research-planner

name	research-planner
description	Turn a question into a bounded evidence plan with subquestions, source strategy, freshness needs, and synthesis criteria. Use when a request requires external facts, current information, comparison, or deep investigation.
license	MIT
metadata	{"category":"research","origin":"custom-synthesis","revision":"1.1.0"}

Research Planner

Trigger

• Research scope is broad or ambiguous.
• Several source types must be combined.

Inputs

• Research question and intended decision
• Geographic/time scope
• Quality and citation requirements
• Known sources

Procedure

1. Define the decision the research must support.
2. Break the question into factual, comparative, causal, implementation, and risk subquestions.
3. Set inclusion/exclusion criteria and freshness threshold.
4. Map each subquestion to primary sources first, then high-quality secondary corroboration.
5. Plan search terms, aliases, repositories, standards, and document versions.
6. Define stop conditions: saturation, source agreement, or decision sufficiency.
7. Create an evidence table before synthesis.

Output contract

• Research plan
• Subquestion/source matrix
• Evidence schema
• Stop and confidence criteria

Quality gates

• The plan can distinguish absence of evidence from evidence of absence.
• Current claims have a freshness strategy.
• Primary sources carry the most weight.
• The plan is bounded enough to complete.

Handoffs and dependencies

• documentation-research
• web-evidence-synthesis
• source-verification
• repository-analysis

Boundaries

• Do not search broadly without a decision question.
• Do not predetermine the conclusion.
• Do not use citation count as a proxy for source quality.