📈 GH Trending Explorer 🎲 Surprise me

Search: moonshot-ai · clear

lordx64 / pentestkit

Autonomous multi-agent pentest framework — plans, exploits, verifies (proof-required), CVSS-scores and writes client-ready reports. 104/104 (100%) on the XBOW validation benchmarks, powered by Kimi K3.

▲ ~0/7d ★ 120 Python 2mo ago
ultralytics / lite

A fast, local workspace for Claude Code, Codex, Gemini CLI, Kimi Code, Qwen Code, DeepSeek, OpenRouter, and your shell.