academic-research-skills

Academic Research Skills for Claude Code

繁體中文版

A comprehensive suite of Claude Code skills for academic research, covering the full pipeline from research to publication.

AI is your copilot, not the pilot. This tool won't write your paper for you. It handles the grunt work — hunting down references, formatting citations, verifying data, checking logical consistency — so you can focus on the parts that actually require your brain: defining the question, choosing the method, interpreting what the data means, and writing the sentence after "I argue that."

Unlike a humanizer, this tool doesn't help you hide the fact that you used AI. It helps you write better. Style Calibration learns your voice from past work. Writing Quality Check catches the patterns that make prose feel machine-generated. The goal is quality, not cheating.

---

Guides & Articles

• Academic Writing Shouldn't Be a Solo Act — A detailed walkthrough of the full pipeline workflow (English)
• 學術寫作不該是一個人的事：一套開源 AI 協作工具如何改變研究者的工作流 — 完整使用指南（繁體中文）

---

Features

• Deep Research — 13-agent research team with Socratic guided mode + systematic review / PRISMA + SCR Loop + intent detection + dialogue health monitoring + optional cross-model DA + argumentation & reasoning cognitive framework
• Academic Paper — 12-agent paper writing with Style Calibration, Writing Quality Check, LaTeX output hardening, visualization, revision coaching, citation conversion, and writing judgment framework
• Academic Paper Reviewer — Multi-perspective peer review with 0-100 quality rubrics (EIC + 3 dynamic reviewers + Devil's Advocate with concession threshold protocol + attack intensity preservation + optional cross-model review) + R&R traceability matrix + read-only constraint + review quality thinking framework
• Academic Pipeline — Full 10-stage pipeline orchestrator with adaptive checkpoints, claim verification, material passport, optional cross-model integrity verification, mid-conversation reinforcement, and self-check questions

Full Pipeline

Research → Write → Integrity Check → Review (5-person) → Socratic Coaching
  → Revise → Re-Review → Re-Revise → Final Integrity Check → Finalize
  → Process Summary (with Collaboration Quality Evaluation)

Key Features:

1. Adaptive checkpoints (FULL / SLIM / MANDATORY) after every stage
2. Pre-review integrity verification — 100% reference, data, and claim validation (Phase A-E)
3. Two-stage review with Devil's Advocate + 0-100 quality rubrics
4. Socratic revision coaching with SCR Loop (State-Challenge-Reflect, user-togglable) between review and revision stages
5. Final integrity verification before publication
6. Output: MD + DOCX + LaTeX (APA 7.0 apa7 class / IEEE / Chicago) → PDF via tectonic
7. Post-pipeline process summary with 6-dimension collaboration quality scoring (1–100)
8. Material passport for mid-entry provenance tracking
9. Cross-skill mode advisor (14 scenarios + user archetypes)
10. Style Calibration — learn the author's writing voice from past papers (optional, intake Step 10)
11. Writing Quality Check — writing quality checklist catching overused AI-typical patterns
12. Cross-model verification (optional) — use GPT-5.4 Pro or Gemini 3.1 Pro as an independent second reviewer for integrity checks, DA challenges, and peer review

---

Cross-Model Verification (Optional)

ARS works with Claude Opus 4.6 alone. For higher confidence, you can optionally enable a second AI model to independently verify integrity checks and challenge the devil's advocate.

Quick Setup

# Step 1: Set your API key (choose one or both)
export OPENAI_API_KEY="sk-your-key-here"        # For GPT-5.4 Pro
export GOOGLE_AI_API_KEY="AIza-your-key-here"    # For Gemini 3.1 Pro

# Step 2: Choose your cross-verification model
export ARS_CROSS_MODEL="gpt-5.4-pro"            # Best reasoning
# or: export ARS_CROSS_MODEL="gemini-3.1-pro-preview"  # Strong at factual verification

# Step 3: Run Claude Code as normal — cross-verification activates automatically
claude

What Changes When Enabled

Feature	Without Cross-Model	With Cross-Model
Integrity verification	Single-model 100% check	+ 30% sample independently verified by 2nd model
Devil's Advocate	Single-model DA	+ Cross-model generates independent critique, novel findings added
Peer Review	5 reviewers (same model)	+ 6th independent reviewer from 2nd model

Cost

Full pipeline adds ~$0.60-1.10 in cross-model API costs (GPT-5.4 Pro pricing). See shared/cross_model_verification.md for detailed breakdown.

