Fix LLM validation and dashboard health hot paths

backend/src/api/routes/dashboards.py

@@ -547,12 +547,12 @@ async def get_dashboards(
         )
 
         try:
-            profile_preference = profile_service.get_my_preference(current_user).preference
+            profile_preference = profile_service.get_dashboard_filter_binding(current_user)
             normalized_username = str(
-                getattr(profile_preference, "superset_username_normalized", None) or ""
+                profile_preference.get("superset_username_normalized") or ""
             ).strip().lower()
             raw_username = str(
-                getattr(profile_preference, "superset_username", None) or ""
+                profile_preference.get("superset_username") or ""
             ).strip().lower()
             bound_username = normalized_username or raw_username or None
 
@@ -560,14 +560,14 @@ async def get_dashboards(
                 page_context == "dashboards_main"
                 and bool(apply_profile_default)
                 and not bool(override_show_all)
-                and bool(getattr(profile_preference, "show_only_my_dashboards", False))
+                and bool(profile_preference.get("show_only_my_dashboards", False))
                 and bool(bound_username)
             )
             can_apply_slug_filter = (
                 page_context == "dashboards_main"
                 and bool(apply_profile_default)
                 and not bool(override_show_all)
-                and bool(getattr(profile_preference, "show_only_slug_dashboards", True))
+                and bool(profile_preference.get("show_only_slug_dashboards", True))
             )
 
             profile_match_logic = None

backend/src/api/routes/health.py

@@ -5,27 +5,58 @@
 # @LAYER: UI/API
 # @RELATION: DEPENDS_ON -> health_service
 
-from fastapi import APIRouter, Depends, Query
+from fastapi import APIRouter, Depends, Query, HTTPException, status
 from typing import List, Optional
 from sqlalchemy.orm import Session
 from ...core.database import get_db
 from ...services.health_service import HealthService
 from ...schemas.health import HealthSummaryResponse
-from ...dependencies import has_permission
+from ...dependencies import has_permission, get_config_manager, get_task_manager
 
 router = APIRouter(prefix="/api/health", tags=["Health"])
 
+# [DEF:get_health_summary:Function]
+# @PURPOSE: Get aggregated health status for all dashboards.
+# @PRE: Caller has read permission for dashboard health view.
+# @POST: Returns HealthSummaryResponse.
+# @RELATION: CALLS -> backend.src.services.health_service.HealthService
 @router.get("/summary", response_model=HealthSummaryResponse)
 async def get_health_summary(
     environment_id: Optional[str] = Query(None),
     db: Session = Depends(get_db),
+    config_manager = Depends(get_config_manager),
     _ = Depends(has_permission("plugin:migration", "READ"))
 ):
     """
     @PURPOSE: Get aggregated health status for all dashboards.
     @POST: Returns HealthSummaryResponse
     """
-    service = HealthService(db)
+    service = HealthService(db, config_manager=config_manager)
     return await service.get_health_summary(environment_id=environment_id)
+# [/DEF:get_health_summary:Function]
 
-# [/DEF:health_router:Module]
+
+# [DEF:delete_health_report:Function]
+# @PURPOSE: Delete one persisted dashboard validation report from health summary.
+# @PRE: Caller has write permission for tasks/report maintenance.
+# @POST: Validation record is removed; linked task/logs are cleaned when available.
+# @RELATION: CALLS -> backend.src.services.health_service.HealthService
+@router.delete("/summary/{record_id}", status_code=status.HTTP_204_NO_CONTENT)
+async def delete_health_report(
+    record_id: str,
+    db: Session = Depends(get_db),
+    config_manager = Depends(get_config_manager),
+    task_manager = Depends(get_task_manager),
+    _ = Depends(has_permission("tasks", "WRITE")),
+):
+    """
+    @PURPOSE: Delete a persisted dashboard validation report from health summary.
+    @POST: Validation record is removed; linked task/logs are deleted when present.
+    """
+    service = HealthService(db, config_manager=config_manager)
+    if not service.delete_validation_report(record_id, task_manager=task_manager):
+        raise HTTPException(status_code=404, detail="Health report not found")
+    return
+# [/DEF:delete_health_report:Function]
+
+# [/DEF:health_router:Module]

backend/src/api/routes/llm.py

@@ -205,8 +205,7 @@ async def test_connection(
     )
     
     try:
-        # Simple test call
-        await client.client.models.list()
+        await client.test_runtime_connection()
         return {"success": True, "message": "Connection successful"}
     except Exception as e:
         return {"success": False, "error": str(e)}
@@ -242,8 +241,7 @@ async def test_provider_config(
     )
     
     try:
-        # Simple test call
-        await client.client.models.list()
+        await client.test_runtime_connection()
         return {"success": True, "message": "Connection successful"}
     except Exception as e:
         return {"success": False, "error": str(e)}

backend/src/core/auth/repository.py

@@ -13,7 +13,7 @@
 # [SECTION: IMPORTS]
 from typing import List, Optional
 
-from sqlalchemy.orm import Session
+from sqlalchemy.orm import Session, selectinload
 
 from ...models.auth import Permission, Role, User
 from ...models.profile import UserDashboardPreference
@@ -53,7 +53,12 @@ class AuthRepository:
                 raise ValueError("username must be a non-empty string")
             
             logger.reason(f"Querying user by username: {username}")
-            user = self.db.query(User).filter(User.username == username).first()
+            user = (
+                self.db.query(User)
+                .options(selectinload(User.roles).selectinload(Role.permissions))
+                .filter(User.username == username)
+                .first()
+            )
             
             if user:
                 logger.reflect(f"User found: {username}")
@@ -199,4 +204,4 @@ class AuthRepository:
 
