Executor Lets You Share Tools and Auth Across All AI Coding Agents

Mar 15, 2026 · Updated Apr 25, 2026

Executor, a local-first execution environment for AI agents, connects external tool sources — MCP servers, OpenAPI APIs, and GraphQL endpoints — once and shares them across every coding agent you use. Auth, credentials, and approval flows live in a single local runtime. Agents access the shared catalog through the executor CLI or MCP bridge, so Claude Code, OpenCode, and other agents draw from the same workspace without separate setup.

The typical friction: switching coding agents means re-pasting MCP configs, re-authenticating, and rebuilding tool access from scratch. Executor centralizes this into one workspace tool catalog any agent can tap into. As developers mix specialized coding agents for different tasks, a shared execution layer cuts the overhead of per-agent configuration.

Connect your MCP servers once through Executor and access them from whichever coding agent fits the task — tool configuration and auth carry over as you switch.

View the full update on github.com

Rhys

@RhysSullivanMar 10

This is why you want an execution layer share your tools, auth and approvals across all agents you use Check out me adding the Axiom MCP server, authing to it, and calling it from Claude Code and OpenCode in under a minute - all locally on my computer `npm i -g executor` https://t.co/JmZlNR7qii

View on X