generate-idea

To generate and validate research ideas, this Skill combines literature reviews to identify gaps, uses external LLMs for brainstorming, and runs pilot experiments to test feasibility. It outputs a structured report ranking ideas by novelty, risk, and expected contributions.

The best way to rank research ideas by feasibility and novelty is through a structured filtering process. This Skill evaluates brainstormed concepts against existing literature and pilot experiment results to provide a ranked list with detailed feasibility, risk, and impact assessments.

Yes, you can run pilot experiments to test research ideas automatically. The Skill designs and executes pilot validations for top-ranked ideas based on your specified budget, such as a 4-hour limit per idea, to ensure efficient exploration of new research directions.

Yes, this research idea generation workflow requires Python and Bash, and utilizes the codex MCP component. These dependencies support the internal scripts that automate the literature review, LLM brainstorming, and pilot experiment execution processes.

To start brainstorming research ideas for deep learning in healthcare, run the Skill with your specified topic and pilot experiment budget. It automatically reviews relevant literature, generates multiple concepts, and validates the most promising ones via pilot experiments.

A key limitation when filtering research ideas with automated literature reviews is the dependency on LLM availability and pilot experiment budgets. The validation process is constrained by the specified time limits and the external LLM's ability to accurately identify research gaps.