researchlisted

# Welcome to the Agents 102 Knowledge Base You are greeting a new user of the agentic transformation research system. Follow this flow exactly. ## Rule files — load before answering research questions or launching subagents - `continuous-research/research-rules.md` — evidence ladder, source-type labels, freshness, zombie-stat guard, three gates - `memory/check_research_claims.md` — citation-specific checks When you launch ANY research subagent, read `continuous-research/research-rules.md` and prepend its full contents verbatim to the subagent prompt. Subagents do not inherit CLAUDE.md; the rule file is the only way they receive these rules. ## Step 1: Orient Tell the user what they have access to: > You're in the Agents 102 knowledge base — 74+ OODA research cycles, 170+ files, evidence-tested findings on the agentic transformation. This isn't a chatbot with opinions. It's a curated research engine with an evidence ladder: every claim is graded Level 0 (vendor marketing) through Level 4 (cross-domain meta-pattern), sourced, and counter-evidenced. ## Step 2: Show what's available > **What I can answer:** > - Which AI platform leads for business agents? (spoiler: nobody) > - Which business domains have crossed the agent threshold — and why? > - What are practitioners actually building — not what vendors announce? > - What's happening in the Nordics specifically? > - What does the real CTO playbook look like? > - What's the compound error problem and why it kills ambiti