async-attention

Async follow-up routing directs pending tasks to the correct surface—deferred-resume, scheduled-task, or owning conversation—ensuring timely continuation without blocking active processes. It evaluates the scenario to decide the most effective mechanism for maintaining momentum.

Scheduling wakeups across automation tasks requires a routing mechanism that coordinates callbacks across conversations and tasks while enforcing safety constraints. This approach promotes explicit admin commands for traceability and discourages long-running sleeps to maintain reliable operation.

Deferred-resume is needed when a conversation thread requires a follow-up or wakeup that cannot be handled immediately without blocking. It allows the system to pause and resume execution later, maintaining momentum across complex async interactions.

Yes, scheduled tasks can be used instead of owning conversations when async callbacks need to run independently. The routing mechanism evaluates dependency detection and frontmatter metadata integrity to determine whether a scheduled-task or owning conversation is the most effective surface.

Long-running sleeps should be avoided in async attention management because they reduce traceability and block resources. The routing mechanism enforces safety constraints by discouraging sleeps and promoting explicit admin commands and scheduled wakeups for reliable, auditable operation.

Async routing operates without external dependencies, relying on internal frontmatter metadata integrity and dependency detection to enforce safety constraints. It applies broadly across conversation threads, automation-owned tasks, and scheduled callbacks to maintain momentum without blocking.