 
 # [/DEF:AuthRepository:Class]
-# [/DEF:backend.src.core.auth.repository:Module]
+# [/DEF:backend.src.core.auth.repository:Module]

backend/src/core/task_manager/__tests__/test_context.py

@@ -0,0 +1,29 @@
+# [DEF:backend.src.core.task_manager.__tests__.test_context:Module]
+# @TIER: STANDARD
+# @SEMANTICS: tests, task-context, background-tasks, sub-context
+# @PURPOSE: Verify TaskContext preserves optional background task scheduler across sub-context creation.
+
+from unittest.mock import MagicMock
+
+from src.core.task_manager.context import TaskContext
+
+
+# [DEF:test_task_context_preserves_background_tasks_across_sub_context:Function]
+# @PURPOSE: Plugins must be able to access background_tasks from both root and sub-context loggers.
+# @PRE: TaskContext is initialized with a BackgroundTasks-like object.
+# @POST: background_tasks remains available on root and derived sub-contexts.
+def test_task_context_preserves_background_tasks_across_sub_context():
+    background_tasks = MagicMock()
+    context = TaskContext(
+        task_id="task-1",
+        add_log_fn=lambda **_kwargs: None,
+        params={"x": 1},
+        background_tasks=background_tasks,
+    )
+
+    sub_context = context.create_sub_context("llm")
+
+    assert context.background_tasks is background_tasks
+    assert sub_context.background_tasks is background_tasks
+# [/DEF:test_task_context_preserves_background_tasks_across_sub_context:Function]
+# [/DEF:backend.src.core.task_manager.__tests__.test_context:Module]

backend/src/core/task_manager/context.py

@@ -8,7 +8,7 @@
 
 # [SECTION: IMPORTS]
 # [SECTION: IMPORTS]
-from typing import Dict, Any, Callable
+from typing import Dict, Any, Callable, Optional
 from .task_logger import TaskLogger
 from ..logger import belief_scope
 # [/SECTION]
@@ -58,11 +58,13 @@ class TaskContext:
         task_id: str,
         add_log_fn: Callable,
         params: Dict[str, Any],
-        default_source: str = "plugin"
+        default_source: str = "plugin",
+        background_tasks: Optional[Any] = None,
     ):
         with belief_scope("__init__"):
             self._task_id = task_id
             self._params = params
+            self._background_tasks = background_tasks
             self._logger = TaskLogger(
                 task_id=task_id,
                 add_log_fn=add_log_fn,
@@ -102,6 +104,16 @@ class TaskContext:
         with belief_scope("params"):
             return self._params
     # [/DEF:params:Function]
+
+    # [DEF:background_tasks:Function]
+    # @PURPOSE: Expose optional background task scheduler for plugins that dispatch deferred side effects.
+    # @PRE: TaskContext must be initialized.
+    # @POST: Returns BackgroundTasks-like object or None.
+    @property
+    def background_tasks(self) -> Optional[Any]:
+        with belief_scope("background_tasks"):
+            return self._background_tasks
+    # [/DEF:background_tasks:Function]
     
     # [DEF:get_param:Function]
     # @PURPOSE: Get a specific parameter value with optional default.
@@ -128,7 +140,8 @@ class TaskContext:
                 task_id=self._task_id,
                 add_log_fn=self._logger._add_log,
                 params=self._params,
-                default_source=source
+                default_source=source,
+                background_tasks=self._background_tasks,
             )
     # [/DEF:create_sub_context:Function]
 

backend/src/core/task_manager/manager.py

@@ -208,7 +208,8 @@ class TaskManager:
                         task_id=task_id,
                         add_log_fn=self._add_log,
                         params=params,
-                        default_source="plugin"
+                        default_source="plugin",
+                        background_tasks=None,
                     )
                     
                     if asyncio.iscoroutinefunction(plugin.execute):

backend/src/plugins/llm_analysis/__tests__/test_client_headers.py

@@ -0,0 +1,30 @@
+# [DEF:backend.src.plugins.llm_analysis.__tests__.test_client_headers:Module]
+# @TIER: STANDARD
+# @SEMANTICS: tests, llm-client, openrouter, headers
+# @PURPOSE: Verify OpenRouter client initialization includes provider-specific headers.
+
+from src.plugins.llm_analysis.models import LLMProviderType
+from src.plugins.llm_analysis.service import LLMClient
+
+
+# [DEF:test_openrouter_client_includes_referer_and_title_headers:Function]
+# @PURPOSE: OpenRouter requests should carry site/app attribution headers for compatibility.
+# @PRE: Client is initialized for OPENROUTER provider.
+# @POST: Async client headers include Authorization, HTTP-Referer, and X-Title.
+def test_openrouter_client_includes_referer_and_title_headers(monkeypatch):
+    monkeypatch.setenv("OPENROUTER_SITE_URL", "http://localhost:8000")
+    monkeypatch.setenv("OPENROUTER_APP_NAME", "ss-tools-test")
+
+    client = LLMClient(
+        provider_type=LLMProviderType.OPENROUTER,
+        api_key="sk-test-provider-key-123456",
+        base_url="https://openrouter.ai/api/v1",
+        default_model="nvidia/nemotron-nano-12b-v2-vl:free",
+    )
+
+    headers = dict(client.client.default_headers)
+    assert headers["Authorization"] == "Bearer sk-test-provider-key-123456"
+    assert headers["HTTP-Referer"] == "http://localhost:8000"
+    assert headers["X-Title"] == "ss-tools-test"
+# [/DEF:test_openrouter_client_includes_referer_and_title_headers:Function]
+# [/DEF:backend.src.plugins.llm_analysis.__tests__.test_client_headers:Module]

