Skip to content
Back to search

research-companion

Strategic research thinking agents for Claude Code — idea evaluation, project triage, and structured brainstorming. Helps you decide which papers to write, not just how to write them.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/andrehuang--research-companion/badge.svg)](/listings/andrehuang--research-companion)

HTML

<a href="/listings/andrehuang--research-companion"><img src="/api/listings/andrehuang--research-companion/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

Claude Code

claude plugin marketplace add https://github.com/andrehuang/research-companion
claude plugin install research-companion@andrehuang-research-companion

Codex

See [docs/codex-installation.md](./docs/codex-installation.md) for the full Codex installation guide, including:

- repo marketplace installation - personal local marketplace installation - official Codex plugin docs links

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 19 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. {
  2. "name": "andrehuang-research-companion",
  3. "description": "Strategic research thinking agents for idea evaluation, project triage, and structured brainstorming",
  4. "owner": {
  5. "name": "Andre Huang",
  6. "email": "andre@example.com"
  7. },
  8. "plugins": [
  9. {
  10. "name": "research-companion",
  11. "description": "Idea critic, research strategist, enhanced brainstormer, and structured brainstorming skill for research ideation",
  12. "source": "./",
  13. "category": "productivity",
  14. "version": "1.0.0",
  15. "license": "MIT",
  16. "keywords": ["research", "brainstorming", "strategy", "academic", "idea-evaluation"]
  17. }
  18. ]
  19. }
README.md · README · 200 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # Research Companion
  2. **Strategic research thinking agents for [Claude Code](https://docs.anthropic.com/en/docs/claude-code)** — idea evaluation, project triage, and structured brainstorming to help you do research that matters.
  3. This repository now also includes native [Codex](https://openai.com/codex/) support.
  4. Most AI writing tools help you *write* papers. This plugin helps you decide *which* papers to write.
  5. Inspired by Nicholas Carlini's essay ["How to Win a Best Paper Award"](https://nicholas.carlini.com/writing/2026/how-to-win-a-best-paper-award.html) — which argues that great research starts with taste, strategic problem selection, honest self-evaluation, and knowing when to kill your darlings.
  6. > **📢 Update (April 2026): Research Companion is now also a component of [`researcher-pack`](https://github.com/andrehuang/researcher-pack)** — a more advanced, all-in-one Claude Code plugin for researchers that bundles strategic brainstorming, paper reading, literature tracking, academic writing, and project management into a single integrated workflow. This standalone repo is still actively maintained, and any updates to Research Companion will be synced between both repos — so you can use it here on its own, or get the full toolkit over at [`researcher-pack`](https://github.com/andrehuang/researcher-pack). Feedback, issues, and PRs are very welcome in either place! 🙏
  7. ## The Problem
  8. Researchers don't lack the ability to write papers. They lack a trusted colleague who will:
  9. - Tell them an idea isn't worth 6 months of their life — *before* they invest those months
  10. - Ask "who else is working on this and what's your unfair advantage?"
  11. - Challenge them to state the key insight in one sentence (and refuse to move on until they can)
  12. - Help them find the unexpected cross-field connection that makes a contribution truly novel
  13. - Evaluate whether a struggling project should be continued, pivoted, or killed
  14. This plugin provides that colleague.
  15. ## What's Inside
  16. ### Agents
  17. | Agent | What it does |
  18. |-------|-------------|
  19. | **Idea Critic** | Stress-tests research ideas along 7 dimensions: novelty, impact, timing, feasibility, competitive landscape, the nugget, and narrative potential. Returns a Pursue / Refine / Kill verdict. |
  20. | **Research Strategist** | Project-level strategic thinking — triage (continue/pivot/kill), comparative advantage mapping, impact forecasting, opportunity cost analysis, and scooping risk assessment. |
