Claude Scholar Prompt

scholar-summary.md

Google Scholar Alert Analysis Prompt

When I say "ANALYZE SCHOLAR ALERTS", perform these specific steps:

1. SEARCH AND RETRIEVE ALL EMAILS

Use gmail_search tool with: from:[email protected] newer_than:7d IMPORTANT: Gmail returns results in pages. You MUST:

• Check if nextPageToken is present in the response
• If yes, continue searching with the page_token parameter until no nextPageToken is returned
• Count total emails retrieved and report the number
• Example: If you get 5 emails and a nextPageToken, search again with that token to get remaining emails

2. READ EACH ALERT EMAIL

Use read_gmail_thread tool to access full content of EVERY email found

3. EXTRACT ARTICLE INFORMATION

From each alert email, capture:

• Article title (exact, complete)
• Authors (all listed)
• Publication venue and date
• LINKS (HIGHEST PRIORITY): a) Direct article URL - DOI links or publisher URLs b) PDF download links - direct PDF access c) Google Scholar entry URL - always capture as fallback
• Brief description/abstract if available

4. PRIORITY FOCUS AREAS

VIRTUAL EMPLOYEE (VE) RESEARCH (TOP PRIORITY)

Papers on: digital workers/workforce, multi-agent systems/frameworks, autonomous business agents, AI knowledge workers, LLM-powered agents/orchestration, cognitive assistants (enterprise), agent-based automation, task-oriented autonomous systems, digital labor/AI labor solutions

HIGH-REPUTATION AUTHORS/INSTITUTIONS

• Top institutions: MIT, Stanford, CMU, Google DeepMind, Anthropic, OpenAI, Microsoft Research, IBM Research
• Multi-institution collaborations (3+ institutions)
• Note industry connections and prior contributions

ENTERPRISE RELEVANCE

• Enterprise AI adoption/implementation
• Salesforce ecosystem developments
• Mid-tier business implementation metrics

5. ORGANIZE FINDINGS AS:

• TOTAL EMAILS PROCESSED: [Number] Google Scholar alerts analyzed
• VE-ALIGNED RESEARCH (top section)
• HIGH-PRIORITY RESEARCH (3-5 most relevant papers)
• CATEGORY INSIGHTS (grouped by area with pattern analysis)
• INDUSTRY APPLICATIONS (implementation examples)
• POLICY DEVELOPMENTS (governance frameworks)
• ACCESS STRATEGY (valuable paywalled content with rationale)

6. CREATE OUTPUT AS ARTIFACT

• Use artifacts tool with command="create"
• Title: "SCHOLAR ALERT ANALYSIS: [Current Date]"
• Type: "text/markdown"
• Include ALL sections with complete metadata and links

PAGINATION EXAMPLE:

First search returns: 5 emails + nextPageToken: "09705118503109634064"
Second search with page_token: 5 more emails + nextPageToken: "04403907045970183217"
Third search with page_token: 4 more emails + no nextPageToken
Total: 14 emails to process

IMPORTANT: DO NOT search the web. ONLY use Gmail search tools to find and process Google Scholar alert emails in my inbox. Continue pagination until ALL emails from the past 7 days are retrieved.