Spike: Power Automate

Summary

Power Automate is a valid proof-of-concept path for Outlook and Microsoft 365 calendars, but it is not the best foundation for a long-lived multi-provider system that must also include Apple calendars and scripted deployment.

Why it is relevant

The repo you called out, MShekow/outlook-calendar-sync, packages a real Power Automate flow for synchronizing two Outlook calendars.

Its core design is sound:

• scheduled synchronization
• blocker-event mirroring
• event-id tracking
• loop prevention with a subject prefix

That design can be generalized, even if Power Automate itself is not the final platform.

What the published flow does well

• Handles two Outlook calendars with a scheduled sync.
• Uses source event identifiers to find corresponding mirror events.
• Avoids infinite loop behavior by skipping sync-marker events.
• Supports privacy reduction by hiding subject and body details.

Known drawbacks

• Slow execution on larger event sets.
• Action-count limits and throttling.
• Low-code maintenance overhead.
• Weak story for scripted deployment and versioned infrastructure.
• Apple Calendar support is not a clean first-class path.

Best use in Project Marvin

Use it as:

• a reference implementation for the sync algorithm
• a temporary fallback for Outlook-to-Outlook mirroring
• a way to validate privacy rules and blocker behavior with minimal code

Do not use it as:

• the final cross-provider architecture
• the main deployment target if you want scripted, repeatable setup

Design takeaways to carry forward

• Scheduled reconciliation is simpler than raw event-trigger fanout.
• Mirror events need a stable mapping to their source.
• Loop prevention must be explicit.
• Privacy policy should be configurable per destination calendar.

Spike outcome

Verdict: useful reference, not preferred final platform.