agentic-mobile-control/openspec/changes/task-execution-progress-visibility/proposal.md

Why

Host Agent now wires a local TaskMetadataStore and Timeline into its TaskRunner, but its Cloud-assignment path still creates a Task and immediately calls runner.run(task). TaskRunner only updates an existing metadata row, so every update affects zero rows and the Host Agent task page remains empty. The standalone Runtime REST service/UI is a separate process with unrelated storage, so it cannot be the inspection surface for Host Agent executions. An operator who submits or receives a task therefore has no authoritative web view of the actions that actually ran on that Host.

What Changes

• Make TaskRunner.run() create its metadata row idempotently before its first status update, so every execution path, including workflow-owned tasks, persists history when a metadata store is configured. Have Host Agent register Cloud task/attempt correlation before goal execution so operators can match a submitted Cloud task to its local execution record.
• Expose that step-level detail through Host Agent's local console: extend AgentStatusTracker//api/status (or add a focused endpoint) with current step index, step status, and a short in-progress step log; render it in the dashboard's "Current assignment" section instead of the static started-at-only view.
• Add a progress-reporting path from Host Agent to Cloud Control Plane so the control plane learns step-level state near-real-time rather than only at claim/heartbeat/terminal-result. Extend the existing cloud.internal_api.models Pydantic schema (the single source of truth Host Agent already imports directly) rather than introducing a parallel schema.
• Persist and expose the latest per-assignment progress on the Cloud Control Plane side, and surface it in Cloud Console so an operator watching a remote task sees live step progress, not just "dispatched" / "succeeded" / "failed".
• Make Host Agent's own authenticated, server-rendered :8765/tasks pages the authoritative web entry point for actual execution history. They render the shared Timeline records directly and retain the existing same-origin session/CSRF boundary.
• All three surfaces continue to use polling (matching current behavior); this change does not introduce SSE/WebSocket infrastructure unless design.md finds a compelling reason to.
• Fix the shared Timeline/TaskRunner/AIPlanner recording path so a persisted step's "prompt" is the actual prompt sent to the LLM for that step (not the task's overall goal) and the model's resulting decision is captured too — this pre-existing gap affects Runtime and Host Agent alike and undermines the step-level detail this change otherwise adds.
• Persist a durable, per-step log of full LLM prompt/response content on the Cloud Control Plane, for centralized troubleshooting — reusing the already-existing cloud-planner-proxy decide endpoint as the capture point (no new protocol/endpoint) rather than the coarse lease-renewal piggyback used for live index/status. This durable log is populated only for hosts using the cloud planner transport; hosts on the direct transport still get only the coarse index/status via lease renewal. Cloud Console gains a view to browse a task's full LLM interaction history.
• Keep before/after screenshots, operation detail, raw OCR observations, and normalized UI-tree results in the shared Runtime Timeline; render all of that evidence in the Host Agent task-detail page.
• Retire the standalone Runtime REST service, its unauthenticated console/UI, and Host Agent Runtime-supervision settings. Preserve the shared runtime/, storage/, and non-REST api/ library modules used by the Host Agent and MCP integrations.

Capabilities

New Capabilities

• host-agent-task-progress: Host Agent captures step-level execution progress for every in-process task, correlates Cloud assignments with local records, and exposes it through its local console/API.
• cloud-task-progress-visibility: Cloud Control Plane receives, persists, and exposes near-real-time step-level progress for assignments it has dispatched to a Host Agent, and Cloud Console renders it.
• host-agent-console-task-pages: Host Agent's local server-rendered console gains the complete before/after evidence, OCR, and UI-tree views for Host-Agent-executed tasks and is the only web inspection surface for those executions.
• runtime-task-evidence: Runtime task history retains complete pre/post action evidence, raw OCR observations, and existing UI-tree inspection results for Host Agent rendering.
• runtime-standalone-service: the standalone Runtime REST service and its UI are removed; Runtime remains an execution library rather than a second operational console.

Modified Capabilities

• host-agent-protocol: add a requirement that the Host Agent reports in-progress step-level status updates to the control plane (in addition to the existing heartbeat/claim/renewal/result operations), and that the control plane accepts and stores them per active assignment.
• cloud-planner-proxy: the existing planner-decision endpoint additionally persists each resolved decision's prompt/response into a durable, bounded-retention per-step log, instead of discarding it after the response is returned.

Impact

• apps/device-host-agent/host_agent/app.py, execution.py, assignment.py, status.py, web/app.py — wire a real metadata/timeline store into the in-process TaskRunner, extend status tracking and the local console UI/API.
• runtime/task.py, runtime/ai_planner.py, runtime/tool_calling_client.py, storage/timeline.py, storage/artifact_store.py, core/models.py, perception/scene_builder.py — fix the shared step-recording path so the real per-step prompt and the model's response are captured, retain pre/post action screenshots and raw OCR observations, not just the task goal and parsed tool call.
• packages/cloud-platform/cloud/internal_api/models.py, packages/cloud-platform/cloud/internal_api/api.py (decide_planner_call), repository.py/sql_repository.py, a new Alembic migration — new progress-reporting request/response models and a persistence + query path for latest per-assignment coarse progress, and a new durable per-step planner_decision_log table (with its own retention job) populated from the existing planner-decision endpoint.
• cloud-console/ (Vue3 SPA) — new UI to render live per-assignment progress, and a new view to browse a task's full LLM interaction history.
• api/rest.py, api/console.py, api/console_web.py, api/templates/runtime_console/*, api/static/runtime_console/*, their tests, package data, and Runtime-supervision configuration — removed.
• apps/device-host-agent/host_agent/assignment.py, execution.py, web/app.py, and its task templates — create correlated execution records and render the complete shared Timeline evidence.
• runtime/task.py, storage/task_metadata.py, and storage/timeline.py — make durable task creation an execution invariant while retaining generic storage boundaries.