research

Identify research questions and evaluate evidence across web and codebase sources.

198|19|Updated May 7, 2026
One-click install
npx skills add https://github.com/testdouble/han --skill research-testdouble
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: research
Source: https://github.com/testdouble/han/tree/main/han-core/skills/research
Command: npx skills add https://github.com/testdouble/han --skill research-testdouble

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Streamlines open-ended research by coordinating evidence gathering, diverse sources, and adversarial validation to surface an informed recommendation without committing to a single artifact.

Core Features & Use Cases

  • Orchestrates structured open-ended investigations across web sources, codebase context, and provided materials.
  • Produces a durable, evidence-backed report with an explicit recommendation and traceable sources.
  • Use cases include evaluating design trade-offs, surveying prior art, and understanding how a system works before choosing a direction.

Quick Start

Ask it to research architectural approaches for a chosen problem and return a compact, evidence-backed report with sources and a recommendation.

Frequently Asked Questions about research

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I conduct evidence-backed research for architectural decisions?

To conduct evidence-backed research for architectural decisions, identify an open-ended question and evaluate the minimal evidence required across web sources, codebase context, and provided materials. This process surfaces discrete options or recommendations with traceable sources and confidence levels.

What is the best way to survey prior art before choosing a design direction?

The best way to survey prior art before choosing a design direction is by orchestrating structured open-ended investigations across diverse sources. This approach produces a durable, evidence-backed report with an explicit recommendation or no clear winner if evidence is insufficient.

Can I evaluate design trade-offs using codebase context and provided materials?

You can evaluate design trade-offs using codebase context and provided materials by coordinating evidence gathering and adversarial validation. This method returns a compact, embedding-friendly paragraph with source references and an explicit recommendation.

How does adversarial validation work when researching a codebase?

When researching a codebase, adversarial validation works by determining the minimal evidence required to answer an open-ended question and evaluating it across multiple sources. It surfaces discrete options with confidence levels and an explicit recommendation or no clear winner.

What happens when evidence is insufficient to answer an open-ended research question?

When evidence is insufficient to answer an open-ended research question, the research process returns no clear winner instead of forcing a recommendation. This ensures the final compact paragraph maintains traceable source references and honest confidence levels.