feat: surface task execution progress across Host Agent and Cloud

Host Agent now persists step-level execution detail locally (via a real
TaskMetadataStore/Timeline wired into TaskRunner) and reports a bounded
in-progress snapshot piggybacked on lease renewal. Cloud persists that
snapshot per active assignment and exposes it through the existing task
list/detail query path; Cloud Console renders it as a live badge. Host
Agent's local console gains authenticated, read-only task list and
detail/timeline pages (same-origin, server-rendered) with inlined
screenshots.

Also fixes a pre-existing gap in the shared Timeline: the actual
per-step LLM prompt is now recorded instead of the task goal, benefiting
both Runtime and Host Agent consoles. When a host uses the cloud planner
transport, each decide call's prompt and resulting tool decision are
durably logged in a new planner_decision_log table (with bounded
retention) and browsable from Cloud Console; direct-transport hosts
explicitly surface a "not reported" state.

Includes Alembic migrations 0008 (progress columns on scheduled_tasks)
and 0009 (planner_decision_log), bounded Host-Agent-local retention,
dual-backend repository parity, and Vitest + pytest coverage. Task 6.5
(manual end-to-end device verification) remains.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

storage/artifact_store.py

@@ -1,6 +1,7 @@
 from __future__ import annotations
 
 import json
+import shutil
 from dataclasses import asdict, is_dataclass
 from datetime import date, datetime
 from pathlib import Path
@@ -53,6 +54,10 @@ class ArtifactStore:
             steps.append(json.loads(path.read_text(encoding="utf-8")))
         return steps
 
+    def delete_task(self, task_id: str) -> None:
+        """Remove all on-disk artifacts (JSON + screenshots) for a task."""
+        shutil.rmtree(self.task_dir(task_id), ignore_errors=True)
+
 
 def _jsonable(value: Any) -> Any:
     if hasattr(value, "to_dict"):
@@ -68,4 +73,3 @@ def _jsonable(value: Any) -> Any:
     if isinstance(value, (datetime, date)):
         return value.isoformat()
     return value
-

storage/task_metadata.py

@@ -73,6 +73,21 @@ class TaskMetadataStore:
             ).fetchall()
         return [dict(row) for row in rows]
 
+    def list_task_ids(self) -> list[str]:
+        with self._connect() as connection:
+            rows = connection.execute(
+                "select id from tasks order by created_at desc"
+            ).fetchall()
+        return [row["id"] for row in rows]
+
+    def delete_task(self, task_id: str) -> None:
+        """Delete a task row. No-op if the task does not exist."""
+        with self._connect() as connection:
+            connection.execute(
+                "delete from tasks where id = ?",
+                (task_id,),
+            )
+
     def _ensure_schema(self) -> None:
         with self._connect() as connection:
             connection.execute(
@@ -94,4 +109,3 @@ class TaskMetadataStore:
         connection = sqlite3.connect(self.db_path)
         connection.row_factory = sqlite3.Row
         return connection
-

storage/timeline.py

@@ -11,6 +11,8 @@ from storage.artifact_store import ArtifactStore
 class TimelineRecord:
     index: int
     scene: dict[str, Any]
+    # The prompt actually sent to the LLM for this step. For non-LLM planners
+    # (or older records persisted before D9), this falls back to the task goal.
     prompt: str
     tool_call: dict[str, Any]
     result: dict[str, Any]
@@ -60,3 +62,6 @@ class Timeline:
     def read(self, task_id: str) -> list[dict[str, Any]]:
         return self.artifact_store.read_steps(task_id)
 
+    def delete_task(self, task_id: str) -> None:
+        """Delete all timeline records and screenshots for a task."""
+        self.artifact_store.delete_task(task_id)