research-refine

Refine vague research directions into implementation-ready method plans.

Updated Apr 21, 2026
One-click install
npx skills add https://github.com/Shallow-W/llm-wiki --skill research-refine-shallow-w
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: research-refine
Source: https://github.com/Shallow-W/llm-wiki/tree/main/.claude/skills/research-refine
Command: npx skills add https://github.com/Shallow-W/llm-wiki --skill research-refine-shallow-w

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It helps you convert an initially vague research direction into a concrete, problem-anchored, implementation-ready method plan that is frontier-aware and suitable for top-venue scrutiny.

Core Features & Use Cases

  • Problem Anchor Preservation: Freezes an immutable problem statement and carries it through every iteration to prevent drift.
  • Method-First, Elegance-First Refinement Loop: Builds an initial proposal from targeted grounding, then iteratively reviews and refines it with emphasis on mechanism clarity and contribution focus.
  • Claim-Driven Validation Design: Produces minimal validation experiments tied to the core claims, avoiding benchmark or module sprawl.
  • Checkpoint Recovery: Persists state across phases to support interruption tolerance and safe resumption.

Quick Start

Use the research-refine skill when you want to refine my approach for a research problem into a concrete, implementation-oriented plan, and include your problem plus your current vague approach.

Frequently Asked Questions about research-refine

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

FAQPage Schema
How do I refine a vague research idea into a paper-ready method plan?

Problem decomposition prevents research drift by freezing an immutable problem statement carried through every iteration. It anchors the refining process so the final method plan remains focused on the original issue rather than sprawling into unrelated technical routes.

How do I design validation experiments for a new research method?

Design validation experiments for a new research method by using a claim-driven approach that ties minimal tests directly to core contributions. This avoids benchmark sprawl and ensures your validation path strictly supports the specific claims of your method plan.

Can I resume my research proposal refinement after an interruption?

You can resume research proposal refinement after an interruption because the process uses checkpointed round-by-round output. Checkpoint recovery persists state across phases, supporting interruption tolerance and safe resumption without losing your iterative review progress.

What is the best way to specify technical mechanisms for a research proposal?

The best way to specify technical mechanisms for a research proposal is through concrete mechanism specification during an iterative refinement loop. This requires detailing interfaces, losses, and inference paths to ensure the method plan is implementation-ready and frontier-aware.

When should I not use an iterative refinement loop for paper writing?

You should not use an iterative refinement loop for paper writing if your technical route is already concrete rather than fuzzy. This approach specifically targets scenarios where the problem is visible but the implementation-oriented method plan lacks front-aware technical leverage.