idea-creator

Generate and rank publishable research ideas from a broad research direction.

Updated May 29, 2026
One-click install
npx skills add https://github.com/Mang30/myskills --skill idea-creator-mang30
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: idea-creator
Source: https://github.com/Mang30/myskills/tree/main/skills/idea-creator
Command: npx skills add https://github.com/Mang30/myskills --skill idea-creator-mang30

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Idea-creator helps you quickly transform a broad research direction into concrete, publishable research ideas with feasibility and novelty filtering, so you can spend less time brainstorming and more time testing.

Core Features & Use Cases

  • Landscape survey: Scans existing local papers and recent online literature to map sub-directions, gaps, and open problems.
  • Idea generation & ranking: Produces 8–12 differentiated ideas, filters them for feasibility and potential impact, then deep-validates top candidates with novelty checks and critical review.
  • Pilot validation: Runs lightweight pilot experiments in parallel (when GPUs are available) to rank ideas based on empirical signal, not just intuition.

Quick Start

Generate ranked publishable research ideas for your topic by running: /idea-creator "your research direction"

Frequently Asked Questions about idea-creator

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

FAQPage Schema
How do I generate and rank publishable research ideas from a broad topic?

Research idea generation surveys literature to map gaps, brainstorms 8-12 hypotheses, and filters them by feasibility, novelty, and impact to produce ranked publishable directions.

How does novelty-check filtering work during research ideation?

Novelty-check filtering deep-validates top research candidates by scanning existing literature and recent papers to ensure generated ideas have not been previously published or explored.

Can I run pilot experiments to validate research ideas before committing to full tests?

Pilot validation runs lightweight parallel experiments when GPUs are available, ranking research ideas based on empirical signal rather than intuition before committing to full experiments.

Do I need a research-wiki folder to use the literature survey workflow?

A research-wiki folder is optional for literature surveys, enabling integration if the directory exists, while standard outputs are reported to idea-stage/IDEA_REPORT.md without requiring it.

What's the best way to manage GPU budgeting for parallel pilot validation?

GPU budgeting allocates available compute resources to run lightweight pilot experiments in parallel, validating multiple research ideas simultaneously to rank them by empirical signal.