idea_refinery

Automate research idea refinement from seed capture to final proposal documents.

2|Updated Mar 20, 2026
One-click install
npx skills add https://github.com/Gonglitian/agent-skills --skill idea-refinery
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: idea_refinery
Source: https://github.com/Gonglitian/agent-skills/tree/main/skills/idea_refinery
Command: npx skills add https://github.com/Gonglitian/agent-skills --skill idea-refinery

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It guides researchers from a rough seed idea to a concrete, validated proposal by orchestrating a structured, phase-based workflow that captures insights and decisions in an auditable trail.

Core Features & Use Cases

  • Phase-driven workflow: workspace setup, seed capture, topic survey, seed validation, direction exploration, and converge.
  • Artifact generation: creates idea cards, survey cache, direction reports, validation documents, and final proposal.
  • Parallel exploration: supports evaluating multiple directions in parallel and merging results.

Quick Start

Provide the seed idea and constraints to initialize the workspace; the system will generate the v0_seed card and set up the phased workflow.

Frequently Asked Questions about idea_refinery

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

FAQPage Schema
How do I turn a rough research idea into a polished proposal?

To validate a research idea, the system performs topic surveys and seed validation to test feasibility before committing to full proposal generation. This structured validation phase ensures your direction is viable and creates an auditable trail of insights and decisions.

Can I explore multiple research directions in parallel during proposal generation?

Yes, you can explore multiple research directions in parallel during proposal generation. The system supports evaluating multiple branches simultaneously and merging the results, allowing rapid convergence on the most viable path while generating comprehensive direction reports.

What is the best way to structure a literature survey for idea refinement?

The best way to structure a literature survey for idea refinement is to automate the generation of a survey cache within a phase-driven workspace. This cache captures topic insights and integrates them into an auditable workflow, ensuring your validation and convergence phases are grounded in existing research.

How do I start refining a seed concept into a research workflow?

To start refining a seed concept, you initialize the workspace by providing your rough idea and any constraints. The system then generates a v0_seed card and sets up the phased research workflow, guiding you from initial capture through topic survey and validation to final convergence.

Does idea refinement require external dependencies to generate direction reports?

No external dependencies are required to generate direction reports or conduct parallel exploration. The system handles the end-to-end refinement internally, creating structured artifacts like idea cards and validation documents without relying on external components.