Hottest AI repositories

ARTEX is a Go+Next.js autonomous penetration-testing system that won Baidu's "agent+" attack-defense challenge — and it keeps a human in the loop.

iPhone Use wires Codex into a physical iPhone via WebDriverAgent, so natural-language tasks become taps, swipes, and typed text on actual hardware.

artex-ko translates the outputs of ARTEX, an LLM multi-agent system that runs penetration tests on its own, while keeping its internal reasoning prompts untouched.

A MIT-licensed pipeline that reverse-engineers any app, rebuilds it clean-room, and mines real reviews for a competitive angle.

Moli is a standalone Rust headless browser that avoids continuous rendering by computing layout and paint only when an operation explicitly demands geometry or pixels.

Whirl open-sources the full code behind whirl.chat: every top model via OpenRouter, real memory, and encrypted locked chats you can run yourself.

An agent skill that turns hard questions into readable one-page HTML while the model writes only the content, not the markup.

OpenQodex points your existing Claude Code or Codex at your unpushed change, feeds it scanner findings and every changed line, and refuses to call a review complete until it has seen all of it.

unlazy forces AI agents to prove their work through runnable acceptance gates instead of trusting their own exit reports.

Modding knowledge is scattered forum lore; this repo turns it into agent skills, a fal-powered asset pipeline, and a knowledge base that every modding agent writes back into.

Because editing a dozen config files just to try a new model is a waste of keystrokes.

Most browser agents chat at you; this one returns editable spreadsheets, slides, and sites you can actually use.

ArtCraft swaps dice-roll prompting for deliberate scene blocking, compositing, and character posing across a dozen generative models.

A Rust-based Chinese IME that treats your daily typing as a language lesson, because opening a study app is a chore.

Agent Smith routes routine Codex work to cheaper models and shows the receipts — a local router plus a native macOS usage monitor.

EasyRead translates English papers into typeset Chinese page by page — and keeps the original one click away, because a translation you can't check is a translation you can't trust.

yomiyasu is an agent skill that deodorizes LLM-written Japanese, repairing the sentence structure underneath instead of playing buzzword whack-a-mole.

OpenRig wraps AI coding agents in persistent, YAML-defined teams with roles, shared context, and tmux sessions you can actually attach to.

A hunter open-sources the LLM prompt chain behind an evidence-first bounty process — but leaves the infrastructure as an exercise for you.

Oil UI makes coding agents produce several genuinely distinct UI directions, put them side by side, and refine against real screenshots.