Improve user-facing error messages for context canceled/deadline exceeded

[ianhodge]

Problem

When a cloud agent task fails due to a context.Canceled error (e.g., infrastructure-level context cancellation after 34 minutes of execution), the user sees the opaque message:

Failed to execute task: context canceled

This is the raw Go error string formatted with fmt.Sprintf("Failed to execute task: %v", err), providing no actionable information.

Solution

Add userFacingTaskError() to translate well-known infrastructure errors into clear, user-facing messages before sending them as TaskFailed messages to the server.

New error messages

• context.Canceled:

  "The task was interrupted due to an infrastructure issue (context canceled). This is typically transient — please try again."

• context.DeadlineExceeded:

  "The task exceeded its maximum allowed execution time and was terminated. Consider breaking the task into smaller steps or increasing the timeout."

• Other errors: Preserve existing "Failed to execute task: <err>" format.

Changes

• errors.go: Added userFacingTaskError() function
• worker.go: Updated executeTask() to use userFacingTaskError(err) instead of fmt.Sprintf("Failed to execute task: %v", err)
• worker_test.go: Added tests for context canceled/deadline exceeded scenarios and unit tests for userFacingTaskError()

Companion PR

Server-side defense-in-depth for these errors is in the warp-server companion PR, which adds context.Canceled/context.DeadlineExceeded cases to FromError() in the platformerrors package.

Conversation: https://staging.warp.dev/conversation/827c8486-edd0-4431-bb34-bda17e042c18
Run: https://oz.staging.warp.dev/runs/019dfe67-8ddc-7a9a-a8f4-cf95f003a682
This PR was generated with Oz.