Each lens researches the same topic from a different angle (cyan), and everything converges into a single briefing. It's the mind map for Track 1.
Course path map
Detailed content
🎯 The single-prompt problem
Why a single prompt leaves gaps in research — and what the STORM method does differently.
STORM is a research method that brings together multiple perspectives on a topic instead of relying on a single angle. Here, it becomes the skill storm-research for Claude Code.
This is the foundation of everything: understanding the core idea explains why the pipeline has five lenses and a verification phase.
Multiple perspectives · synthesis · source verification · HTML briefing.
A single prompt carries a single frame—and everything that frame fails to see is left out of the research.
Recognizing the blind spot is what motivates asking for more than one perspective.
Framing bias · coverage · source diversity.
Deep research launches hundreds of generic agents; STORM uses a few deliberate, verified perspectives.
Shows that more agents don’t mean better research—structure matters more than volume.
Cost · rate limit · structure vs. volume.
A skill is a set of instructions (a master prompt) that Claude reads and executes when you invoke it.
Understanding that "skill = prompt" demystifies installation and shows that you can edit it.
SKILL.md · frontmatter · invocation · self-contained.
A standalone HTML file with a summary, findings ranked by reliability, and verified references.
Knowing what the deliverable looks like helps you judge whether STORM is what you need.
Autonomous · ranked · verified · actionable.
STORM shines on contested, multi-perspective topics; for a simple factual query, it’s overkill.
Using the right tool saves time and agent costs.
Contested topic · cost · fit.
🔭 The five expert lenses
Professional, academic, skeptic, economist, and historian — and why each one finds a hole the others miss.
The lens of someone who works with the topic every day: real-world friction, what works, and what tends to be overlooked.
Reveals the gap between theory and operational practice.
Friction · case study · operational reality.
The lens that cares about peer-reviewed studies and effect size, not anecdotes.
Grounds claims in rigorous evidence and flags where it is weak.
Peer review · effect size · preprint.
The lens that builds the strongest counterargument—rigorous, not contrarian for sport.
Exposes what supporters conveniently ignore.
Refutation · failure · contradictory data.
The lens that tracks revenue, valuation, funding, and incentives behind the narrative.
Shows who profits from the hype and how money shapes research.
Incentive · unit economics · market size.
The lens that looks for genuine historical parallels and what we learned from how they played out.
Set expectations: who won, who lost, and what stabilized.
Cycle · analogy · pattern.
Each lens has distinct priorities, so it sees what the others ignore. Together, they cover more.
This is the principle that justifies the entire panel.
Complementary coverage · diversity · blind spot.
🧱 Subagents, contradictions, and verification
The three pillars that make the result reliable—and why verification is what sets STORM apart from an ordinary report.
In STORM, the main session talks to five subagents — which don’t talk to one another (unlike an agent team).
Defines the pipeline’s cost and behavior.
Subagent · orchestration · agent team.
The five lenses were written by the same author; agreement among them is a strong hypothesis, not independent proof.
This is the method’s most important intellectual safeguard.
Author bias · convergence · honesty.
The step where direct conflicts between lenses are named and the stronger evidence is weighed.
This is the raw material for the findings and the “supports/challenges” chips.
Direct conflict · evidence hierarchy.
What all the lenses confirm (likely finding) and what none addressed (the missing sixth lens).
Feeds the "boundary question" and shows the panel’s limits.
Universal agreement · blind spot · sixth lens.
Rate by source quality: peer-reviewed causal evidence > official data > single research study > analogy > preprint.
Reliability isn’t subjective confidence—it’s the strength of the evidence.
Hierarchy · causal · official data.
Check every citation against the actual primary source, correcting, downgrading, or removing anything that doesn’t hold up.
Without this phase, it isn’t a STORM report.
Primary source · verdict · verification banner.