backend/src/plugins/llm_analysis/__tests__/test_screenshot_service.py

@@ -0,0 +1,344 @@
+# [DEF:backend.src.plugins.llm_analysis.__tests__.test_screenshot_service:Module]
+# @TIER: STANDARD
+# @SEMANTICS: tests, screenshot-service, navigation, timeout-regression
+# @PURPOSE: Protect dashboard screenshot navigation from brittle networkidle waits.
+
+import pytest
+
+from src.plugins.llm_analysis.service import ScreenshotService
+
+
+# [DEF:test_iter_login_roots_includes_child_frames:Function]
+# @PURPOSE: Login discovery must search embedded auth frames, not only the main page.
+# @PRE: Page exposes child frames list.
+# @POST: Returned roots include page plus child frames in order.
+def test_iter_login_roots_includes_child_frames():
+    frame_a = object()
+    frame_b = object()
+    fake_page = type("FakePage", (), {"frames": [frame_a, frame_b]})()
+    service = ScreenshotService(env=type("Env", (), {})())
+
+    roots = service._iter_login_roots(fake_page)
+
+    assert roots == [fake_page, frame_a, frame_b]
+# [/DEF:test_iter_login_roots_includes_child_frames:Function]
+
+
+# [DEF:test_response_looks_like_login_page_detects_login_markup:Function]
+# @PURPOSE: Direct login fallback must reject responses that render the login screen again.
+# @PRE: Response body contains stable login-page markers.
+# @POST: Helper returns True so caller treats fallback as failed authentication.
+def test_response_looks_like_login_page_detects_login_markup():
+    service = ScreenshotService(env=type("Env", (), {})())
+
+    result = service._response_looks_like_login_page(
+        """
+        <html>
+          <body>
+            <p>Enter your login and password below</p>
+            <label>Username:</label>
+            <label>Password:</label>
+            <button>Sign in</button>
+          </body>
+        </html>
+        """
+    )
+
+    assert result is True
+# [/DEF:test_response_looks_like_login_page_detects_login_markup:Function]
+
+
+# [DEF:test_find_first_visible_locator_skips_hidden_first_match:Function]
+# @PURPOSE: Locator helper must not reject a selector collection just because its first element is hidden.
+# @PRE: First matched element is hidden and second matched element is visible.
+# @POST: Helper returns the second visible candidate.
+@pytest.mark.anyio
+async def test_find_first_visible_locator_skips_hidden_first_match():
+    class _FakeElement:
+        def __init__(self, visible, label):
+            self._visible = visible
+            self.label = label
+
+        async def is_visible(self):
+            return self._visible
+
+    class _FakeLocator:
+        def __init__(self, elements):
+            self._elements = elements
+
+        async def count(self):
+            return len(self._elements)
+
+        def nth(self, index):
+            return self._elements[index]
+
+    service = ScreenshotService(env=type("Env", (), {})())
+    hidden_then_visible = _FakeLocator([
+        _FakeElement(False, "hidden"),
+        _FakeElement(True, "visible"),
+    ])
+
+    result = await service._find_first_visible_locator([hidden_then_visible])
+
+    assert result.label == "visible"
+# [/DEF:test_find_first_visible_locator_skips_hidden_first_match:Function]
+
+
+# [DEF:test_submit_login_via_form_post_uses_browser_context_request:Function]
+# @PURPOSE: Fallback login must submit hidden fields and credentials through the context request cookie jar.
+# @PRE: Login DOM exposes csrf hidden field and request context returns authenticated HTML.
+# @POST: Helper returns True and request payload contains csrf_token plus credentials plus request options.
+@pytest.mark.anyio
+async def test_submit_login_via_form_post_uses_browser_context_request():
+    class _FakeInput:
+        def __init__(self, name, value):
+            self._name = name
+            self._value = value
+
+        async def get_attribute(self, attr_name):
+            return self._name if attr_name == "name" else None
+
+        async def input_value(self):
+            return self._value
+
+    class _FakeLocator:
+        def __init__(self, items):
+            self._items = items
+
+        async def count(self):
+            return len(self._items)
+
+        def nth(self, index):
+            return self._items[index]
+
+    class _FakeResponse:
+        status = 200
+        url = "https://example.test/welcome/"
+
+        async def text(self):
+            return "<html><body>Welcome</body></html>"
+
+    class _FakeRequest:
+        def __init__(self):
+            self.calls = []
+
+        async def post(self, url, form=None, headers=None, timeout=None, fail_on_status_code=None, max_redirects=None):
+            self.calls.append({
+                "url": url,
+                "form": dict(form or {}),
+                "headers": dict(headers or {}),
+                "timeout": timeout,
+                "fail_on_status_code": fail_on_status_code,
+                "max_redirects": max_redirects,
+            })
+            return _FakeResponse()
+
+    class _FakeContext:
+        def __init__(self):
+            self.request = _FakeRequest()
+
+    class _FakePage:
+        def __init__(self):
+            self.frames = []
+            self.context = _FakeContext()
+
+        def locator(self, selector):
+            if selector == "input[type='hidden'][name]":
+                return _FakeLocator([
+                    _FakeInput("csrf_token", "csrf-123"),
+                    _FakeInput("next", "/superset/welcome/"),
+                ])
+            return _FakeLocator([])
+
+    env = type("Env", (), {"username": "admin", "password": "secret"})()
+    service = ScreenshotService(env=env)
+    page = _FakePage()
+
+    result = await service._submit_login_via_form_post(page, "https://example.test/login/")
+
+    assert result is True
+    assert page.context.request.calls == [{
+        "url": "https://example.test/login/",
+        "form": {
+            "csrf_token": "csrf-123",
+            "next": "/superset/welcome/",
+            "username": "admin",
+            "password": "secret",
+        },
+        "headers": {
