Context
During review of #144 (reminders), we identified a deferred requirement: reminders should eventually let users choose which agent definition executes the reminder.
That implies:
- selecting an agent profile/definition at reminder creation time
- storing that selection in reminder payload/config
- executing reminders with that agent's system prompt and tool scope
Why this matters
Right now reminders implicitly run with the default runtime behavior. For real-world usage, users will want explicit control over behavior and policy per reminder (effectively system-prompt control via agent definition).
Scope (deferred)
Not for PR #144. Track this for a follow-up once sub-agent implementation matures.
Acceptance criteria (future)
- Reminder create UX includes optional Agent Definition field
- API + persistence include selected agent definition ID
- Reminder execution path uses that agent definition context
- Validation and error path for missing or deleted agent definitions
Related: #144
Context
During review of #144 (reminders), we identified a deferred requirement: reminders should eventually let users choose which agent definition executes the reminder.
That implies:
Why this matters
Right now reminders implicitly run with the default runtime behavior. For real-world usage, users will want explicit control over behavior and policy per reminder (effectively system-prompt control via agent definition).
Scope (deferred)
Not for PR #144. Track this for a follow-up once sub-agent implementation matures.
Acceptance criteria (future)
Related: #144