quodeq — code quality compass
Quodeq is an AI agent that inspects your codebase using read-only tools, scores it against the six ISO 25010 quality dimensions, and maps issues to CWE classifications. It rewards good code as well as flags violations, then generates exact fixes you can paste into your IDE or AI assistant. You can run it offline with Ollama or connect to cloud models without sending your code offsite.
- Quodeq scores code with a Q² formula that rewards good patterns instead of just penalizing violations, mapping issues to ISO 25010 dimensions and CWE identifiers with exact fixes
- Runs fully local via Ollama or connects to cloud models (Claude, Codex, Gemini) without sending code offsite
- MIT licensed and fully open for inspection, extension, and custom quality criteria beyond the built-in ISO 25010/CWE checks
- Builds a structural model of the codebase across multiple languages/frameworks rather than acting as a simple linter
Quodeq is an AI agent that audits your entire codebase against ISO 25010’s six structural quality dimensions—functional suitability, performance efficiency, compatibility, usability, reliability, security, maintainability, and portability—all mapped to Common Weakness Enumeration (CWE) classes. You point it at any project and it uses read-only tools to follow imports, open files, and understand context across multiple languages and frameworks. It isn’t a simple linter; it builds a model of your code’s structure before scoring.
Its scoring uses a Q² formula that factors in both violations and compliances. Instead of only penalizing bad code, it rewards good patterns you’ve implemented. After analysis, Quodeq generates a report showing every issue alongside its ISO dimension, CWE identifier, exact code snippet, and a step-by-step fix plan. You can copy those fix plans directly into your IDE or feed them to another AI assistant for one-click remediation paths.
Quodeq runs entirely on your machine if you use Ollama, or you can connect it to cloud models like Claude, Codex, or Gemini. The entire code is MIT licensed—nothing’s hidden. You can read through its logic, extend or tweak the quality dimensions, and contribute improvements back to the project. No closed waters here.
Out of the box, you get ISO 25010 and CWE checks, but you can also define your own criteria—whether for individual projects, team standards, or company-wide policies. It integrates with popular languages and frameworks, giving you a flexible, standards-based way to enforce reliability, security, and maintainability at scale.