+            "Origin": "https://example.test",
+            "Referer": "https://example.test/login/",
+        },
+        "timeout": 10000,
+        "fail_on_status_code": False,
+        "max_redirects": 0,
+    }]
+# [/DEF:test_submit_login_via_form_post_uses_browser_context_request:Function]
+
+
+# [DEF:test_submit_login_via_form_post_accepts_authenticated_redirect:Function]
+# @PURPOSE: Fallback login must treat non-login 302 redirect as success without waiting for redirect target.
+# @PRE: Request response is 302 with Location outside login path.
+# @POST: Helper returns True.
+@pytest.mark.anyio
+async def test_submit_login_via_form_post_accepts_authenticated_redirect():
+    class _FakeInput:
+        def __init__(self, name, value):
+            self._name = name
+            self._value = value
+
+        async def get_attribute(self, attr_name):
+            return self._name if attr_name == "name" else None
+
+        async def input_value(self):
+            return self._value
+
+    class _FakeLocator:
+        def __init__(self, items):
+            self._items = items
+
+        async def count(self):
+            return len(self._items)
+
+        def nth(self, index):
+            return self._items[index]
+
+    class _FakeResponse:
+        status = 302
+        url = "https://example.test/login/"
+        headers = {"location": "/superset/welcome/"}
+
+        async def text(self):
+            return ""
+
+    class _FakeRequest:
+        async def post(self, url, form=None, headers=None, timeout=None, fail_on_status_code=None, max_redirects=None):
+            return _FakeResponse()
+
+    class _FakeContext:
+        def __init__(self):
+            self.request = _FakeRequest()
+
+    class _FakePage:
+        def __init__(self):
+            self.frames = []
+            self.context = _FakeContext()
+
+        def locator(self, selector):
+            if selector == "input[type='hidden'][name]":
+                return _FakeLocator([_FakeInput("csrf_token", "csrf-123")])
+            return _FakeLocator([])
+
+    env = type("Env", (), {"username": "admin", "password": "secret"})()
+    service = ScreenshotService(env=env)
+
+    result = await service._submit_login_via_form_post(_FakePage(), "https://example.test/login/")
+
+    assert result is True
+# [/DEF:test_submit_login_via_form_post_accepts_authenticated_redirect:Function]
+
+
+# [DEF:test_submit_login_via_form_post_rejects_login_markup_response:Function]
+# @PURPOSE: Fallback login must fail when POST response still contains login form content.
+# @PRE: Login DOM exposes csrf hidden field and request response renders login markup.
+# @POST: Helper returns False.
+@pytest.mark.anyio
+async def test_submit_login_via_form_post_rejects_login_markup_response():
+    class _FakeInput:
+        def __init__(self, name, value):
+            self._name = name
+            self._value = value
+
+        async def get_attribute(self, attr_name):
+            return self._name if attr_name == "name" else None
+
+        async def input_value(self):
+            return self._value
+
+    class _FakeLocator:
+        def __init__(self, items):
+            self._items = items
+
+        async def count(self):
+            return len(self._items)
+
+        def nth(self, index):
+            return self._items[index]
+
+    class _FakeResponse:
+        status = 200
+        url = "https://example.test/login/"
+
+        async def text(self):
+            return """
+            <html>
+              <body>
+                Enter your login and password below
+                Username:
+                Password:
+                Sign in
+              </body>
+            </html>
+            """
+
+    class _FakeRequest:
+        async def post(self, url, form=None, headers=None, timeout=None, fail_on_status_code=None, max_redirects=None):
+            return _FakeResponse()
+
+    class _FakeContext:
+        def __init__(self):
+            self.request = _FakeRequest()
+
+    class _FakePage:
+        def __init__(self):
+            self.frames = []
+            self.context = _FakeContext()
+
+        def locator(self, selector):
+            if selector == "input[type='hidden'][name]":
+                return _FakeLocator([_FakeInput("csrf_token", "csrf-123")])
+            return _FakeLocator([])
+
+    env = type("Env", (), {"username": "admin", "password": "secret"})()
+    service = ScreenshotService(env=env)
+
+    result = await service._submit_login_via_form_post(_FakePage(), "https://example.test/login/")
+
+    assert result is False
+# [/DEF:test_submit_login_via_form_post_rejects_login_markup_response:Function]
+
+
+# [DEF:test_goto_resilient_falls_back_from_domcontentloaded_to_load:Function]
+# @PURPOSE: Pages with unstable primary wait must retry with fallback wait strategy.
+# @PRE: First page.goto call raises; second succeeds.
+# @POST: Helper returns second response and attempts both wait modes in order.
+@pytest.mark.anyio
+async def test_goto_resilient_falls_back_from_domcontentloaded_to_load():
+    class _FakePage:
+        def __init__(self):
+            self.calls = []
+
+        async def goto(self, url, wait_until, timeout):
+            self.calls.append((url, wait_until, timeout))
+            if wait_until == "domcontentloaded":
+                raise RuntimeError("primary wait failed")
+            return {"ok": True, "url": url, "wait_until": wait_until}
+
+    page = _FakePage()
+    service = ScreenshotService(env=type("Env", (), {})())
+
+    response = await service._goto_resilient(
+        page,
+        "https://example.test/dashboard",
+        primary_wait_until="domcontentloaded",
+        fallback_wait_until="load",
+        timeout=1234,
+    )
+
+    assert response["ok"] is True
+    assert page.calls == [
+        ("https://example.test/dashboard", "domcontentloaded", 1234),
+        ("https://example.test/dashboard", "load", 1234),
+    ]
+# [/DEF:test_goto_resilient_falls_back_from_domcontentloaded_to_load:Function]
+# [/DEF:backend.src.plugins.llm_analysis.__tests__.test_screenshot_service:Module]

