|
| 1 | +--- |
| 2 | +date: 2026-04-21 |
| 3 | +topic: mainline-reconciliation-after-m8-50-priority-consumer |
| 4 | +--- |
| 5 | + |
| 6 | +# Mainline Reconciliation After M8.50 Priority Consumer Requirements |
| 7 | + |
| 8 | +## Problem Frame |
| 9 | + |
| 10 | +`M8.50` now ships on `main`. |
| 11 | +The bounded continuity-aware triage prioritization step is closed: |
| 12 | + |
| 13 | +- `topReplayReports` no longer stay replay-rate-only, |
| 14 | +- continuity-boosted risk can now outrank raw higher replay-rate peers, |
| 15 | +- triage payloads now expose explainable `triagePriority` metadata derived from already-normalized fast-path fields, |
| 16 | +- triage history now stays chronological while still exposing the same explainable priority view per point. |
| 17 | + |
| 18 | +That changes the next seam again. |
| 19 | +The active question is no longer whether continuity-aware prioritization should exist. |
| 20 | +The active question is whether any additional operator consumer should reuse that shipped priority surface before broader dashboard work resumes. |
| 21 | + |
| 22 | +## Verified Mainline Reality (2026-04-21) |
| 23 | + |
| 24 | +- `/triage` now orders `topReplayReports` by continuity-aware priority instead of raw replay rate alone. |
| 25 | +- `/triage` and `/triage/history` now expose `triagePriority` with: |
| 26 | + - `continuityBoosted`, |
| 27 | + - `persistentContinuity`, |
| 28 | + - `freshPersistentTransition`, |
| 29 | + - bounded continuity breadth/read-count metadata, |
| 30 | + - `rationale`. |
| 31 | +- No new persistence model, route family, or dashboard surface was introduced. |
| 32 | +- Backend sufficiency remains green and dormant-by-default. |
| 33 | + |
| 34 | +## Critical Findings |
| 35 | + |
| 36 | +1. `M8.50` closed the ordering seam. |
| 37 | + Triage no longer claims continuity-aware severity while still ranking by raw replay rate alone. |
| 38 | + |
| 39 | +2. A broader dashboard jump is still not the sharpest next move. |
| 40 | + The highest-leverage next question is still consumer reuse, not surface-area expansion. |
| 41 | + |
| 42 | +3. The next bounded seam is priority-consumer alignment. |
| 43 | + If another operator consumer needs this signal, it should reuse the shipped `triagePriority` surface instead of recomputing ranking ad hoc. |
| 44 | + |
| 45 | +4. New persistence would be wasteful here. |
| 46 | + The current payload already carries the explainable ordering ingredients. |
| 47 | + |
| 48 | +## Requirements |
| 49 | + |
| 50 | +**Truth Refresh** |
| 51 | +- R1. This document becomes the new current source-of-truth item for post-`M8.50` direction. |
| 52 | +- R2. Current-facing docs must describe `M8.50` as shipped continuity-aware triage prioritization. |
| 53 | + |
| 54 | +**Next-Lane Narrowing** |
| 55 | +- R3. Current-facing docs must define `M8.51` as bounded triage priority-consumer alignment. |
| 56 | +- R4. `M8.51` must stay inside already-shipped fast-path payloads: |
| 57 | + - `triagePriority`, |
| 58 | + - `replayRiskLevel`, |
| 59 | + - `replayRiskSignals`, |
| 60 | + - `replayCandidateRate`, |
| 61 | + - `managedConversationContinuitySummary`. |
| 62 | +- R5. `M8.51` must not require new persistence, full-report joins, or a broader dashboard shell. |
| 63 | +- R6. `M8.51` must prefer reuse of shipped priority payloads over fresh ranking logic. |
| 64 | + |
| 65 | +**Scope Control** |
| 66 | +- R7. Do not reopen backend-escalation or CI-repair narration unless fresh evidence appears. |
| 67 | +- R8. Do not move markdown-reader work ahead of `M8.51` unless it compounds the same operator/study loop more directly. |
| 68 | +- R9. Do not jump to broader operator dashboards before bounded priority-consumer alignment lands or is explicitly rejected. |
| 69 | + |
| 70 | +## Success Criteria |
| 71 | + |
| 72 | +- this document is listed first under `docs/brainstorms/index.md`, |
| 73 | +- EN/ZH current-facing docs describe `M8.50` as shipped, |
| 74 | +- EN/ZH current-facing docs describe `M8.51` as bounded priority-consumer alignment work, |
| 75 | +- docs verification passes, |
| 76 | +- `main` stays clean after verification. |
| 77 | + |
| 78 | +## Scope Boundaries |
| 79 | + |
| 80 | +- This document does not reopen backend sufficiency work. |
| 81 | +- This document does not add a new diagnostics persistence model. |
| 82 | +- This document does not claim that broader dashboard work now ships. |
| 83 | +- This document does not require a new operator dashboard. |
| 84 | + |
| 85 | +## Next Steps |
| 86 | + |
| 87 | +-> `/ce:plan` for `post-M8.50 triage priority-consumer alignment`, keeping `M8.51` bounded to reuse of the shipped explainable priority surface. |
0 commit comments