What documentation does the Research skill generate?

It generates a central FEATURE.md file for high-level design and individual spec-research.md files for every decomposed unit of work.

How does it help with large feature requests?

It breaks down large changes into smaller, independent technical specs (200-400 lines each) to ensure the resulting code is mergeable and easy to review.

When should I use the Research skill?

Use this skill at the start of any new work, feature investigation, or refactoring project to ensure a clear architectural plan exists before implementation starts.

Does this skill interact with the user during the process?

Yes, it is designed to be interactive, using AskUserQuestion at key decision points like scope confirmation, architectural approach selection, and testing strategy.

Benji Research Workflow

Name: Benji Research Workflow
Author: AI-Builder-Team

byAI-Builder-Team

•

生産性とワークフロー

Explores complex problem spaces, designs architectural approaches, and decomposes large-scale features into implementable technical specifications.

The Research skill is a comprehensive workflow designed for the 'fuzzy front end' of software development. It guides Claude through a structured process of understanding requirements, performing deep codebase exploration to identify patterns, and engaging in interactive decision-making with the user. By formalizing architectural designs, interface contracts, and test plans before coding begins, it ensures that complex changes are decomposed into manageable, high-quality technical specs that are ready for implementation and PR review.

主な機能

012 GitHub stars

02Automated test plan generation derived directly from interface contracts

03Automated codebase exploration and API pattern discovery

04Multi-spec decomposition with dependency mapping and complexity estimation

05Formal interface contract definition with source tracing and TBD resolution

06Interactive decision checkpoints for scope and architecture confirmation

ユースケース

01Planning a major new feature or enhancement within an existing codebase

02Decomposing large-scale architectural refactors into small, reviewable pull requests

03Investigating complex bugs and designing a robust, multi-step fix strategy

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ai-builder-team/ai-builder-plugin-marketplace research

For use in Claude.ai and ChatGPT

主な機能

012 GitHub stars

02Automated test plan generation derived directly from interface contracts

03Automated codebase exploration and API pattern discovery

04Multi-spec decomposition with dependency mapping and complexity estimation

05Formal interface contract definition with source tracing and TBD resolution

06Interactive decision checkpoints for scope and architecture confirmation

ユースケース

01Planning a major new feature or enhancement within an existing codebase

02Decomposing large-scale architectural refactors into small, reviewable pull requests

03Investigating complex bugs and designing a robust, multi-step fix strategy

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ai-builder-team/ai-builder-plugin-marketplace research

For use in Claude.ai and ChatGPT