backend/src/plugins/llm_analysis/__tests__/test_service.py

@@ -0,0 +1,67 @@
+# [DEF:backend.src.plugins.llm_analysis.__tests__.test_service:Module]
+# @TIER: STANDARD
+# @SEMANTICS: tests, llm-analysis, fallback, provider-error, unknown-status
+# @PURPOSE: Verify LLM analysis transport/provider failures do not masquerade as dashboard FAIL results.
+
+import pytest
+
+from src.plugins.llm_analysis.models import LLMProviderType
+from src.plugins.llm_analysis.service import LLMClient
+
+
+# [DEF:test_test_runtime_connection_uses_json_completion_transport:Function]
+# @PURPOSE: Provider self-test must exercise the same chat completion transport as runtime analysis.
+# @PRE: get_json_completion is available on initialized client.
+# @POST: Self-test forwards a lightweight user message into get_json_completion and returns its payload.
+@pytest.mark.anyio
+async def test_test_runtime_connection_uses_json_completion_transport(monkeypatch):
+    client = LLMClient(
+        provider_type=LLMProviderType.OPENROUTER,
+        api_key="sk-test-provider-key-123456",
+        base_url="https://openrouter.ai/api/v1",
+        default_model="nvidia/nemotron-nano-12b-v2-vl:free",
+    )
+    recorded = {}
+
+    async def _fake_get_json_completion(messages):
+        recorded["messages"] = messages
+        return {"ok": True}
+
+    monkeypatch.setattr(client, "get_json_completion", _fake_get_json_completion)
+
+    result = await client.test_runtime_connection()
+
+    assert result == {"ok": True}
+    assert recorded["messages"][0]["role"] == "user"
+    assert "Return exactly this JSON object" in recorded["messages"][0]["content"]
+# [/DEF:test_test_runtime_connection_uses_json_completion_transport:Function]
+
+
+# [DEF:test_analyze_dashboard_provider_error_maps_to_unknown:Function]
+# @PURPOSE: Infrastructure/provider failures must produce UNKNOWN analysis status rather than FAIL.
+# @PRE: LLMClient.get_json_completion raises provider/auth exception.
+# @POST: Returned payload uses status=UNKNOWN and issue severity UNKNOWN.
+@pytest.mark.anyio
+async def test_analyze_dashboard_provider_error_maps_to_unknown(monkeypatch, tmp_path):
+    screenshot_path = tmp_path / "shot.jpg"
+    screenshot_path.write_bytes(b"fake-image")
+
+    client = LLMClient(
+        provider_type=LLMProviderType.OPENROUTER,
+        api_key="sk-test-provider-key-123456",
+        base_url="https://openrouter.ai/api/v1",
+        default_model="nvidia/nemotron-nano-12b-v2-vl:free",
+    )
+
+    async def _raise_provider_error(_messages):
+        raise RuntimeError("Error code: 401 - {'error': {'message': 'User not found.', 'code': 401}}")
+
+    monkeypatch.setattr(client, "get_json_completion", _raise_provider_error)
+
+    result = await client.analyze_dashboard(str(screenshot_path), logs=["line-1"])
+
+    assert result["status"] == "UNKNOWN"
+    assert "Failed to get response from LLM" in result["summary"]
+    assert result["issues"][0]["severity"] == "UNKNOWN"
+# [/DEF:test_analyze_dashboard_provider_error_maps_to_unknown:Function]
+# [/DEF:backend.src.plugins.llm_analysis.__tests__.test_service:Module]

backend/src/plugins/llm_analysis/models.py

@@ -35,6 +35,7 @@ class ValidationStatus(str, Enum):
     PASS = "PASS"
     WARN = "WARN"
     FAIL = "FAIL"
+    UNKNOWN = "UNKNOWN"
 # [/DEF:ValidationStatus:Class]
 
 # [DEF:DetectedIssue:Class]

backend/src/plugins/llm_analysis/plugin.py

@@ -307,7 +307,7 @@ class DashboardValidationPlugin(PluginBase):
                     await notification_service.dispatch_report(
                         record=db_record,
                         policy=policy,
-                        background_tasks=context.background_tasks if context else None
+                        background_tasks=getattr(context, "background_tasks", None) if context else None
                     )
                 except Exception as e:
                     log.error(f"Failed to dispatch notifications: {e}")

backend/src/plugins/llm_analysis/service.py

@@ -12,6 +12,8 @@ import asyncio
 import base64
 import json
 import io
+import os
+from urllib.parse import urlsplit
 from typing import List, Dict, Any
 import httpx
 from PIL import Image
@@ -33,6 +35,242 @@ class ScreenshotService:
         self.env = env
     # [/DEF:ScreenshotService.__init__:Function]
 