No API Key? No Problem

Without ARS_CROSS_MODEL set, everything works exactly as before. The cross-model features are invisible and add zero overhead.

---

Showcase: Real Pipeline Output

See the complete artifacts from a real 10-stage pipeline run — including peer review reports, integrity verification reports, and the final paper:

Browse all pipeline artifacts →

Artifact	Description
Final Paper (EN)	APA 7.0 formatted, LaTeX-compiled
Final Paper (ZH)	Chinese version, APA 7.0
Integrity Report — Pre-Review	Stage 2.5: caught 15 fabricated refs + 3 statistical errors
Integrity Report — Final	Stage 4.5: zero regressions confirmed
Peer Review Round 1	EIC + 3 Reviewers + Devil's Advocate
Re-Review	Verification after revisions
Peer Review Round 2	Follow-up review
Response to Reviewers	Point-by-point author response
Post-Publication Audit Report	Independent full-reference audit: found 21/68 issues missed by 3 rounds of integrity checks

---

Performance Notes

Recommended model: Claude Opus 4.6 with Max plan (or equivalent extended-thinking configuration).

The full academic pipeline (10 stages) consumes a large amount of tokens — a single end-to-end run can exceed 200K input + 100K output tokens depending on paper length and revision rounds. Budget accordingly.

Individual skills (e.g., deep-research alone, or academic-paper-reviewer alone) consume significantly less.

Estimated Token Usage by Mode

Skill / Mode	Input Tokens	Output Tokens	Estimated Cost (Opus 4.6)
deep-research socratic	~30K	~15K	~$0.60
deep-research full	~60K	~30K	~$1.20
deep-research systematic-review	~100K	~50K	~$2.00
academic-paper plan	~40K	~20K	~$0.80
academic-paper full	~80K	~50K	~$1.80
academic-paper-reviewer full	~50K	~30K	~$1.10
academic-paper-reviewer quick	~15K	~8K	~$0.30
Full pipeline (10 stages)	~200K+	~100K+	~$4-6
+ Cross-model verification	+~10K (external)	+~5K (external)	+~$0.60-1.10

Estimates based on a ~15,000-word paper with ~60 references. Actual usage varies with paper length, revision rounds, and dialogue depth. Costs at Anthropic API pricing as of April 2026.

Recommended Settings

For the best experience with these skills, enable the following Claude Code features:

Setting	What it does	How to enable	Docs
Agent Team	Spawns subagents for parallel research, writing, and review — critical for multi-agent pipelines	Set CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1 (research preview)	Experimental feature — no stable docs yet
Ralph Loop	Keeps the session alive during long-running pipeline stages so Claude can work autonomously without timing out	Use /ralph-loop to activate	Community plugin — experimental
Skip Permissions	Bypasses per-tool confirmation prompts, enabling uninterrupted autonomous execution across all pipeline stages	Launch with claude --dangerously-skip-permissions	Permissions · Advanced Usage

⚠️ Skip Permissions: This flag disables all tool-use confirmation dialogs. Use at your own discretion — it is convenient for trusted, long-running pipelines but removes the safety net of manual approval. Only enable this in environments where you are comfortable with Claude executing file reads, writes, and shell commands without asking first.

---

Prerequisites

Install Claude Code

Recommended: Native installer (no Node.js required, auto-updates):

# macOS / Linux
curl -fsSL https://claude.ai/install.sh | bash

# Windows (PowerShell)
irm https://claude.ai/install.ps1 | iex

Requires Node.js 18+.

npm install -g @anthropic-ai/claude-code

Set Up API Key

You need an Anthropic API key. Get one at https://console.anthropic.com/

# Claude Code will prompt for your API key on first run
claude

Or set it as an environment variable:

export ANTHROPIC_API_KEY=sk-ant-xxxxx

LaTeX / PDF Output (Optional)

PDF output requires tectonic and specific fonts. This is optional — MD and DOCX output work without any of this.