  21. | **Brainstormer** | Enhanced creative brainstormer with explicit focus on cross-field connections, "strategic ignorance" (challenging flawed assumptions the field follows uncritically), and the skeptical-reader test. |
  22. ### Skill
  23. | Skill | What it does |
  24. |-------|-------------|
  25. | `/research-companion` | A structured multi-phase ideation session that orchestrates all three agents through: **Seed** → **Diverge** → **Evaluate** → **Deepen** → **Frame** → **Decide**. Includes Carlini's "conclusion-first test." |
  26. ### Principles
  27. 8 research strategy principles organized into three categories (Problem Selection, Execution Strategy, Strategic Positioning) that guide the agents' evaluations.
  28. ### Codex Support
  29. This repository also includes a native Codex plugin manifest and interface metadata:
  30. - `.codex-plugin/plugin.json`
  31. - `agents/openai.yaml`
  32. - `.agents/plugins/marketplace.json`
  33. The same core prompts, agents, and principles are shared across Claude Code and Codex, with the orchestration instructions written to degrade gracefully if delegated subagents are unavailable.
  34. Codex supports plugin marketplaces and installation through the Plugin Directory. This repo now includes a repo-scoped marketplace file so the plugin can be installed after cloning, without manually creating marketplace JSON. Detailed Codex setup lives in [docs/codex-installation.md](./docs/codex-installation.md).
  35. ## Installation
  36. ### Claude Code
  37. ```bash
  38. claude plugin marketplace add https://github.com/andrehuang/research-companion
  39. claude plugin install research-companion@andrehuang-research-companion
  40. ```
  41. ### Codex
  42. See [docs/codex-installation.md](./docs/codex-installation.md) for the full Codex installation guide, including:
  43. - repo marketplace installation
  44. - personal local marketplace installation
  45. - official Codex plugin docs links
  46. ## Usage
  47. ### Evaluate a research idea
  48. Just describe your idea and ask for evaluation:
  49. ```
  50. I'm thinking about studying how LLM-generated code introduces subtle security
  51. vulnerabilities that pass standard code review. Can you evaluate this idea?
  52. ```
  53. The **Idea Critic** will evaluate across 7 dimensions and give you a verdict with the single most important question to resolve next.
  54. ### Decide whether to continue a project
  55. ```
  56. I've been working on adversarial attacks against multimodal models for 3 months.
  57. I have some results but they're incremental. Two other groups just posted preprints
  58. in the same area. Should I continue?
  59. ```
  60. The **Research Strategist** will assess your competitive position, impact potential, and opportunity cost, then recommend Continue / Pivot / Kill.
  61. ### Run a full brainstorming session
  62. In Claude Code:
  63. ```
  64. /research-companion I'm interested in the intersection of program synthesis
  65. and scientific discovery
  66. ```
  67. In Codex:
  68. ```
  69. $research-companion I'm interested in the intersection of program synthesis
  70. and scientific discovery
  71. ```
  72. This launches a 6-phase guided session:
  73. 1. **Seed** — Understand your problem space, interests, and what bugs you about the field
  74. 2. **Diverge** — Generate ideas, alternative framings, and cross-field connections
  75. 3. **Evaluate** — Stress-test the top 2-3 ideas with the Idea Critic
  76. 4. **Deepen** — Check novelty, positioning, and competitive landscape
  77. 5. **Frame** — Write the abstract and conclusion as if the paper is done (the conclusion-first test: if you can't write a compelling conclusion now, the idea isn't ready)
  78. 6. **Decide** — Final assessment with next steps (starting with the single riskiest assumption to test first)
  79. ### Find cross-field connections
  80. ```
  81. I work on differential privacy. What ideas from other fields
  82. (cryptography, economics, ecology, etc.) could lead to novel approaches?
  83. ```
  84. The **Brainstormer** is designed to bridge distant fields — the kind of connection that led to applying differential cryptanalysis to model stealing, or semi-supervised learning to poisoning attacks.
  85. ### Stress-test with a skeptical reviewer
  86. ```
  87. Here's my draft abstract. Play devil's advocate — what would a skeptical
  88. Area Chair say?
  89. ```
