idea-discovery

Generate validated research ideas with literature surveys, novelty checks, and experiment plans.

Updated May 25, 2026
One-click install
npx skills add https://github.com/duypham2801/ThS_LLM --skill idea-discovery-duypham2801
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: idea-discovery
Source: https://github.com/duypham2801/ThS_LLM/tree/main/.claude/skills/idea-discovery
Command: npx skills add https://github.com/duypham2801/ThS_LLM --skill idea-discovery-duypham2801

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It solves the problem of turning a broad research direction into a set of validated, pilot-tested ideas with a concrete experiment plan, instead of getting stuck at vague brainstorming.

Core Features & Use Cases

  • End-to-end Idea Discovery Pipeline: Runs a complete workflow that chains literature survey, idea generation, novelty verification, critical review, and method refinement into a final proposal.
  • Checkpointed Validation with Pilots: Filters and ranks ideas, runs time/budget-bounded pilot experiments, then performs deeper novelty checks to eliminate already-published directions.
  • Deliverables for Implementation: Produces a ranked idea-stage/IDEA_REPORT.md plus refine-logs/FINAL_PROPOSAL.md and refine-logs/EXPERIMENT_PLAN.md for the top idea.

Quick Start

Run idea discovery for a topic by executing: use the idea-discovery skill with the research direction you want to explore, for example: /idea-discovery "chest X-ray phrase grounding with bounding boxes — semi-supervised grounded detection".

Frequently Asked Questions about idea-discovery

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

FAQPage Schema
How do I generate validated research ideas from a broad topic?

Research idea generation works by orchestrating literature surveys, ideation, and novelty verification to transform broad directions into validated concepts. It chains critical review and method refinement into a final, publication-aware proposal.

How do I check if my research idea is novel before writing a proposal?

Novelty verification filters and ranks generated ideas by running deeper checks against existing literature. This eliminates already-published directions during the academic workflow, ensuring your proposal remains unique.

How do I plan pilot experiments for a research hypothesis?

Pilot experiment planning involves running time and budget-bounded tests on filtered ideas to validate hypotheses. The process outputs a concrete experiment plan and final proposal for the top-ranked idea.

What is the best way to automate an academic literature survey for research ideation?

Automating a literature survey for research ideation involves chaining it with idea generation and critical review. This end-to-end pipeline produces ranked, publication-aware proposals and deliverables like IDEA_REPORT.md.

Can I use this research ideation workflow for machine learning domains?

Yes, the research ideation workflow applies to academic R&D workflows across machine learning and other research domains. It generates validated ideas and experiment plans tailored to these specific fields.

What deliverables should I expect from an end-to-end research ideation pipeline?

Deliverables from a research ideation pipeline include a ranked IDEA_REPORT.md, a FINAL_PROPOSAL.md, and an EXPERIMENT_PLAN.md. These outputs provide concrete implementation steps for the top validated idea.