(CODUX-69) Project ratings into task contracts

[numnx]

🧩 Task: Project ratings into task contracts

Automated task PR opened by Code UX for T02.

📋 Summary

Sprint: Sprint 180 · Self Reflection Rating · Task: T02

Expose each task agent's self-reflection as a persisted 5-star rating, surface the latest rating on Tasks and Live task cards, and show per-section ratings in an accessible hover/focus overlay.

🤖 Model & Provider

Provider	Codex CLI
Model	gpt-5.5

⏱️ Timing

Started	2026-07-07 12:52:49 UTC
Finished	2026-07-07 12:59:14 UTC
Duration	6m 25s

📊 Token Usage

Metric	Value
Input tokens	221,313
Cached input tokens	1,740,416
Output tokens	12,699
Total tokens	1,974,428
Invocations	1
Tool calls	65
Active time	5m 41s

Included usage estimate (subscription/local login): $2.36

✅ QA Review

⏳ QA review runs after this PR is opened. Check the task card in the dashboard for the outcome once it lands.

📝 Full Task Prompt

## Objective
Attach the latest persisted self-reflection rating to the task contracts consumed by `/tasks` and `/live` without changing existing task status semantics.

## Scope
- Edit `src/contracts/project-management-types.ts`
- Edit `src/contracts/app-types.ts`
- Edit `src/repositories/project-management-repository.ts`
- Edit `src/repositories/project-runtime/runtime-status-projection.ts`
- Edit `src/app/dependency-factory/core-factory.ts` only if additional repository wiring from T01 is needed by projections
- Update `tests/backend/repositories/project-management-repository.test.ts`
- Update `tests/backend/repositories/project-runtime/runtime-status-projection.test.ts`
- Update `tests/backend/server/dashboard-project-api.test.ts` if API serialization needs coverage

## Implementation Requirements
1. Add optional `selfReflectionRating` fields to `TaskRecord` and `Subtask` using the contract introduced in T01.
2. In `ProjectManagementRepository.inflateTasks`, batch-load latest ratings for the listed task ids and attach them to each `TaskRecord`.
3. In `RuntimeStatusProjection.getMappedTasks` and `buildProjectStatus`, batch-load latest ratings and attach them to each live `Subtask`.
4. Ensure latest selection is deterministic by captured timestamp and row order when multiple task runs exist.
5. Add regression tests proving unreviewed/unrated tasks omit the field while rated tasks include overall and section ratings.

## Constraints
- Depends on T01's persisted table, repository, and rating contract.
- Do not add new dashboard endpoints; reuse existing `/api/projects/:projectId/tasks` and `/api/status` or project runtime projections.
- Do not perform per-task N+1 queries; use batch lookup for task ids.
- Preserve existing `latestReview`, `mergeIndicator`, dependency, and task ordering behavior.

## Verification
- Run `pnpm run test:backend -- tests/backend/repositories/project-management-repository.test.ts tests/backend/repositories/project-runtime/runtime-status-projection.test.ts tests/backend/server/dashboard-project-api.test.ts`.
- Run `pnpm run lint`.
- Success means both project task records and live subtasks include the same latest rating shape when ratings exist and remain unchanged when ratings do not exist.

🌿 Branch Info

Base: feature/CODUX-180-self-reflection-rating
Head: task/feature-codux-180-self-reflection-rating-t02-codex-mranfzcq

---

🤖 Generated by Code UX