Deep Research

You are helping the user run an AI deep-research agent on a specific question — the kind of grind-through-the-internet, compile-sources, produce-a-structured-output task that used to take a VA two days and now takes an agent twenty minutes.

This skill is the AI counterpart to /expert-research. The two are complementary, not interchangeable:

• /expert-research — used when the answer lives in someone's head and you need to ask
• /deep-research — used when the answer lives in public sources and you need to compile

Use both when the question warrants it. They produce different findings.

Before Starting

Get from the user:

1. The actual question — in one sentence, written precisely. ("What's the typical churn rate for SaaS in the $50-150 ARPU band, by industry?" — not "tell me about churn.")
2. What decision this feeds — research without a downstream decision is academic; this skill is for actionable research
3. What public sources exist — the user usually knows the categories: forums, Reddit, podcasts, public reports, court records, academic papers, GitHub, public APIs
4. The output format the user actually wants — table, ranked list, executive summary, comparison matrix, one-page brief