About AgentQuadrant

AgentQuadrant is a curated directory of AI agent extensions, organised into five quadrants: MCP servers, Claude plugins, Custom GPTs, Claude Skills, and autonomous agents. Each quadrant covers a different layer of the stack, from protocol-level integrations and reusable instruction packages up to full autonomous frameworks. The point is to give developers one trustworthy index of what exists and which of it is worth installing.

We curate rather than aggregate. The tooling ecosystem grows faster than anyone can track it, and dumping everything into one list just produces noise. So every entry gets one of three tiers (Editor's Pick, Verified, or Listed), and a tool only moves up once we have tested it or sourced it from a trusted origin. An entry we cannot independently confirm stays at Listed until someone re-checks it. Nothing publishes itself.

Data comes from official registries (the Model Context Protocol registry, the OpenAI GPT Store), public GitHub repositories with marketplace manifests and SKILL.md frontmatter, and a licensed dataset of notable GPTs. We don't scrape marketing copy or ingest unverified community lists. Every entry carries a last_verified date, and stale ones get flagged for re-review on a rolling schedule. Built something that belongs here? Send it through the submission form; everything lands in an editorial queue and gets reviewed before it goes live.