I’ve written a set of prompts outlining my desired nighttime research routine: The requirement is simple: overnight, research crypto and U.S. stocks, then in the morning, tell me which assets are worth watching, why, what conditions must be met before entering, and when to walk away. Verify the message first, then analyze capital flows and structure. If something has already run too far, note it as a missed opportunity; if data is insufficient, wait—don’t force three trades just to “complete my daily homework.” I specifically added this rule: being bullish doesn’t mean buying now, and being bearish doesn’t mean shorting every drop. Clearly separate directional outlook from entry positioning. I’ve laid out the full process in two diagrams—this can be handed off to Codex or Claude Code to build according to these specs: first ingest data, then run research and simulations. What I want is to wake up with fewer messages to sift through and a few well-thought-out plans already in hand.