mcp-cocktail
AI & Memory
Local
Provides a framework to benchmark, evaluate, and inject real-time guardrails across competing MCP servers, CLIs, and AI agent tools. It enables recursive self-improvement by mining failures and synthesizing weakest valid guardrails.
How to connect
1
Glama registry
View https://glama.ai/mcp/servers/ct488yobch for deploy options, or install from https://github.com/asavs/mcp-cocktail (see README for MCP config).
2
GitHub
Install from https://github.com/asavs/mcp-cocktail and add the server to your MCP client configuration (see repository README).
Tools
Tool names are not listed in our registry for this server. Connect it in your MCP client to see the live tool list.