+    # [DEF:ScreenshotService._find_first_visible_locator:Function]
+    # @PURPOSE: Resolve the first visible locator from multiple Playwright locator strategies.
+    # @PRE: candidates is a non-empty list of locator-like objects.
+    # @POST: Returns a locator ready for interaction or None when nothing matches.
+    async def _find_first_visible_locator(self, candidates) -> Any:
+        for locator in candidates:
+            try:
+                match_count = await locator.count()
+                for index in range(match_count):
+                    candidate = locator.nth(index)
+                    if await candidate.is_visible():
+                        return candidate
+            except Exception:
+                continue
+        return None
+    # [/DEF:ScreenshotService._find_first_visible_locator:Function]
+
+    # [DEF:ScreenshotService._iter_login_roots:Function]
+    # @PURPOSE: Enumerate page and child frames where login controls may be rendered.
+    # @PRE: page is a Playwright page-like object.
+    # @POST: Returns ordered roots starting with main page followed by frames.
+    def _iter_login_roots(self, page) -> List[Any]:
+        roots = [page]
+        page_frames = getattr(page, "frames", [])
+        try:
+            for frame in page_frames:
+                if frame not in roots:
+                    roots.append(frame)
+        except Exception:
+            pass
+        return roots
+    # [/DEF:ScreenshotService._iter_login_roots:Function]
+
+    # [DEF:ScreenshotService._extract_hidden_login_fields:Function]
+    # @PURPOSE: Collect hidden form fields required for direct login POST fallback.
+    # @PRE: Login page is loaded.
+    # @POST: Returns hidden input name/value mapping aggregated from page and child frames.
+    async def _extract_hidden_login_fields(self, page) -> Dict[str, str]:
+        hidden_fields: Dict[str, str] = {}
+        for root in self._iter_login_roots(page):
+            try:
+                locator = root.locator("input[type='hidden'][name]")
+                count = await locator.count()
+                for index in range(count):
+                    candidate = locator.nth(index)
+                    field_name = str(await candidate.get_attribute("name") or "").strip()
+                    if not field_name or field_name in hidden_fields:
+                        continue
+                    hidden_fields[field_name] = str(await candidate.input_value()).strip()
+            except Exception:
+                continue
+        return hidden_fields
+    # [/DEF:ScreenshotService._extract_hidden_login_fields:Function]
+
+    # [DEF:ScreenshotService._extract_csrf_token:Function]
+    # @PURPOSE: Resolve CSRF token value from main page or embedded login frame.
+    # @PRE: Login page is loaded.
+    # @POST: Returns first non-empty csrf token or empty string.
+    async def _extract_csrf_token(self, page) -> str:
+        hidden_fields = await self._extract_hidden_login_fields(page)
+        return str(hidden_fields.get("csrf_token") or "").strip()
+    # [/DEF:ScreenshotService._extract_csrf_token:Function]
+
+    # [DEF:ScreenshotService._response_looks_like_login_page:Function]
+    # @PURPOSE: Detect when fallback login POST returned the login form again instead of an authenticated page.
+    # @PRE: response_text is normalized HTML or text from login POST response.
+    # @POST: Returns True when login-page markers dominate the response body.
+    def _response_looks_like_login_page(self, response_text: str) -> bool:
+        normalized = str(response_text or "").strip().lower()
+        if not normalized:
+            return False
+
+        markers = [
+            "enter your login and password below",
+            "username:",
+            "password:",
+            "sign in",
+            'name="csrf_token"',
+        ]
+        return sum(marker in normalized for marker in markers) >= 3
+    # [/DEF:ScreenshotService._response_looks_like_login_page:Function]
+
+    # [DEF:ScreenshotService._redirect_looks_authenticated:Function]
+    # @PURPOSE: Treat non-login redirects after form POST as successful authentication without waiting for redirect target.
+    # @PRE: redirect_location may be empty or relative.
+    # @POST: Returns True when redirect target does not point back to login flow.
+    def _redirect_looks_authenticated(self, redirect_location: str) -> bool:
+        normalized = str(redirect_location or "").strip().lower()
+        if not normalized:
+            return True
+        return "/login" not in normalized
+    # [/DEF:ScreenshotService._redirect_looks_authenticated:Function]
+
+    # [DEF:ScreenshotService._submit_login_via_form_post:Function]
+    # @PURPOSE: Fallback login path that submits credentials directly with csrf token.
+    # @PRE: login_url is same-origin and csrf token can be read from DOM.
+    # @POST: Browser context receives authenticated cookies when login succeeds.
+    async def _submit_login_via_form_post(self, page, login_url: str) -> bool:
+        hidden_fields = await self._extract_hidden_login_fields(page)
+        csrf_token = str(hidden_fields.get("csrf_token") or "").strip()
+        if not csrf_token:
+            logger.warning("[DEBUG] Direct form login fallback skipped: csrf_token not found")
+            return False
+
+        try:
+            request_context = page.context.request
+        except Exception as context_error:
+            logger.warning(f"[DEBUG] Direct form login fallback skipped: request context unavailable: {context_error}")
+            return False
+
+        parsed_url = urlsplit(login_url)
+        origin = f"{parsed_url.scheme}://{parsed_url.netloc}" if parsed_url.scheme and parsed_url.netloc else login_url
+        payload = dict(hidden_fields)
+        payload["username"] = self.env.username
+        payload["password"] = self.env.password
+        logger.info(
+            f"[DEBUG] Attempting direct form login fallback via browser context request with hidden fields: {sorted(hidden_fields.keys())}"
+        )
+
+        response = await request_context.post(
+            login_url,
+            form=payload,
+            headers={
+                "Origin": origin,
+                "Referer": login_url,
+            },
+            timeout=10000,
+            fail_on_status_code=False,
+            max_redirects=0,
+        )
+        response_url = str(getattr(response, "url", "") or "")
+        response_status = int(getattr(response, "status", 0) or 0)
+        response_headers = dict(getattr(response, "headers", {}) or {})
+        redirect_location = str(
+            response_headers.get("location")
+            or response_headers.get("Location")
+            or ""
+        ).strip()
+        redirect_statuses = {301, 302, 303, 307, 308}
+        if response_status in redirect_statuses:
+            redirect_authenticated = self._redirect_looks_authenticated(redirect_location)
+            logger.info(
+                f"[DEBUG] Direct form login fallback redirect response: status={response_status} url={response_url} location={redirect_location!r} authenticated={redirect_authenticated}"
+            )
+            return redirect_authenticated
+
+        response_text = await response.text()
+        text_snippet = " ".join(response_text.split())[:200]
+        looks_like_login_page = self._response_looks_like_login_page(response_text)
+        logger.info(
+            f"[DEBUG] Direct form login fallback response: status={response_status} url={response_url} login_markup={looks_like_login_page} snippet={text_snippet!r}"
+        )
+        return not looks_like_login_page
+    # [/DEF:ScreenshotService._submit_login_via_form_post:Function]
+
+    # [DEF:ScreenshotService._find_login_field_locator:Function]
+    # @PURPOSE: Resolve login form input using semantic label text plus generic visible-input fallbacks.
+    # @PRE: field_name is `username` or `password`.
+    # @POST: Returns a locator for the corresponding input or None.
+    async def _find_login_field_locator(self, page, field_name: str) -> Any:
+        normalized = str(field_name or "").strip().lower()
+        for root in self._iter_login_roots(page):
+            if normalized == "username":
+                input_candidates = [
+                    root.get_by_label("Username", exact=False),
+                    root.get_by_label("Login", exact=False),
+                    root.locator("label:text-matches('Username|Login', 'i')").locator("xpath=following::input[1]"),
+                    root.locator("text=/Username|Login/i").locator("xpath=following::input[1]"),
+                    root.locator("input[name='username']"),
+                    root.locator("input#username"),
+                    root.locator("input[placeholder*='Username']"),
+                    root.locator("input[type='text']"),
+                    root.locator("input:not([type='password'])"),
+                ]
+                locator = await self._find_first_visible_locator(input_candidates)
+                if locator:
+                    return locator
+
+            if normalized == "password":
+                input_candidates = [
+                    root.get_by_label("Password", exact=False),
+                    root.locator("label:text-matches('Password', 'i')").locator("xpath=following::input[1]"),
+                    root.locator("text=/Password/i").locator("xpath=following::input[1]"),
+                    root.locator("input[name='password']"),
+                    root.locator("input#password"),
+                    root.locator("input[placeholder*='Password']"),
+                    root.locator("input[type='password']"),
+                ]
+                locator = await self._find_first_visible_locator(input_candidates)
+                if locator:
+                    return locator
+
+        return None
+    # [/DEF:ScreenshotService._find_login_field_locator:Function]
+
+    # [DEF:ScreenshotService._find_submit_locator:Function]
+    # @PURPOSE: Resolve login submit button from main page or embedded auth frame.
+    # @PRE: page is ready for login interaction.
+    # @POST: Returns visible submit locator or None.
+    async def _find_submit_locator(self, page) -> Any:
+        selectors = [
+            lambda root: root.get_by_role("button", name="Sign in", exact=False),
+            lambda root: root.get_by_role("button", name="Login", exact=False),
+            lambda root: root.locator("button[type='submit']"),
+            lambda root: root.locator("button#submit"),
+            lambda root: root.locator(".btn-primary"),
+            lambda root: root.locator("input[type='submit']"),
+        ]
+        for root in self._iter_login_roots(page):
+            locator = await self._find_first_visible_locator([factory(root) for factory in selectors])
+            if locator:
+                return locator
+        return None
+    # [/DEF:ScreenshotService._find_submit_locator:Function]
+
+    # [DEF:ScreenshotService._goto_resilient:Function]
+    # @PURPOSE: Navigate without relying on networkidle for pages with long-polling or persistent requests.
+    # @PRE: page is a valid Playwright page and url is non-empty.
+    # @POST: Returns last navigation response or raises when both primary and fallback waits fail.
+    async def _goto_resilient(
+        self,
+        page,
+        url: str,
+        primary_wait_until: str = "domcontentloaded",
+        fallback_wait_until: str = "load",
+        timeout: int = 60000,
+    ):
+        try:
+            return await page.goto(url, wait_until=primary_wait_until, timeout=timeout)
+        except Exception as primary_error:
+            logger.warning(
+                f"[ScreenshotService._goto_resilient] Primary navigation wait '{primary_wait_until}' failed for {url}: {primary_error}"
+            )
+            return await page.goto(url, wait_until=fallback_wait_until, timeout=timeout)
+    # [/DEF:ScreenshotService._goto_resilient:Function]
+
     # [DEF:ScreenshotService.capture_dashboard:Function]
     # @PURPOSE: Captures a full-page screenshot of a dashboard using Playwright and CDP.
     # @PRE: dashboard_id is a valid string, output_path is a writable path.
@@ -85,7 +323,13 @@ class ScreenshotService:
                 login_url = f"{base_ui_url.rstrip('/')}/login/"
                 logger.info(f"[DEBUG] Navigating to login page: {login_url}")
                 
