Most of us keep rewriting the same prompts: "review this like a senior engineer", "plan this feature", "write tests for this".
So I packaged 243 of them (13 categories: planning, testing, code review, performance, architecture, security, docs…) behind one MCP URL:
https://hamza-agent-hub-mcp.loukilihamza78.workers.dev/mcp
Add it to ChatGPT (custom MCP server), Claude (custom connector) or Cursor (mcp.json), then ask things like "Use the security auditor agent on this route."
- 4 read-only tools, no auth, no install
- MIT, self-host on Cloudflare Workers
- Works as plain Markdown too
30-second quickstart
ChatGPT: Settings → Plugins → Add custom MCP server → paste the URL → Authentication: None.
Claude: Settings → Connectors → Add custom connector → paste the URL. Claude Code:
claude mcp add --transport http agent-hub https://hamza-agent-hub-mcp.loukilihamza78.workers.dev/mcp
Cursor: add this to ~/.cursor/mcp.json:
{
"mcpServers": {
"hamza-agent-hub": { "url": "https://hamza-agent-hub-mcp.loukilihamza78.workers.dev/mcp" }
}
}
The 4 tools
-
list_categories: the 13 categories and their counts -
search_agents: find agents by keyword or category -
get_agent: full role prompt for one agent -
run_agent_prompt: builds a ready-to-use prompt from agent role + your task
To be clear: these are well-written role prompts, not autonomous bots. They make the AI you already use behave like a focused specialist.
Repo + demo: https://github.com/hamzachallenge3-dev/hamza-agent-hub-252
Top comments (0)