Papersflow Skills

papersflow-skills

papersflow-skills packages reusable PapersFlow research workflows on top of the hosted papersflow-mcp server.

It is designed for:

• agent and assistant clients that benefit from guided research workflows instead of raw tool discovery
• Anthropic skill review, with a public GitHub URL that clearly shows the packaged skills
• PapersFlow users who want Claude to automatically choose the right MCP tools for literature search, citation mapping, and DeepScan monitoring

What Is Included

• a Claude Code plugin manifest in .claude-plugin/plugin.json
• a bundled remote MCP configuration in .mcp.json
• four skills:
• research-briefing
• deepscan-monitor
• citation-verifier
• comparative-synthesis

Included MCP Server

This plugin points Claude Code at the production PapersFlow MCP server:

• https://doxa.papersflow.ai/mcp

The included .mcp.json uses the hosted remote MCP endpoint so Claude can authenticate with OAuth when needed.

Skills

research-briefing

Use when the user wants:

• literature search
• citation verification
• related-paper discovery
• citation graph exploration
• a concise reading brief from PapersFlow data

This skill guides Claude toward these papersflow-mcp tools:

• search
• fetch
• verify_citation
• search_literature
• find_related_papers
• get_citation_graph
• get_paper_neighbors
• expand_citation_graph

deepscan-monitor

Use when the user wants:

• an async DeepScan started
• progress updates while it runs
• key findings before completion
• the final report once the run is done
• optional plotting of final results

This skill guides Claude toward these papersflow-mcp tools:

• run_deepscan
• get_deepscan_status
• get_deepscan_live_snapshot
• get_deepscan_report
• summarize_evidence
• run_python_plot

citation-verifier

Use when the user wants:

• a DOI, URL, arXiv ID, or citation string verified
• a normalized, enriched paper record from a raw identifier
• a quick check without running a full literature search

This skill guides Claude toward these papersflow-mcp tools:

• verify_citation
• fetch

comparative-synthesis

Use when the user wants:

• cross-run comparison of multiple DeepScan reports
• a unified summary from several research sessions
• trend analysis or gap identification across completed runs
• side-by-side visualizations from report data

This skill guides Claude toward these papersflow-mcp tools:

• summarize_evidence
• get_deepscan_report
• run_python_plot

Install For Local Testing In Claude Code

claude --plugin-dir ./papersflow-skills

Once Claude Code starts:

• run /help to see the namespaced skills
• authenticate the remote MCP server if prompted
• use /reload-plugins after local edits

Expected Skill Names

With the current plugin manifest, the skills appear in Claude Code as:

• /papersflow-skills:research-briefing
• /papersflow-skills:deepscan-monitor
• /papersflow-skills:citation-verifier
• /papersflow-skills:comparative-synthesis

OAuth And Access

Public PapersFlow tools can be used without account access in some flows, but Claude should authenticate with PapersFlow to unlock the full surface:

• summarize_evidence
• run_deepscan
• get_deepscan_status
• get_deepscan_live_snapshot
• get_deepscan_report
• run_python_plot

Support

• Website: https://papersflow.ai
• MCP server: https://doxa.papersflow.ai/mcp
• Privacy: https://papersflow.ai/privacy
• Terms: https://papersflow.ai/terms
• Support: https://papersflow.ai/contact
• Support email: developer@papersflow.ai

Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: papersflow-ai
• Source: papersflow-ai/papersflow-skills
• License: MIT

Install and usage instructions live in the source repository linked above.