely offline using only Python's built in tools, with no other dependencies and no calls to any AI service, so your code never leaves your machine. Every problem it finds comes with a specific suggested fix rather than just a complaint, and it produces an overall ship readiness score from 0 to 100 that you can track over time or add to your continuous integration pipeline so a bad score fails a build automatically. There are three ways to use it. Non technical users can download the repository and double click a launcher file for Mac or Windows, which installs everything and asks you to drag your project folder in before writing a report next to your project. Developers comfortable with a terminal can install it with pip and run it against their project folder directly, or try a demo mode that builds a deliberately messy sample project to scan first. It can also be added as a single command in a CI pipeline that fails the build if the score drops below a chosen threshold. The checks cover things like unreachable dead modules, hardcoded secrets from providers such as OpenAI or AWS, duplicated code blocks, undeclared dependencies, version numbers that disagree across files, silent error handling, circular imports, oversized files and functions, leftover working files, and missing basics like a README or tests. The README is upfront that these are conservative static checks meant to prompt human judgment, not a full security audit, and that Python gets the deepest analysis while JavaScript and TypeScript get a narrower set of checks. The project is released under the MIT license.
An offline scanner that checks a vibe-coded project for dead code, leaked secrets, and missing basics, then gives it a ship-readiness score out of 100.
Mainly Python. The stack also includes Python.
Use freely for any purpose, including commercial use, as long as you keep the copyright notice.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly vibe coder.
This repo across BitVibe Labs
Verify against the repo before relying on details.