Trusty Squire MCP

Trusty Squire connects coding agents to an account and credential broker for service setup, authenticated API workflows and approved purchases.

A broker beside the coding agent

The project provides a local MCP package and a broker used for service onboarding, sign-in and API access with stored credentials. Its documentation also describes purchase workflows that require user approval. Evaluate those capabilities separately: a connection used for one API does not imply that an agent should be allowed to open accounts or make purchases freely.

Installation starts a local service

The documented connect command signs the user in and writes the selected agent's configuration. On Linux it registers a systemd user service; on macOS it uses a launchd user agent. A working user service manager is required. The broker must be running when the MCP client connects.

Start with limited access

Choose one test service and confirm the broker, account and permitted workflow before exposing broader credentials. Inspect the resulting service state after a setup action. Claims about vault isolation should be assessed against the implementation and deployment you use; this directory profile does not independently certify a security boundary. Review approvals and permissions for each consequential operation.

Project documentation

Repository and setup guide