Week 3: Paper Survey - Claude Code Study

SITUATION

You need a quick literature review for a new research topic. Use MCP to connect to academic paper APIs, Subagents to analyze papers in parallel, and create a Custom Skill to automate survey writing.

🔒

Challenge Details Locked

Start your timer to reveal the challenge requirements, scoring rubric, and submission form.

Your time will be recorded for the leaderboard ranking.

WHAT TO DO

1

Set Up MCP Server

Install and configure an academic paper MCP server (choose one):

Option A: arXiv MCP (Recommended)

uv tool install arxiv-mcp-server claude mcp add arxiv -- uv tool run arxiv-mcp-server

Option B: Semantic Scholar MCP

npx -y @smithery/cli@latest install @smithery-ai/semantic-scholar --client claude

💡 Verify installation: claude mcp list

2

Create Survey Writing Skill

Create a custom Skill in .claude/skills/survey/SKILL.md to automate survey generation.

3

Collect and Analyze Papers with Subagents

Use Task tool to spawn subagents for parallel paper analysis. Collect 10+ papers.

4

Write Survey Document

Use your Skill to write a structured survey with introduction, critical analysis, and research insights.

WHAT YOU'LL LEARN

⚡

Skills (Core)

Create custom Skills for automated survey writing workflow

.claude/skills/*/SKILL.md

🤖

Subagents (Core)

Use Task tool to spawn parallel agents for paper analysis

Task tool with subagent_type

🔌

MCP Setup

Install and configure external API servers for Claude

claude mcp add

📖

Academic APIs

arXiv or Semantic Scholar for paper search (no hallucination)

arxiv-mcp-server

DELIVERABLES

📄

survey.md (Required)

• Introduction & scope
• Critical analysis (10+ papers)
• Research insights & trends
• Conclusion & future directions

📚

references.bib (Required)

• BibTeX format, 10+ entries
• Real papers only (arXiv ID or DOI)

⚡

.claude/skills/*/SKILL.md (Required)

• Custom survey-writer skill
• YAML frontmatter + instructions

🔌

.claude/settings.local.json (Optional)

• MCP server config (if used)
• Bonus points for proper setup

SCORING RUBRIC

Category	Points
A. Tool Setup Custom Skill (10) + MCP config (5) + Subagent usage (5)	20
B. Writing Structure Introduction, organization, references, conclusion	35
C. Research Insight Summaries, critical analysis, synthesis	35
⏱️ Time Bonus	+10
Total	100

⚠️ Note: Fake/hallucinated papers will result in 0 points for that paper. Use MCP to fetch real papers with arXiv ID or DOI.

📚 Paper Survey (10+ papers)