docs(roadmap): add #68 — internal reinjection/resume path opacity

OMX lanes leaking internal control prose like [OMX_TMUX_INJECT]
instead of operator-meaningful state. Adding requirement for
structured recovery/reinject events with clear cause, preserved
state, and target lane info.

Also fixes merge conflict in test_isolation.rs.

Source: gaebal-gajae dogfood analysis 2026-04-12
This commit is contained in:
YeonGyu-Kim
2026-04-12 08:53:00 +09:00
parent 4f84607ad6
commit 06d1b8ac87
4 changed files with 5 additions and 2 deletions

View File

@@ -500,3 +500,5 @@ Model name prefix now wins unconditionally over env-var presence. Regression tes
66. **Completion-aware reminder shutdown missing** — dogfooded 2026-04-12. Ultraclaw batch completed and was reported as done, but 10-minute cron reminder (`roadmap-nudge-10min`) kept firing into channel as if work still pending. Reminder/cron state not coupled to terminal task state. **Fix shape:** (a) cron jobs should check task completion state before firing; (b) or, provide explicit `cron.remove` on task completion; (c) or, reminders should include "work complete" detection and auto-expire. Blocker: none. Source: gaebal-gajae dogfood analysis 2026-04-12.
67. **Scoped review lanes do not emit structured verdicts** — dogfooded 2026-04-12. OMX review lanes now have improved scope (specific ROADMAP items, specific files, explicit APPROVE/REJECT contract), but the stop event only contains the review request — not the actual verdict. Operators must infer approval/rejection/blockage from later git commits or surrounding chatter. **Fix shape:** emit structured review result on stop with: `verdict: approve|reject|blocked`, `target: commit/diff reviewed`, `rationale: short summary`. Blocker: none. Source: gaebal-gajae dogfood analysis 2026-04-12.
68. **Internal reinjection/resume paths leak opaque control prose** — dogfooded 2026-04-12. OMX lanes stopping with `Continue from current mode state. [OMX_TMUX_INJECT]` expose internal implementation details instead of operator-meaningful state. The event tells us *that* tmux reinjection happened, but not *why* (retry after failure? resume after idle? manual recovery?), *what state was preserved*, or *what the lane was trying to do*. **Fix shape:** recovery/reinject events should emit structured cause like: `resume_after_stop`, `retry_after_tool_failure`, `tmux_reinject_after_idle`, `manual_recovery` plus preserved state / target lane info. Never leak bare internal markers like `[OMX_TMUX_INJECT]` as the primary summary. Blocker: none. Source: gaebal-gajae dogfood analysis 2026-04-12.