-                response = await page.goto(login_url, wait_until="networkidle", timeout=60000)
+                response = await self._goto_resilient(
+                    page,
+                    login_url,
+                    primary_wait_until="domcontentloaded",
+                    fallback_wait_until="load",
+                    timeout=60000,
+                )
                 if response:
                     logger.info(f"[DEBUG] Login page response status: {response.status}")
                 
@@ -101,57 +345,59 @@ class ScreenshotService:
                 logger.info("[DEBUG] Attempting to find login form elements...")
                 
                 try:
+                    used_direct_form_login = False
                     # Find and fill username
-                    u_selector = None
-                    for s in selectors["username"]:
-                        count = await page.locator(s).count()
-                        logger.info(f"[DEBUG] Selector '{s}': {count} elements found")
-                        if count > 0:
-                            u_selector = s
-                            break
+                    username_locator = await self._find_login_field_locator(page, "username")
+
+                    if not username_locator:
+                        roots = self._iter_login_roots(page)
+                        logger.info(f"[DEBUG] Found {len(roots)} login roots including child frames")
+                        for root_index, root in enumerate(roots[:5]):
+                            all_inputs = await root.locator('input').all()
+                            logger.info(f"[DEBUG] Root {root_index}: found {len(all_inputs)} input fields")
+                            for i, inp in enumerate(all_inputs[:5]):  # Log first 5
+                                inp_type = await inp.get_attribute('type')
+                                inp_name = await inp.get_attribute('name')
+                                inp_id = await inp.get_attribute('id')
+                                logger.info(f"[DEBUG] Root {root_index} input {i}: type={inp_type}, name={inp_name}, id={inp_id}")
+                        used_direct_form_login = await self._submit_login_via_form_post(page, login_url)
+                        if not used_direct_form_login:
+                            raise RuntimeError("Could not find username input field on login page")
+                        username_locator = None
                     
-                    if not u_selector:
-                        # Log all input fields on the page for debugging
-                        all_inputs = await page.locator('input').all()
-                        logger.info(f"[DEBUG] Found {len(all_inputs)} input fields on page")
-                        for i, inp in enumerate(all_inputs[:5]):  # Log first 5
-                            inp_type = await inp.get_attribute('type')
-                            inp_name = await inp.get_attribute('name')
-                            inp_id = await inp.get_attribute('id')
-                            logger.info(f"[DEBUG] Input {i}: type={inp_type}, name={inp_name}, id={inp_id}")
-                        raise RuntimeError("Could not find username input field on login page")
-                    
-                    logger.info(f"[DEBUG] Filling username field with selector: {u_selector}")
-                    await page.fill(u_selector, self.env.username)
+                    if username_locator is not None:
+                        logger.info("[DEBUG] Filling username field")
+                        await username_locator.fill(self.env.username)
                     
                     # Find and fill password
-                    p_selector = None
-                    for s in selectors["password"]:
-                        if await page.locator(s).count() > 0:
-                            p_selector = s
-                            break
-                    
-                    if not p_selector:
+                    password_locator = await self._find_login_field_locator(page, "password") if username_locator is not None else None
+
+                    if username_locator is not None and not password_locator:
                         raise RuntimeError("Could not find password input field on login page")
                     
-                    logger.info(f"[DEBUG] Filling password field with selector: {p_selector}")
-                    await page.fill(p_selector, self.env.password)
+                    if password_locator is not None:
+                        logger.info("[DEBUG] Filling password field")
+                        await password_locator.fill(self.env.password)
                     
