pragman-researchlisted

# Pragman Research Research starts with the decision it must enable. Match effort to consequence, use the strongest available sources, and finish with an actionable synthesis rather than a reading list. ## Use this skill when - A choice depends on current facts, primary documentation, market evidence, technical comparisons, or bounded internal context. - You need a source plan, freshness boundary, confidence, and explicit stopping rule rather than an unbounded reading list. - Multiple research capabilities may help, but private data and external egress must remain deliberate. ## Acceleration payoff It spends research effort where it can change the decision, separates facts from inference, and finishes with a clear action and revisit signal. ## Workflow 1. Define the decision, audience, deliverable, deadline, required freshness, acceptable uncertainty, and excluded scope. Ask only for missing constraints that materially change method, access, privacy, or cost. For quick reconnaissance, state before retrieval the minimum evidence needed to choose between stopping and deeper research, plus a time or source cap. 2. Select the lightest sufficient method from [references/research-methods.md](references/research-methods.md). Use independent fan-out only when lanes do not depend on one another and their outputs can be reconciled deterministically. 3. Route needed capabilities through `pragman-router`. Reuse configured research, browser, repository, and workspace providers. If