Research Companion
Adapted from https://github.com/andrehuang/research-companion
Strategic research thinking agents for Claude Code — idea evaluation, project triage, and structured brainstorming to help you do research that matters.
Most AI writing tools help you write papers. This plugin helps you decide which papers to write.
Inspired by Nicholas Carlini's essay "How to Win a Best Paper Award" — which argues that great research starts with taste, strategic problem selection, honest self-evaluation, and knowing when to kill your darlings.
The Problem
Researchers don't lack the ability to write papers. They lack a trusted colleague who will:
- Tell them an idea isn't worth 6 months of their life — before they invest those months
- Ask "who else is working on this and what's your unfair advantage?"
- Challenge them to state the key insight in one sentence (and refuse to move on until they can)
- Help them find the unexpected cross-field connection that makes a contribution truly novel
- Evaluate whether a struggling project should be continued, pivoted, or killed
This plugin provides that colleague.
What's Inside
Agents
| Agent | What it does |
|---|
| Brainstormer | Creative brainstormer with emphasis on cross-field connections, strategic ignorance (challenging flawed assumptions), and surprising ideas. |
| Idea Deepener | Takes rough brainstorm-level ideas and fleshes them out into concrete one-pager elevator pitches. |
| Paper Crawler | Systematic collection of related papers from DBLP and OpenAlex APIs. |
| Research Analyst | Maps the research landscape around an idea — closest related work, landscape density, key differentiators. |
| Idea Critic | Stress-tests research ideas along 6 dimensions: novelty, impact, timing, feasibility, the nugget, and narrative potential. Returns a Pursue / Park / Kill verdict. |
| Idea Reviser | Revises a pitch based on landscape analysis and evaluation feedback — generates a meaningful variation addressing identified weaknesses. |
| Proposal Drafter | Expands a one-pager into a concrete two-page research proposal with draft abstract, conclusion-first test, and experimental plan. |
| Research Strategist | On-demand project-level strategic thinking — triage (continue/pivot/kill), comparative advantage mapping, impact forecasting, and scooping risk assessment. |
Skill
| Skill | What it does |
|---|
/research-companion | A structured multi-phase ideation session that orchestrates specialized agents through: Seed → Diverge → Deepen → Compare → Evaluate → Finalize. Includes researcher checkpoints, a revision loop, and Carlini's conclusion-first test. |
Principles
8 research strategy principles organized into three categories (Problem Selection, Execution Strategy, Strategic Positioning) that guide the agents' evaluations.
Installation
Claude Code
claude plugin marketplace add https://github.com/frasalvi/research-companion
claude plugin install research-companion@frasalvi-research-companion
Usage
Evaluate a research idea
Just describe your idea and ask for evaluation:
I'm thinking about studying how LLM-generated code introduces subtle security
vulnerabilities that pass standard code review. Can you evaluate this idea?
The Idea Critic will evaluate across 6 dimensions and give you a verdict.
Decide whether to continue a project
I've been working on adversarial attacks against multimodal models for 3 months.
I have some results but they're incremental. Two other groups just posted preprints
in the same area. Should I continue?
The Research Strategist will assess your competitive position, impact potential, and opportunity cost, then recommend Continue / Pivot / Kill.
Run a full brainstorming session
In Claude Code:
/research-companion I'm interested in the intersection of program synthesis
and scientific discovery
This launches a 6-phase guided session:
- Seed — Understand your problem space, interests, and what bugs you about the field
- Diverge — Generate ideas, alternative framings, and cross-field connections
- Deepen — Flesh out the top 2-3 ideas into concrete one-pager elevator pitches
- Compare — Map the research landscape: related work, closest competitors, landscape density
- Evaluate — Stress-test ideas along 6 dimensions, deliver Pursue / Park / Kill verdicts. You decide per idea: advance, revise, or drop.
- Finalize — Expand surviving ideas into two-page research proposals with draft abstract, conclusion-first test, and experimental plan
Find cross-field connections
frasalvi