  90. The **Idea Critic** will identify the strongest counter-arguments, the weakest assumptions, and what a hostile but fair reviewer would target.
  91. ## The 7 Evaluation Dimensions
  92. When the Idea Critic evaluates your research idea, it assesses:
  93. | # | Dimension | Key Question |
  94. |---|-----------|-------------|
  95. | 1 | **Novelty** | If you don't do this, how long until someone else does? |
  96. | 2 | **Impact** | Can you write a compelling conclusion *right now*, without doing the work? |
  97. | 3 | **Timing** | Is the field ready for this? Too early? Already crowded? |
  98. | 4 | **Feasibility** | What's the single riskiest assumption? Can you test it in a week? |
  99. | 5 | **Competitive Landscape** | Who else is working on this? What's your unfair advantage? |
  100. | 6 | **The Nugget** | Can you state the key insight in one sentence? |
  101. | 7 | **Narrative Potential** | Can you tell a story that makes a skeptical reader care? |
  102. ## The 8 Research Strategy Principles
  103. These principles, derived from patterns in high-impact research, guide all agents in this plugin:
  104. **Problem Selection**
  105. - **RS1: The Novelty Test** — "If you don't do this, how many months until someone else does?"
  106. - **RS2: The Conclusion-First Test** — "Can you write a compelling conclusion right now?"
  107. - **RS3: The Nugget Test** — "Can you state the key insight in one sentence?"
  108. **Execution Strategy**
  109. - **RS4: Fail Fast** — "Start with the sub-problem most likely to kill the project"
  110. - **RS5: Kill Early** — "A working project with low impact is worse than a killed project"
  111. - **RS6: Unreasonable Effort** — "Strengthen 'sometimes' to 'usually' through additional work"
  112. **Strategic Positioning**
  113. - **RS7: Comparative Advantage** — "Research space is high-dimensional; find your unique corner"
  114. - **RS8: Timing Awareness** — "Impact = skill x domain importance at this moment"
  115. ## Persistent Evaluations (NEW in v1.1)
  116. Evaluation results are now **saved to disk** so they persist across sessions:
  117. - After each session, verdicts are written to `research-evaluations/YYYY-MM-DD-<topic>.md`
  118. - On subsequent sessions, the system **checks for prior evaluations** of similar topics before starting fresh
  119. - PARK'd ideas include "revisit conditions" — what would need to change to reconsider
  120. - The Research Strategist now outputs a **watch list** (search terms, key researchers, venues to monitor) for competitive tracking
  121. - The Idea Critic checks for prior evaluations to avoid re-evaluating killed ideas unless conditions have changed
  122. This means your research thinking accumulates over time rather than starting from scratch each session.
  123. ## Pairs Well With
  124. - [**researcher-pack**](https://github.com/andrehuang/researcher-pack) ⭐ — **The recommended way to use Research Companion.** An integrated Claude Code plugin for researchers that incorporates Research Companion as its strategic-thinking component, alongside paper reading, literature tracking, academic writing, and session management. If you use more than one of the tools below, install `researcher-pack` instead of stitching them together by hand.
  125. - [**Academic Writing Agents**](https://github.com/andrehuang/academic-writing-agents) — 12 agents for reviewing, auditing, drafting, and polishing academic papers. Research Companion handles the "what to write" question; Academic Writing Agents handles the "how to write it" question.
  126. - [**Claude-Claw**](https://github.com/andrehuang/claude-claw) — OpenClaw-inspired enhancements for session continuity, task management, and memory architecture. Provides `/handoff` for saving session state and `/manager` for tracking research tasks.
  127. ## Philosophy
  128. Most researchers optimize for publication count. The best researchers optimize for impact — and papers follow naturally. This plugin is built around that distinction.
  129. It won't help you produce more papers. It will help you produce *better* ones by being honest about which ideas are worth your finite time and by stress-testing your thinking before you commit months of effort.
  130. The agents are deliberately opinionated and direct. They'd rather save you 3 months than spare your feelings. If an idea should be killed, they'll say so — and explain why.
  131. ## License
  132. MIT

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.