researchlisted

# Research Skill - Multi-Agent Ralph v2.88 **Smart Research with Zai MCP Integration** - Leverages Zai MCP for comprehensive web search, article fetching, and content analysis. Based on the principle that research should be thorough, well-sourced, and actionable. ## Quick Start ```bash # Via skill invocation /research Latest React 19 patterns and best practices # Via CLI ralph research "TypeScript 5.0 performance optimizations" # Focus on specific sources /research "Next.js App Router" --sources github,docs ``` ## Key Changes in v2.88 - **Zai MCP as primary**: Uses `mcp__web-search-prime__webSearchPrime` instead of Minimax - **Native WebFetch**: Uses Claude Code's native WebFetch for content retrieval - **Multi-source fetching**: Supports GitHub, CSDN, Juejin, Linux.do article extraction - **Model-agnostic**: Works with any model configured in settings.json ## Available Tools ### 1. mcp__web-search-prime__webSearchPrime (Primary Search) **Purpose:** Comprehensive web search with high-quality results **Parameters:** ```yaml search_query: string # The search query (max 70 characters recommended) content_size: string # "medium" (400-600 words) or "high" (2500 words) location: string # "cn" (Chinese) or "us" (non-Chinese) search_recency_filter: string # oneDay, oneWeek, oneMonth, oneYear, noLimit search_domain_filter: string # Optional: limit to specific domain ``` **Optimal Patterns:** ```yaml # Good: Specific, time-bounded mcp__web-search-prim