Recovery
Use the first-response sequence
This sequence keeps recovery small and preserves the project context you already have.
- 01
Check whether the runtime is idle
Many stalled states only need the workspace to wake before the next action can continue.
- 02
Retry the exact failing action
Preview, sync, and publish issues are easier to reason about when the retry matches the original step.
- 03
Inspect before resetting
Use the current thread, files, and sync state before assuming the workspace needs to be replaced.
Blockers
Common blockers and what to try next
These are the recurring issues most teams hit first.
Preview is unavailable
Start or restart preview and give the runtime a chance to reconnect.
GitHub sync stopped
Read the current sync state first, then retry the action that matches it.
Publishing is blocked
Check whether the project has a deployable snapshot and whether preview is already correct.
The workspace feels stale
Continue in the same project first and re-ground the thread before creating something new.
