Decide whether wakeup-poll should regain a triage-only tier for some issue class #83

Closed
opened 2026-07-18 18:41:46 +00:00 by jared · 1 comment
Owner

Deferred decision from the ADR-0046 readiness-label review (2026-07-18). The label collapse retired the old semi triage-only wakeup tier: wakeup-poll now spawns acting sessions for ready-for-agent and is notify-only for everything else (ready-for-human and unlabeled). Question deferred: should some class of new issue — most plausibly unlabeled ones — get a triage-only session again (agent labels, comments a proposed approach, stops; never implements)?

Deliberately deferred: current focus is building os-sdlc, and a wakeup+triage loop running now would interfere with that. Also no production signal yet — the wakeup pilot has no cron installed and no opted-in project (ADR-0035).

Revisit trigger: after os-sdlc build and the first real wakeup rollout, if unlabeled issues pile up unrouted in ~/.cc-os/wakeup-summary.md.

Options sketched in the 2026-07-18 session: (1) keep notify-only status quo; (2) triage-only sessions keyed to ABSENCE of labels — no new label, so ADR-0046's two-state vocabulary is untouched; (3) a dedicated triage label — reverses ADR-0046's rejected alternative, would need a superseding ADR. ready-for-human stays notify-only in all options.

Code touchpoints if implemented: plugins/os-backlog/lib/backlog/wakeup.rb (tier(), session_prompt), bin/wakeup-poll, tests/wakeup_test.rb, hooks/session_start.py NOTE text, openspec/specs/issue-state-labels/spec.md.

Deferred decision from the ADR-0046 readiness-label review (2026-07-18). The label collapse retired the old semi triage-only wakeup tier: wakeup-poll now spawns acting sessions for ready-for-agent and is notify-only for everything else (ready-for-human and unlabeled). Question deferred: should some class of new issue — most plausibly unlabeled ones — get a triage-only session again (agent labels, comments a proposed approach, stops; never implements)? Deliberately deferred: current focus is building os-sdlc, and a wakeup+triage loop running now would interfere with that. Also no production signal yet — the wakeup pilot has no cron installed and no opted-in project (ADR-0035). Revisit trigger: after os-sdlc build and the first real wakeup rollout, if unlabeled issues pile up unrouted in ~/.cc-os/wakeup-summary.md. Options sketched in the 2026-07-18 session: (1) keep notify-only status quo; (2) triage-only sessions keyed to ABSENCE of labels — no new label, so ADR-0046's two-state vocabulary is untouched; (3) a dedicated triage label — reverses ADR-0046's rejected alternative, would need a superseding ADR. ready-for-human stays notify-only in all options. Code touchpoints if implemented: plugins/os-backlog/lib/backlog/wakeup.rb (tier(), session_prompt), bin/wakeup-poll, tests/wakeup_test.rb, hooks/session_start.py NOTE text, openspec/specs/issue-state-labels/spec.md.
Author
Owner

Discarded in 2026-07-21 bloat cull: speculative deferred decision

Discarded in 2026-07-21 bloat cull: speculative deferred decision
jared closed this issue 2026-07-21 18:21:45 +00:00
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
jared/cc-os#83
No description provided.