I see you reported a similar Cloud Agent issue previously, and I’m sorry you’re running into more trouble. Your Dockerfile and environment.json are correct — this is a server-side infrastructure issue, not anything on your side.
Our team is aware of Docker build failures affecting Cloud Agent setup and is actively working on improvements to the build pipeline. There have been recent infrastructure changes rolling out that should help stabilize this.
Could you try setting up the agent again? If it still fails, please share the Cloud Agent session URL from your browser’s address bar (it looks like cursor.com/agents/bc-``...). That will let me trace the specific failure in our backend logs.