Question 1

How does discover.oops.zone score GitHub repositories?

Accepted Answer

Each repository is analyzed by Qwen3-32B, a large language model, across 5 dimensions: Usefulness (30% weight), Innovation (25%), Code Quality (20%), Documentation (15%), and Community health (10%). The model evaluates the repo's description, language, stars, forks, topics, license, README presence, and recency to produce scores from 0-100.

Question 2

How often are new repositories discovered?

Accepted Answer

The platform checks GitHub every 30 minutes for newly created repositories (within the last 1-3 days) that are gaining traction. It focuses on fresh projects in AI coding, LLM tools, agents, security, and general development — not established projects like React or VSCode.

Question 3

What categories of repositories are tracked?

Accepted Answer

Repositories are automatically categorized by AI into: AI Coding (Claude Code clones, coding assistants), AI Agents, LLM Tools, Security, Web, DevTools, Data, Infrastructure, CLI, Library, Mobile, Database, and Other.

Discover Repos