# Awesome-Auto-Research-Tools

> Comprehensive curation and orchestration of automated scientific research tools: AI scientists, autonomous literature review engines, paper generation frameworks, simulation harnesses, and benchmark suites.

- Skill: `marvin-lijunzan/awesome-auto-research-tools` (Agent Skill)
- Install (CLI): `npx skillmds@latest add marvin-lijunzan/awesome-auto-research-tools`
- Raw SKILL.md: https://api.skillmd.com/api/skills/marvin-lijunzan/awesome-auto-research-tools/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Research & Search
- Author: Marvin-LiJunZan (https://skillmd.com/u/marvin-lijunzan)
- Updated: 2026-09-21
- Page: https://skillmd.com/skills/marvin-lijunzan/awesome-auto-research-tools

---


# Awesome-Auto-Research-Tools

A curated, systematic directory and operational playbook for state-of-the-art **Automated Research Systems (ARS)**, Autonomous AI Scientists, Literature Harvesting Frameworks, and Scientific Reasoning Agents.

---

## 1. Categorized Landscape of Auto-Research Frameworks

| Category | Representative Projects | Core Value Proposition | Integration Point |
|---|---|---|---|
| **Autonomous End-to-End Scientists** | The AI Scientist (Sakana AI), AutoResearchClaw, ARIS (Auto-Research-in-Sleep) | From hypothesis generation, code execution, experiment plotting to complete LaTeX writeup | Full pipeline execution |
| **Literature Discovery & Deep Mining** | Storm (Stanford), OpenAlex API, Semantic Scholar API, PaperQA2, ScholarAI | Multi-hop citation graph traversal, evidence synthesis, grounded citations | Literature Review phase |
| **Citation Verification & Integrity** | Cite-Verify, Crossref DOI Engine, Retraction Watch, arXiv Sanity | Anti-hallucination, DOI confirmation, retracted paper screening | Pre-submission quality gate |
| **Paper Structure & Dissection** | PaperSpine, SciBench, Benchmark-Paper-Template | Reverse-engineering paper blueprints, extracting evidence chains | Writing & Framing phase |
| **Peer Review Simulation** | LLM-Reviewer (Reviewer #2), External Review Protocol | Multi-perspective critique, score calibration, rebuttal testing | Internal peer review audit |

---

## 2. Tool Evaluation & Selection Matrix

When selecting tools for an automated research task, follow this decision tree:

```text
Research Stage
├── Ideation & Hypothesis
│   ├── Unconstrained Discovery ──> The AI Scientist / Ideator
│   └── Focused Gap Analysis   ──> Storm / Deep-Research-skills / PaperSpine
├── Literature & Citations
│   ├── High-throughput Harvest  ──> OpenAlex / arXiv MCP
│   └── Anti-Hallucination Gate ──> Cite Verify / Reference Checker
├── Experimentation & Execution
│   ├── Simulation & Numerical ──> Local Python/Conda Sandbox + Runner
│   └── Benchmarking           ──> Benchmark-Paper-Template
└── Manuscript Synthesis
    ├── High-Impact Framing    ──> Nature Skills
    ├── De-AI / Human Polish   ──> Human-Writing
    └── Review Simulation      ──> Light / Reviewer #2 Protocol
```

---

## 3. Best Practices for Automated Research

1. **Grounded Provenance Over Generative Hallucination**: Never allow an LLM to generate literature citations without API verification.
2. **Deterministic Code Sandboxes**: All numerical data in papers must originate from reproducible Python scripts, not generated directly into Markdown tables.
3. **Multi-Agent Cross-Examination**: Employ adversarial reviewing agents (optimist vs. skeptic) to challenge claims before freezing conclusions.