Questions
As of 2026-11-21, 26 questions have a published answer, each computed from the catalogue at build time: a direct answer with counts first, then the table behind it and the rule used to rank it. None are held back for thin data.
Platforms and compatibility
- What are the best MCP servers for Claude Code?
As of 2026-11-21, the most-starred MCP servers whose publishers name Claude Code are context-mode (34,347 GitHub stars), Auto-claude-code-research-in-sleep (19,548) and CodeBurn (11,879).
- What are the best MCP servers for Cursor?
As of 2026-11-21, the most-starred MCP servers whose publishers name Cursor are n8n-mcp (23,292 GitHub stars), CodeBurn (11,879) and Trail-Code (11,389).
- What are the best MCP servers for Codex?
As of 2026-11-21, the most-starred MCP servers whose publishers name Codex are context-mode (34,347 GitHub stars), Auto-claude-code-research-in-sleep (19,548) and CodeBurn (11,879).
- What are the best agent skills for Codex?
As of 2026-11-21, the most-starred agent skills whose publishers name Codex are claude-mem (118,742 GitHub stars), archify (98,455) and Understand-Anything (93,490).
- What are the best agent skills for Claude Code?
As of 2026-11-21, the most-starred agent skills whose publishers name Claude Code are claude-mem (118,742 GitHub stars), archify (98,455) and Understand-Anything (93,490).
- Which coding agents support Agent Skills?
As of 2026-11-21, 17 of the 28 coding agents in AgentSidekick's curated platform records officially document Agent Skills support, including Claude Code, Codex, Cursor and OpenCode (the ones publishers name most often).
- Which AI platforms support MCP?
As of 2026-11-21, 36 of the 47 AI platforms in AgentSidekick's curated platform records officially document MCP support, including Claude Code, Codex, Cursor and ChatGPT (the ones publishers name most often).
- Claude Code vs Cursor vs Codex: how do they compare?
All 3 document both MCP and Agent Skills support.
Standards and concepts
- MCP vs A2A: what is the difference?
They connect different things.
- What is the difference between remote and local MCP servers?
As of 2026-11-21, AgentSidekick tracks 69,078 active MCP servers.
- What is an MCP server?
An MCP server is a program that exposes tools, resources and prompts to AI applications through the Model Context Protocol, an open JSON-RPC protocol maintained by Agentic AI Foundation (Linux Foundation) under MIT AND Apache-2.0.
- What is an agent skill?
An agent skill is a folder containing a SKILL.md file with instructions, plus optional scripts and resources, that an agent loads on demand when a task calls for it.
- What is the difference between an MCP server, an agent skill and a plugin?
An MCP server is a running program that gives an AI application tools and data over the open Model Context Protocol.
Tools for a specific system
- What are the AI agent tools for Salesforce?
As of 2026-11-21, AgentSidekick lists 27 AI agent tools tagged with Salesforce that pass its completeness bar (21 MCP servers, 4 plugins, 1 ide extension and 1 agent skill); 48 carry the tag across the whole catalogue.
- What are the best MCP servers for PostgreSQL?
As of 2026-11-21, AgentSidekick lists 58 MCP servers tagged with PostgreSQL that pass its completeness bar; 257 carry the tag across the whole catalogue.
- What are the best MCP servers for Supabase?
As of 2026-11-21, AgentSidekick lists 22 MCP servers tagged with Supabase that pass its completeness bar; 45 carry the tag across the whole catalogue.
- What are the best MCP servers for GitHub?
As of 2026-11-21, AgentSidekick lists 49 MCP servers tagged with GitHub that pass its completeness bar; 423 carry the tag across the whole catalogue.
- What are the best MCP servers for Slack?
As of 2026-11-21, AgentSidekick lists 23 MCP servers tagged with Slack that pass its completeness bar; 90 carry the tag across the whole catalogue.
Tool categories compared
- What are the best open-source agent memory systems?
As of 2026-11-21, the most-starred agent memory systems in AgentSidekick's curated records are Mem0 (69,298 GitHub stars), Cognee (33,443) and Graphiti (32,973), ranked by GitHub stars of each project's own repository.
- What are the best AI agent frameworks?
As of 2026-11-21, the most-starred agent frameworks in AgentSidekick's curated records are LangChain (149,989 GitHub stars), AutoGen (61,956) and CrewAI (61,138), ranked by GitHub stars of each project's own repository.
- How do agent observability platforms compare?
As of 2026-11-21, the most-starred observability and tracing tools in AgentSidekick's curated records are Langfuse (37,691 GitHub stars), MLflow (28,920) and Opik (23,478), ranked by GitHub stars of each project's own repository.
- What are the best browser automation tools for AI agents?
As of 2026-11-21, the most-starred browser and computer-use tools in AgentSidekick's curated records are Browser Use (121,649 GitHub stars), Chrome DevTools MCP (58,600) and Lightpanda (42,290), ranked by GitHub stars of each project's own repository.
- How do LLM model gateways compare?
As of 2026-11-21, the most-starred model gateways and routers in AgentSidekick's curated records are LiteLLM (63,646 GitHub stars), Kong AI Gateway (44,403) and Portkey AI Gateway (13,424), ranked by GitHub stars of each project's own repository.
- Which vector databases are used for AI agents?
As of 2026-11-21, the most-starred vector databases in AgentSidekick's curated records are Milvus (46,712 GitHub stars), Faiss (41,841) and Qdrant (35,652), ranked by GitHub stars of each project's own repository.
- How do AI agent evaluation and testing tools compare?
As of 2026-11-21, the most-starred evaluation and testing tools in AgentSidekick's curated records are promptfoo (27,513 GitHub stars), OpenAI Evals (19,776) and DeepEval (19,582), ranked by GitHub stars of each project's own repository.
- LangGraph vs CrewAI: what is the difference?
LangGraph (LangChain, MIT): Low-level orchestration framework that models agents as stateful graphs with checkpointing, human-in-the-loop interrupts and durable execution.
How answers are written
Each answer is produced by a query, not typed. Rankings say their rule in one sentence (usually GitHub stars of the product's own repository, which measure popularity, not quality). Compatibility uses the statuses set out in the methodology; none has been tested by AgentSidekick.
Page data last updated .