Multi-agent orchestration for AI coding agents — pluggable runtime adapters for Claude Code, Pi, and more
Terminal score: 0–1000 raw, weighted across 4 dimensions. Public score: 0–10 normalized (shown in the 30-day stars chart above).