A scheduled agent works for weeks, then a connector, API or node update changes how a tool behaves. The run still finishes and nothing alerts you, but the output is wrong or missing. You find out days later.
You do not control the connector or the platform. Vendors ship updates without notice to individual users, and tools can time out or miss data from synced sources.
Give every scheduled run a footer with counts: emails read per account, events read, rows processed. Alert when a count is zero or far from normal. Re-run a short test case after platform upgrades. Keep a simple fallback output, such as a list of links without drafts, so a broken step does not remove the whole brief. In n8n, add an error workflow and limits on loops and items.
The poster's Cowork email triage worked for weeks until a connector update removed thread IDs from draft creation, so drafts became standalone emails. Two commenters say the same update broke their setups, including a daily morning-draft skill.
The reporter says the newer AI Agent node version looped on a tool call with prompts growing to over 82,000 tokens, and pinning the older node version worked around it.
Gonzalez dropped a combined Apple connector within an hour because Notes timed out and Calendar missed synced Google events, which would have left gaps in his daily brief without an error.