A lightweight Python MCP server gives AI agents free DuckDuckGo search and webpage fetching, with a new curl backend to slip past Cloudflare bot filters.
A Rust CLI collapses git worktree management into three commands, letting you run Claude Code and Codex agents side by side without stepping on each other.
Nous Research's Lighthouse Attention cuts long-context LLM pre-training time by 1.4–1.7x, running 17x faster than standard attention at 512K context — with no custom sparse kernel and no quality loss at inference.