The Science One framework introduces a Chain-of-Evidence mechanism to make autonomous research verifiable. It forces AI agents to link every scientific claim to a specific, traceable source. This prevents hallucinations in complex data synthesis. Researchers can now audit the logic path of automated discoveries, reducing the manual verification burden for lab scientists.