<h1 align="center">
<a href="https://prompts.chat">
A sophisticated research automation system built on Claude Code that enhances AI capabilities with specialized agents, intelligent search, and comprehensive research workflows.
Sign in to like and favorite skills
A sophisticated research automation system built on Claude Code that enhances AI capabilities with specialized agents, intelligent search, and comprehensive research workflows.
This repository extends Claude Code's native capabilities through custom agents and MCP (Model Context Protocol) integrations, focusing on automated research, data analysis, and knowledge synthesis. The system enables sophisticated multi-hop reasoning, parallel execution, and evidence-based research workflows.
āāāāāāāāāāāāāāāāāāā āāāāāāāāāāāāāāāāāāā āāāāāāāāāāāāāāāāāāā ā Claude Code ā ā MCP Servers ā ā Custom Agents ā ā Base Platform āāāāāŗā (Tavily, Seq) āāāāāŗā & Commands ā āāāāāāāāāāāāāāāāāāā āāāāāāāāāāāāāāāāāāā āāāāāāāāāāāāāāāāāāā ā ā ā ā¼ ā¼ ā¼ āāāāāāāāāāāāāāāāāāā āāāāāāāāāāāāāāāāāāā āāāāāāāāāāāāāāāāāāā ā Configuration ā ā Web Search ā ā Research ā ā & Permissions ā ā & Analysis ā ā Workflows ā āāāāāāāāāāāāāāāāāāā āāāāāāāāāāāāāāāāāāā āāāāāāāāāāāāāāāāāāā ā ā¼ āāāāāāāāāāāāāāāāāāā ā Research ā ā Reports ā ā (claudedocs/) ā āāāāāāāāāāāāāāāāāāā
Clone or setup the repository
# Ensure the repository structure matches the Claude Code requirements
Configure MCP Servers The
.mcp.json file configures the available MCP servers:
{ "mcpServers": { "sequential-thinking": { "type": "stdio", "command": "npx", "args": ["-y", "@modelcontextprotocol/server-sequential-thinking"] }, "tavily": { "type": "stdio", "command": "npx", "args": ["-y", "tavily-mcp@latest"], "env": { "TAVILY_API_KEY": "your_tavily_api_key_here" } } } }
Set Permissions The
.claude/settings.local.json file controls tool permissions:
{ "permissions": { "allow": [ "WebSearch", "Bash(npx:*)", "mcp__tavily__tavily-search", "mcp__sequential-thinking__sequentialthinking" ] } }
/research "your research query here"
# Specify research depth /research "quantum computing developments" --depth deep # Choose planning strategy /research "AI safety frameworks" --strategy unified # Combined options /research "competitive analysis" --depth exhaustive --strategy planning
: Research depth level--depth
quick: Basic search, 1 hop, summary outputstandard: Extended search, 2-3 hops, structured reportdeep: Comprehensive search, 3-4 hops, detailed analysisexhaustive: Maximum depth, 5 hops, complete investigation
: Planning approach--strategy
planning: Direct execution with minimal clarificationintent: Interactive clarification before executionunified: Balanced approach with user feedbackcustom-claude/ āāā .claude/ # Claude Code configuration ā āāā agents/ # Custom agent definitions ā ā āāā deep-research-agent.md # Deep research agent ā āāā commands/ # Custom commands ā ā āāā research.md # Research command implementation ā āāā MODE_DeepResearch.md # Research mindset configuration ā āāā RESEARCH_CONFIG.md # Research system configuration ā āāā MCP_Tavily.md # Tavily MCP server documentation ā āāā MCP_Sequential.md # Sequential thinking MCP documentation ā āāā settings.local.json # Permission settings ā āāā Claude.md # Minimal configuration file āāā claudedocs/ # Research outputs ā āāā research_*_*.md # Generated research reports with timestamps āāā .mcp.json # MCP server configuration āāā package.json # Dependencies (tavily-mcp) āāā README.md # This file
Located in
.claude/agents/deep-research-agent.md, this agent handles:
Located in
.claude/commands/research.md, implements:
The system optimizes for parallel operations:
Research outputs are saved to the
claudedocs/ directory with structured formats including:
This project is designed as a configuration and extension system for Claude Code. Contributions should focus on:
MCP Server Connection Issues
.mcp.jsonPermission Errors
.claude/settings.local.jsonResearch Command Not Found
.claude/commands/This project is configured based on the super-claude which is used MIT License.
Note: This repository represents a sophisticated research automation system built on Claude Code. It's designed for researchers, analysts, and knowledge workers who require comprehensive, evidence-based research capabilities with AI assistance.