Research tools fall into three layers: access (web, social platforms, documentation), structured context (code graphs, symbol-level understanding, up-to-date library docs), and synthesis/verification (sourced reports, citation audits). The most robust setup combines these layers rather than replacing them with one overly broad skill. One important distinction: social signals are not evidence; likes, upvotes, and views do not prove a claim is true. All five tools in your list made the top 10; Context7, Serena, K-Dense, codebase-memory-mcp, and GPT Researcher were added from outside the list.
CLI that turns code and documentation into a queryable knowledge graph
A CLI that gives agents access to cross-platform search, content reading, and video transcripts
An MCP and CLI tool that gives agents current, version-specific library documentation
Provides symbol-level code search, editing, and refactoring tools over MCP.
A multi-agent bundle of 177 skills for scientific research and data analysis
A skill that turns the past 30 days of social, technical, and market signals into a sourced research brief
MCP server that maps code, documentation, and infrastructure relationships in a persistent local knowledge graph
A multi-agent research framework that produces cited reports from web and local documents
A human-approved, source-checked skill package spanning academic research to publication
A plugin that transforms a codebase into an interactive knowledge graph with multi-agent analysis