Provides tools for spinning up isolated E2B sandboxes where AI agents can execute code safely. You get programmatic control over sandbox lifecycle: create environments, run arbitrary code, manage processes, and tear everything down when done. Useful when your agent needs to run untrusted code, test snippets in a clean environment, or execute tasks that require filesystem isolation. Part of MCP Armory's auto-generated collection, so it covers the full E2B API surface. If you're building agents that write and run code autonomously, this gives you the infrastructure layer without rolling your own containerization.
claude mcp add --transport stdio com.mcparmory-e2b -- uvx mcparmory-e2b