The Science One framework introduces a Chain-of-Evidence mechanism to ensure autonomous research is verifiable. It forces agents to link every claim to a specific, traceable source. This reduces hallucinations in complex scientific discovery. Researchers can now audit AI-generated hypotheses with precision, moving beyond the black-box nature of standard LLM reasoning.