AGENTS.MD OUTPERFORMS SKILLS IN OUR AGENT EVALS

Summary from the TLDR Dev newsletter

AI coding agents struggle with outdated framework knowledge, leading to incorrect code generation. Researchers from Vercel evaluated two approaches for providing current Next.js API documentation: "skills" (on-demand knowledge packages) and "AGENTS.md" (a persistent, embedded markdown file). Surprisingly, a compressed 8KB docs index embedded directly in AGENTS.md achieved a perfect 100% pass rate in their evals, outperforming skills, which only reached 79% even with explicit triggering instructions.