(CODUX-220) Reuse execution rows for provider telemetry

[numnx]

🧩 Task: Reuse execution rows for provider telemetry

Automated task PR opened by Code UX for T02.

📋 Summary

Sprint: Sprint 220 · Reduce invocation visibility latency safely · Task: T02

Make CLI task-coding execution invocations server-visible immediately after cancellation registration and before workspace preparation, while preserving truthful provider telemetry, recovery semantics, server-authoritative dashboards, and realtime Stats freshness.

🤖 Model & Provider

Provider	Codex CLI
Model	gpt-5.6-sol

⏱️ Timing

Started	2026-07-13 18:27:19 UTC
Finished	2026-07-13 18:27:51 UTC
Duration	32s

📊 Token Usage

Metric	Value
Input tokens	271,449
Cached input tokens	9,780,224
Output tokens	24,761
Total tokens	10,076,434
Invocations	3
Tool calls	87
Active time	13m 51s

Included usage estimate (subscription/local login): $6.99

✅ 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
Make CLI provider execution reuse the pre-created execution invocation while creating provider usage only when the provider slot is claimed.

## Scope
- `src/services/cli-workflow/pipeline/execute-provider-stage.ts`, especially `executeProviderStage`
- `src/services/provider-execution-service.ts`, especially `ExecutionProviderRunArgs`, `executeProvider`, `runProviderInner`, and terminal updates
- `src/services/provider-concurrency-service.ts`, especially `waitForSlotAndClaim`
- `tests/backend/services/cli-workflow/pipeline/pipeline-stages.test.ts`
- `tests/backend/services/provider-execution-service.test.ts`
- `tests/backend/services/provider-concurrency-service.test.ts`

## Implementation Requirements
1. Pass `PipelineContext`'s execution invocation id to `ProviderExecutionService.executeProvider` and keep the execution row open through Git and PR finalization by using the existing finalization control rather than creating a second row.
2. When an invocation id is supplied, append or refresh the existing transcript, link the single provider usage row after the provider slot is claimed, and guard against claiming or updating provider work after the execution row has been cancelled.
3. Ensure CLI provider usage does not receive a timestamp captured before concurrency waiting; provider `started_at`, duration, and accounting must begin at the actual claim/run phase while preserving explicit timestamp behavior for other callers such as Jules.
4. Add tests for reuse, one-to-one execution/provider linkage, no provider usage during preparation, cancellation before claim, and provider duration excluding preparation and slot-wait time.

## Constraints
- Never create fake provider usage or token telemetry during workspace preparation.
- Preserve existing provider retry, shutdown recovery, redaction, and terminal-status protections.
- Do not change provider concurrency limits or introduce a new client/server protocol.
- Keep documentation changes in the dedicated documentation task.

## Verification
- Run `pnpm exec vitest run tests/backend/services/cli-workflow/pipeline/pipeline-stages.test.ts tests/backend/services/provider-execution-service.test.ts tests/backend/services/provider-concurrency-service.test.ts`.
- Assert that a supplied invocation id prevents fallback execution-row creation and that exactly one linked provider usage row exists after a real provider claim.
- Run `pnpm run lint`.

🌿 Branch Info

Base: feature/CODUX-220-reduce-invocation-visibility-latency-safely
Head: task/feature-codux-22-t02-codex-d591266c-mrjjhe7r

---

🤖 Generated by Code UX