Outcome
Close the remaining gap between the tested AgentOS architecture and a production-confident first Fleet bootstrap.
This tracker is complete when its linked subissues are closed with fresh evidence. Durable rows remain authoritative; live wake mechanisms and integration harnesses must prove recovery rather than only happy-path startup.
Scope
- Review and land the current implementation baseline.
- Prove the complete disposable-cluster lifecycle.
- Validate background-task adapters against live services.
- Add durable PostgreSQL wakeups.
- Decide the supported native Herdr event boundary.
- Make the Bun toolchain reproducible.
No redesign is implied unless live evidence exposes a concrete architectural failure.
Outcome
Close the remaining gap between the tested AgentOS architecture and a production-confident first Fleet bootstrap.
This tracker is complete when its linked subissues are closed with fresh evidence. Durable rows remain authoritative; live wake mechanisms and integration harnesses must prove recovery rather than only happy-path startup.
Scope
No redesign is implied unless live evidence exposes a concrete architectural failure.