LibraNia

Mandatory first step for ANY research, investigation, explanation, planning, comparison, or lookup task in this project (per CLAUDE.md). Auto-searches LibraNia before web, saves findings back. Triggers broadly, not just on exact phrasing — covers: factual/conceptual questions ('what is X', 'why does X happen', 'how does X work'), explain/describe/clarify requests, comparisons ('X vs Y', 'difference between'), documentation/API/reference lookups, learning requests, meta/architecture questions about tools or concepts, investigation of behavior or root causes, and planning/decision questions that need background context. If the request is asking to understand or find out something rather than to write/edit/run code, this skill applies.

Use when researching a complex topic that needs multi-angle deep analysis with cross-validation — "research X thoroughly", "spawn agents to research X", "get multiple perspectives on X", or any research task where a single pass would miss technical, competitive, practical, business, and risk blind spots. Also use when the user asks to verify or fact-check a research document using subagents.