# macOS
brew install tectonic

# Linux (Debian/Ubuntu)
curl --proto '=https' --tlsv1.2 -fsSL https://drop-sh.fullyjustified.net | sh

# Windows
# Download from https://tectonic-typesetting.github.io/en-US/install.html

Required fonts (for APA 7.0 CJK output):

• Times New Roman — usually pre-installed on macOS/Windows; on Linux install ttf-mscorefonts-installer
• Source Han Serif TC VF (思源宋體) — download from Google Fonts or Adobe GitHub
• Courier New — usually pre-installed

If you only need MD/DOCX output, skip this entirely. The pipeline will ask before attempting LaTeX compilation.

---

Installation

Method 1: As Project Skills (Recommended)

Clone this repo into your project's .claude/skills/ directory:

# Navigate to your project root
cd /path/to/your/project

# Create skills directory if it doesn't exist
mkdir -p .claude/skills

# Clone the skills
git clone https://github.com/Imbad0202/academic-research-skills.git .claude/skills/academic-research-skills

Then copy the .claude/CLAUDE.md content into your project's .claude/CLAUDE.md (merge with existing if you have one).

Global installation: To make skills available across all your projects, install to ~/.claude/skills/ instead:

mkdir -p ~/.claude/skills
git clone https://github.com/Imbad0202/academic-research-skills.git ~/.claude/skills/academic-research-skills

Method 2: As a Standalone Project

# Clone the repo
git clone https://github.com/Imbad0202/academic-research-skills.git

# Navigate to the project
cd academic-research-skills

# Start Claude Code
claude

1. Go to https://github.com/Imbad0202/academic-research-skills
2. Click the green Code button → Download ZIP
3. Extract the ZIP to your desired location
4. For Method 1: move the extracted folder to .claude/skills/academic-research-skills inside your project
5. For standalone use: open a terminal in the extracted folder and run claude

Method 3: Claude Cowork (Desktop)

Use these skills in Claude Cowork — Claude Desktop's agentic workspace for knowledge work.

Option A: Folder Access (Quickest)

1. Clone this repo to a local folder:

   git clone https://github.com/Imbad0202/academic-research-skills.git ~/academic-research-skills
   

2. Open Claude Desktop → click Cowork tab (top bar)
3. Select the cloned academic-research-skills folder as the working directory
4. Claude will auto-detect the skills from SKILL.md files and load them as needed

Option B: As Project Skills

If you already have a project folder in Cowork:

cd /path/to/your/project
mkdir -p .claude/skills
git clone https://github.com/Imbad0202/academic-research-skills.git .claude/skills/academic-research-skills

Skills will auto-load when relevant to your conversation — e.g., saying "help me write a paper" triggers academic-paper.

Requirements:

• Claude Desktop (latest version) with Cowork enabled
• Paid plan (Pro, Max, Team, or Enterprise)

Method 4: Upload to claude.ai

You can load these skills via claude.ai's Project feature without installing Claude Code.

Steps:

1. Download all 4 SKILL.md files from this repo:

   • deep-research/SKILL.md
   • academic-paper/SKILL.md
   • academic-paper-reviewer/SKILL.md
   • academic-pipeline/SKILL.md

2. Sign in to claude.ai

3. Create a new Project:

   • Click Projects → Create Project in the sidebar
   • Name it "Academic Research" (or any name you prefer)

4. Upload SKILL.md files:

   • Open the Project → click Project Knowledge (right panel)
   • Click Add Content → Upload Files
   • Upload all 4 SKILL.md files

5. (Optional) Upload reference and template files for better results:

   • Files under deep-research/references/ (APA guide, methodology templates, etc.)
   • Files under academic-paper/references/ (citation formats, writing style, etc.)
   • Files under academic-paper/templates/ (paper structure templates)

6. Start chatting: Open a new conversation in the Project and say "Guide my research on X" or "Help me write a paper"

claude.ai Limitations:

• Project Knowledge file size limit: 200KB per file
• version and last_updated in SKILL.md YAML frontmatter must be under `metada

...