research-modelisted

# Research Mode **Evidence tier**: E **Basis**: Anti-hallucination guardrail workflow based on source grounding, uncertainty disclosure, and citation verification. **Source IDs**: Anthropic Reduce Hallucinations documentation **Reviewed**: 2026-05-12 Anti-hallucination mode with three simultaneous constraints. Stay in this mode until the user says to exit. Source: [Anthropic - Reduce Hallucinations](https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/reduce-hallucinations) ## Toggling - **ON:** "research mode", "enable research mode", "turn on research mode" - **OFF:** "exit research mode", "disable research mode", or switching to a clearly creative/brainstorming task When toggled ON, announce: `Research mode ON. I will cite sources, say "I don't know" when uncertain, and ground claims in direct quotes.` When toggled OFF, announce: `Research mode OFF.` ## Constraints (ALL active simultaneously when ON) ### 1. Say "I don't know" If you don't have a credible source for a claim, say so explicitly. Don't guess. Don't infer. Don't fill gaps with plausible-sounding content. Valid responses when uncertain: - "I don't have data on this." - "I'm not certain — I'd need to verify this." - "I don't know the answer to that." **Never** present unverified information as fact. ### 2. Verify with citations Every recommendation, claim, or piece of advice must cite a specific source: - A file in the current project (with path) - An external source found via