What happens if a research question remains unresolved?

If a subagent finds no useful information, the question is marked as unresolved and carried forward to be flagged for a user decision or deferred to technical planning.

How does Iterative Research differ from technical planning?

Technical planning focuses on codebase exploration and implementation details, whereas Iterative Research focuses on external constraints, scope requirements, and industry patterns.

Does it make changes to my project files automatically?

No, the skill follows a 'present before committing' principle, showing you findings and proposed document updates for your approval before any changes are applied.

Why does the skill use parallel subagents?

Research questions are typically independent of one another. Spawning parallel subagents allows Claude to investigate multiple unknowns simultaneously, significantly reducing the time spent on discovery.

Can I use this skill without an existing PRD?

Yes, the research skill can be invoked standalone with a list of questions to investigate even if you haven't created a formal PRD yet.

Iterative Research

Name: Iterative Research
Author: tmchow

bytmchow

•

生産性とワークフロー

Investigates and resolves open project questions by conducting parallel research into prior art, constraints, and competitive landscapes.

Iterative Research is a Claude Code skill designed to bridge the gap between initial brainstorming and technical planning by systematically resolving project unknowns. It analyzes Product Requirement Documents (PRDs) or user-provided questions, categorizes them based on research needs, and spawns parallel subagents to investigate external constraints, industry patterns, and competitive analysis. By synthesizing findings and automatically proposing updates to your project documentation, it ensures that your engineering foundation is built on verified data rather than assumptions, while maintaining a human-in-the-loop workflow for all changes.

主な機能

01Intelligent question categorization to distinguish between research, design, and technical planning

02Context-aware investigation across web search, codebase patterns, and documentation

03Parallel research subagents for concurrent investigation of multiple independent questions

0420 GitHub stars

05Interactive approval workflow for all proposed documentation updates and scope changes

06Automatic PRD synchronization that moves resolved questions into formal requirements

ユースケース

01Identifying external regulatory or technical constraints that impact project scope and boundaries

02Resolving 'Open Questions' sections in a PRD before moving to the technical implementation phase

03Conducting competitive analysis to understand industry standards and prior art for a new feature

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tmchow/tmc-marketplace research

For use in Claude.ai and ChatGPT

Download Skill