diagnose
Problem Analysis — Orchestrator
Strategy — Step 1 of 4. Defines the problem, forms testable hypotheses, and identifies root causes.
Core Question: "What's actually causing this, and what does the evidence prove?"
Critical Gates — Read First
- Problem statement MUST be: "[Metric] is [current] instead of [target]." No vague problems. If the user says "things aren't going well," interview for the specific metric, current value, and target value before dispatching any agent.
- Do NOT skip external factors — 30%+ of problems have external causes. The external-check-agent runs in Layer 1 alongside the tree builder. Skipping it leads to treating a symptom when the cause is environmental.
- If/Then/Because format required — hypotheses without "because" are unfalsifiable. The "because" clause explains the mechanism. Without it, a rejected hypothesis teaches nothing.
- Do NOT confirm hypotheses without evidence — "seems likely" is not Confirmed. Every verdict must cite a specific data point that matches or contradicts the "then" clause. Inconclusive is a valid verdict.
Inputs Required
- A problem (metric decline, performance gap, strategic question)
More from hungv47/research-skills
market-research
Analyzes market landscapes, competitive dynamics, TAM/SAM/SOM sizing, and whitespace opportunities for a product or category. Produces `research/market-research.md`. Not for building customer personas (use icp-research) or planning marketing campaigns (use campaign-plan). For diagnosing a specific business problem, see diagnose. For prioritizing what to build from market data, see prioritize.
10icp-research
Builds ideal customer profiles and buyer personas — analyzes demographics, pain points, jobs-to-be-done, and segmentation for a target market. Produces `research/icp-research.md`. Not for competitive positioning (use prioritize) or campaign planning (use campaign-plan). For brand identity from audience data, see brand-system. For market sizing and competitor landscape, see market-research.
10problem-analysis
Structured diagnosis of business and strategic problems — builds logic trees, forms testable hypotheses, and identifies root causes with evidence. Produces `.agents/problem-analysis.md`. Not for code bugs (use code-cleanup) or brainstorming solutions to a known problem (use solution-design). For market-level trends and competitive context, see market-research. For testing hypotheses with experiments, see experiment.
9funnel-planner
Models business funnels with numeric targets — works backward from revenue goals to required traffic, conversion rates, and unit economics. Produces `.agents/skill-artifacts/meta/records/targets-*.md`. For campaign planning, see campaign-plan.
9solution-design
Brainstorms and prioritizes strategic solutions when the problem or goal is already clear — generates options, scores trade-offs, and recommends a path forward. Produces `.agents/solution-design.md`. Not for diagnosing what the problem is (use problem-analysis) or engineering task lists (use task-breakdown). For setting numeric targets after prioritizing, see funnel-planner. For technical architecture of chosen initiatives, see system-architecture.
9experiment
Designs validation experiments for business ideas — defines hypotheses, success metrics, sample sizes, and minimum viable tests before full commitment. Produces `.agents/experiment-[name].md`. Not for strategic prioritization (use solution-design) or market sizing (use market-research). For diagnosing root causes, see problem-analysis. For campaign planning, see imc-plan.
7