Dynamic MCPs: Stop Hardcoding Your AI Agent's World

Dynamic MCPs let AI agents discover, configure, and compose tools at runtime instead of relying on a fixed, hardcoded toolset. With the Docker MCP Gateway, agents can search, add, configure, remove, and execute tools on demand—securely, via Docker.

AI agents have historically shipped with a fixed set of tools: whatever you wired in at deployment time was all they could ever use. Dynamic MCPs change that model.

By pairing the Model Context Protocol (MCP) with the Docker MCP Gateway, agents can now discover, enable, configure, and orchestrate tools at runtime—without code changes or redeploys.

The Dynamic MCP Solution

The Docker MCP Gateway exposes a catalog of MCP servers and lets an agent manage them through a small set of meta-tools:

mcp-find – Search the catalog for relevant MCP servers based on a natural-language query.
mcp-add – Enable one or more servers for the current session so their tools become available.
mcp-config-set – Adjust configuration for a server (e.g., API keys, endpoints, modes).
mcp-remove – Detach servers from the current session when they’re no longer needed.
mcp-exec – Invoke tools exposed by any connected MCP server.
code-mode – Dynamically author JavaScript functions that can call multiple tools, add logic, and compose complex workflows.

Instead of hardcoding a static tool list, the agent uses these meta-tools to shape its own environment on demand.