Anthropic Brings Software Testing Rigor to AI Agent Skills

Claude's skill-creator update adds evals, benchmarks, and A/B testing for non-engineers building AI agent skills. Here's what it means for the ecosystem.