Autonomous research orchestration with dual-loop architecture

★ 7.7 · research

0-autoresearch-skill is a Claude Code skill that orchestrates end-to-end autonomous AI research projects using a two-loop architecture: the inner loop runs rapid experiment iterations with clear measurable outcomes, while the outer loop synthesizes results, identifies patterns, and steers research direction. On initialization it creates a structured workspace with research-state.yaml, research-log.md, findings.md, and directories for literature, experiments, data, and the final paper. The skill routes execution to domain-specific skills rather than acting as a domain expert itself, sustains continuous agent operation via /loop and OpenClaw heartbeat, and produces HTML/PDF progress presentations along with a final paper through ml-paper-writing. It is designed for starting a research project from scratch, running multi-hypothesis autonomous experiments, or resuming an interrupted research session from saved state.