researchlisted

# Research — Multi-Source Decision Intelligence Systematic research pipeline that gathers evidence from multiple sources before forming a recommendation. Every verdict is filtered through Hai's roadmap, career anchors, and current phase capacity. This skill exists because Claude's training data alone is insufficient for decisions that depend on current market state, personal context, or accumulated past decisions. The pipeline forces evidence-based answers by checking what Hai already knows (brain + Readwise) before hitting the web, then contextualizing findings against who Hai is and where he's headed. **Design principles:** - **Evidence over opinion** — gather first, synthesize second, recommend last - **Cost discipline** — cheap sources first (brain → chub → Context7 → Web). Don't burn tokens on web searches when the answer is already cached - **Clear stance** — no "it depends." State a position with reasoning. If evidence is genuinely split, say so and explain which side you lean toward and why - **Profile-aware** — a technically correct answer that conflicts with Hai's roadmap, values, or capacity is still wrong for him right now ## Pre-flight Run once per session, not per invocation: 1. `brain_recall("roadmap phase brick capacity")` — determine: - Active brick (which Prove slot is occupied) - Current phase (0 Pre-Baby / 1 Survival / 2 Stabilize / 3 Build) - Capacity level 2. `brain_recall("career anchors goals priority")` — load career anchors and priorit