Researching

What a research results document must store — per-claim provenance and a structured container — so a later consumer can trust and reuse it.

Domain-aware verification frameworks, a 0-100 trust rubric with domain ceilings, and the false-positive patterns that mark an untrustworthy claim.

Architecture, sourcing, citation, and verification lessons distilled from production deep-research agents.

The agentic research workflow beyond web search, plus the authoritative source / API / MCP catalogue organized by domain.

Classify a topic by intent and domain before researching; each type carries its own authoritative sources, method, and validation bar.