Freeze Phase 4 preparation order

This commit is contained in:
axiomlogicnexus 2026-05-06 03:19:53 +02:00
parent b30a7ee488
commit 6d37f63c6f

View file

@ -0,0 +1,174 @@
# HyperTwist Phase 4 preparation packet
Created on `2026-05-06`
Status:
- first-party HyperTwist packet
- bounded post-Phase-`3` preparation slice
## Purpose
This packet opens the explicit post-Phase-`3` preparation lane after:
- the Phase `3` end-state freeze packet landed
- the thin generated-mode runtime consumer packet landed
- the final dashboard generated-mode consumer convergence packet landed
The open task is:
- freeze the Phase `4` start order and guardrails so the next implementation lane begins deliberately instead of by drift
It is not:
- a live recognition-provider integration packet
- a broad orchestration rewrite
- a speech or simulation packet
- a cleanup-only planning rewrite
## Scope
Bounded lane:
- define the first implementation order for post-Phase-`3` work
- identify the real first file surfaces for that order
- record what remains preparatory versus what should wait for the first actual Phase `4` implementation packet
Required result:
- the repo has one explicit product-side packet naming the first Phase `4` start order
- the packet distinguishes orchestration-first work from later recognition-provider work
- the packet preserves the roadmap reservations that matter before Phase `4` widens materially
Out of scope:
- implementing the first Phase `4` code packet itself
- provider-specific CV integration
- changing the already-closed Phase `3` generated-mode backend
- broad roadmap rewording outside the current continuation docs
## Current owned product truth
The current first-party product now has:
- closure-complete Phase `3` training/coaching backend work
- durable repository save/load/repair/integrity coverage across the active coaching stack
- bounded imported generated-mode selection, request creation, persistence, query, inspection, and owned execution
- thin generated-mode consumer parity through:
- `UHyperTwistTrainingPanelWidget`
- `AHyperTwistTrainingSessionActor`
- `UHyperTwistCoachDashboardWidget`
The current missing work is no longer Phase `3` closure debt.
The missing work is post-closure productionization:
- production-oriented runtime orchestration
- provider-backed recognition execution beyond the current mock surface
## Phase 4 start-order decision
The first Phase `4` implementation packet should be:
1. `orchestration-first`
The second Phase `4` implementation packet should be:
2. `recognition productionization over the existing vision contract`
## Why orchestration comes first
Current code reality:
- the training/coaching backend is already materially mature
- the main runtime entry points are already spread across the subsystem, runtime library, panel widget, session actor, and coach dashboard widget
- the recognition client surface is still interface-and-mock level, not provider-backed runtime productionization
That means the stronger first move is:
- stabilize the production-facing orchestration entry lane over the already-closed backend
before:
- wiring a real recognition provider into runtime behavior that still lacks its final orchestration posture
## Strongest first implementation targets for the first Phase 4 packet
### Orchestration entry surfaces
- `UnrealHyperTwist/Source/UnrealHyperTwist/Public/HyperTwistTraining/HyperTwistTrainingSubsystem.h`
- `UnrealHyperTwist/Source/UnrealHyperTwist/Private/HyperTwistTraining/HyperTwistTrainingSubsystem.cpp`
- active run start / continue / submit / complete / clear surfaces
- coach-run launch surfaces
- queue-entry launch surfaces
- `UnrealHyperTwist/Source/UnrealHyperTwist/Public/HyperTwistTraining/HyperTwistTrainingRuntimeLibrary.h`
- `UnrealHyperTwist/Source/UnrealHyperTwist/Private/HyperTwistTraining/HyperTwistTrainingRuntimeLibrary.cpp`
- `UnrealHyperTwist/Source/UnrealHyperTwist/Public/HyperTwistTraining/HyperTwistTrainingPanelWidget.h`
- `UnrealHyperTwist/Source/UnrealHyperTwist/Private/HyperTwistTraining/HyperTwistTrainingPanelWidget.cpp`
- `UnrealHyperTwist/Source/UnrealHyperTwist/Public/HyperTwistTraining/HyperTwistTrainingSessionActor.h`
- `UnrealHyperTwist/Source/UnrealHyperTwist/Private/HyperTwistTraining/HyperTwistTrainingSessionActor.cpp`
- `UnrealHyperTwist/Source/UnrealHyperTwist/Public/HyperTwistTraining/HyperTwistCoachDashboardWidget.h`
- `UnrealHyperTwist/Source/UnrealHyperTwist/Private/HyperTwistTraining/HyperTwistCoachDashboardWidget.cpp`
### Recognition-provider follow-on surfaces
- `UnrealHyperTwist/Source/UnrealHyperTwist/Public/HyperTwistRecognition/HyperTwistVisionClient.h`
- `UnrealHyperTwist/Source/UnrealHyperTwist/Private/HyperTwistRecognition/HyperTwistMockVisionClient.cpp`
- `UnrealHyperTwist/Source/UnrealHyperTwist/Public/HyperTwistRecognition/HyperTwistRecognitionTypes.h`
## Strongest first goal for the first Phase 4 implementation packet
The first actual Phase `4` code packet should:
- define one clearer production-facing orchestration lane for starting, continuing, and finalizing runtime training/coaching flows
- avoid reopening already-closed generated-mode backend work
- give later recognition integration one stable place to plug in
It should not yet:
- bring in a provider-specific recognition stack
- widen into speech, simulation, or XR
- perform large structural cleanup for its own sake
## Planning reservations that remain visible
These still matter before Phase `4` widens materially:
1. manifest reconciliation and custody governance for restrictive mirrors and clean-room linkage
2. explicit handling of the retained supplemental mirror queue
Current decision:
- keep those reservations visible
- do not let them block the first bounded orchestration packet unless that packet directly depends on them
## Suggested next packet
The next explicit implementation packet after this preparation slice should be:
- a bounded `Phase 4 orchestration entry packet`
The packet should be framed as:
- production-oriented run orchestration over the stabilized subsystem/runtime/consumer layers
## Suggested file ownership
Primary edit surface:
- `docs/arch/HYPERTWIST_PHASE4_PREPARATION_PACKET_2026-05-06.md`
## Acceptance criteria
- the packet states the first Phase `4` implementation order
- the packet explains why orchestration should precede recognition-provider work
- the packet records the strongest first file surfaces for that order
- the packet preserves the roadmap reservations without turning them into blanket blockers
## Validation checklist
1. read the packet against the current product code reality
2. confirm it does not reopen Phase `3` closure work
3. confirm it names Phase `4` implementation order concretely enough for the next session
That is the packet.