Step 0 Research

Step 0 · Research

Command: /research [topic] · Habit: H5 Seek First to Understand · Position: Optional pre-step before /requirements

Purpose

Investigate the problem space before writing requirements. Prevents specifying the wrong thing — the most expensive bug class.

When to use

• Problem space is unclear or unfamiliar
• Multiple existing solutions exist and you need a comparison
• Domain has non-obvious constraints (regulatory, performance, security)
• You catch yourself thinking "I'll just start coding and see"

When to skip

• You have deep domain knowledge already
• Requirement is crystal clear (e.g. "fix typo in header")
• Prior ADR already covers the decision space

Process

1. Define the question — 1 sentence, specific
2. Gather prior art — existing code, docs, web research, comparable libraries
3. Compare — if evaluating alternatives, produce a trade-off matrix
4. Surface constraints — legal, performance, team skills, budget
5. Recommend (not decide) — AI recommends; human decides in Step 2

Output

A research brief containing:

• Problem statement
• Findings summary
• Comparison matrix (if applicable)
• Constraints
• Recommendation + rationale
• Open questions

Handoff

• Expects: A topic or question from the user
• Produces for /requirements: Research brief that informs PRD scope and success criteria

H5 Checkpoint

"Have I fully understood the problem before proposing a solution?"

See also

• Source: skills/research/SKILL.md
• Habits Reference → H5
• Next: Step 1 · Requirements