Autonoma

Autonoma runs your end-to-end tests against a live preview deployment of your app on every pull request. When a preview fails to build or deploy, or a test fails because the app itself is broken, this connector gives Claude the same live evidence you'd otherwise dig up in a dashboard: deploy status, build and runtime logs, a rule-based diagnosis, and the full env-var/secret surface for the affected service. Claude can also apply the fix directly - set a missing secret value, edit a service's build or wiring config, or reshape the preview (add/remove an app or a service) - then wait for the rebuild and confirm it came back up, all without leaving the conversation. Every tool is scoped to the repo you name (or that Claude infers from context); access is limited to the organizations and repos your Autonoma account belongs to. Requires an Autonoma account with at least one app connected and a pull request with an active preview deployment. This is an MCP to debug and fix broken Autonoma preview deploys.