                     # Click submit
-                    s_selector = selectors["submit"][0]
-                    for s in selectors["submit"]:
-                        if await page.locator(s).count() > 0:
-                            s_selector = s
-                            break
-                    
-                    logger.info(f"[DEBUG] Clicking submit button with selector: {s_selector}")
-                    await page.click(s_selector)
+                    submit_locator = await self._find_submit_locator(page) if username_locator is not None else None
+
+                    if username_locator is not None and not submit_locator:
+                        raise RuntimeError("Could not find submit button on login page")
+
+                    if submit_locator is not None:
+                        logger.info("[DEBUG] Clicking submit button")
+                        await submit_locator.click()
                     
                     # Wait for navigation after login
-                    await page.wait_for_load_state("networkidle", timeout=30000)
+                    if not used_direct_form_login:
+                        try:
+                            await page.wait_for_load_state("load", timeout=30000)
+                        except Exception as load_wait_error:
+                            logger.warning(f"[DEBUG] Login post-submit load wait timed out: {load_wait_error}")
                     
                     # Check if login was successful
-                    if "/login" in page.url:
+                    if not used_direct_form_login and "/login" in page.url:
                         # Check for error messages on page
                         error_msg = await page.locator(".alert-danger, .error-message").text_content() if await page.locator(".alert-danger, .error-message").count() > 0 else "Unknown error"
                         logger.error(f"[DEBUG] Login failed. Still on login page. Error: {error_msg}")
@@ -183,11 +429,24 @@ class ScreenshotService:
 
                 logger.info(f"[DEBUG] Navigating to dashboard: {dashboard_url}")
                 
-                # Use networkidle to ensure all initial assets are loaded
-                response = await page.goto(dashboard_url, wait_until="networkidle", timeout=60000)
+                # Dashboard pages can keep polling/network activity open indefinitely.
+                response = await self._goto_resilient(
+                    page,
+                    dashboard_url,
+                    primary_wait_until="domcontentloaded",
+                    fallback_wait_until="load",
+                    timeout=60000,
+                )
                 
                 if response:
                     logger.info(f"[DEBUG] Dashboard navigation response status: {response.status}, URL: {response.url}")
+
+                if "/login" in page.url:
+                    debug_path = output_path.replace(".png", "_debug_failed_dashboard_auth.png")
+                    await page.screenshot(path=debug_path)
+                    raise RuntimeError(
+                        f"Dashboard navigation redirected to login page after authentication. Debug screenshot saved to {debug_path}"
+                    )
                 
                 try:
                     # Wait for the dashboard grid to be present
@@ -440,6 +699,13 @@ class LLMClient:
 
         # Some OpenAI-compatible gateways are strict about auth header naming.
         default_headers = {"Authorization": f"Bearer {self.api_key}"}
+        if self.provider_type == LLMProviderType.OPENROUTER:
+            default_headers["HTTP-Referer"] = (
+                os.getenv("OPENROUTER_SITE_URL", "").strip()
+                or os.getenv("APP_BASE_URL", "").strip()
+                or "http://localhost:8000"
+            )
+            default_headers["X-Title"] = os.getenv("OPENROUTER_APP_NAME", "").strip() or "ss-tools"
         if self.provider_type == LLMProviderType.KILO:
             default_headers["Authentication"] = f"Bearer {self.api_key}"
             default_headers["X-API-Key"] = self.api_key
@@ -595,6 +861,22 @@ class LLMClient:
                     raise
     # [/DEF:LLMClient.get_json_completion:Function]
 
+    # [DEF:LLMClient.test_runtime_connection:Function]
+    # @PURPOSE: Validate provider credentials using the same chat completions transport as runtime analysis.
+    # @PRE: Client is initialized with provider credentials and default_model.
+    # @POST: Returns lightweight JSON payload when runtime auth/model path is valid.
+    # @SIDE_EFFECT: Calls external LLM API.
+    async def test_runtime_connection(self) -> Dict[str, Any]:
+        with belief_scope("test_runtime_connection"):
+            messages = [
+                {
+                    "role": "user",
+                    "content": 'Return exactly this JSON object and nothing else: {"ok": true}',
+                }
+            ]
+            return await self.get_json_completion(messages)
+    # [/DEF:LLMClient.test_runtime_connection:Function]
+
     # [DEF:LLMClient.analyze_dashboard:Function]
     # @PURPOSE: Sends dashboard data (screenshot + logs) to LLM for health analysis.
     # @PRE: screenshot_path exists, logs is a list of strings.
@@ -661,9 +943,9 @@ class LLMClient:
             except Exception as e:
                 logger.error(f"[analyze_dashboard] Failed to get analysis: {str(e)}")
                 return {
-                    "status": "FAIL",
+                    "status": "UNKNOWN",
                     "summary": f"Failed to get response from LLM: {str(e)}",
-                    "issues": [{"severity": "FAIL", "message": "LLM provider returned empty or invalid response"}]
+                    "issues": [{"severity": "UNKNOWN", "message": "LLM provider returned empty or invalid response"}]
                 }
     # [/DEF:LLMClient.analyze_dashboard:Function]
 # [/DEF:LLMClient:Class]

backend/src/schemas/health.py

@@ -11,7 +11,9 @@ from datetime import datetime
 # [DEF:DashboardHealthItem:Class]
 # @PURPOSE: Represents the latest health status of a single dashboard.
 class DashboardHealthItem(BaseModel):
+    record_id: str
     dashboard_id: str
+    dashboard_slug: Optional[str] = None
     dashboard_title: Optional[str] = None
     environment_id: str
     status: str = Field(..., pattern="^(PASS|WARN|FAIL|UNKNOWN)$")
@@ -30,4 +32,4 @@ class HealthSummaryResponse(BaseModel):
     unknown_count: int
 # [/DEF:HealthSummaryResponse:Class]
 
-# [/DEF:backend.src.schemas.health:Module]
+# [/DEF:backend.src.schemas.health:Module]

backend/src/services/__tests__/test_health_service.py

@@ -1,6 +1,6 @@
 import pytest
 from datetime import datetime, timedelta
-from unittest.mock import MagicMock
+from unittest.mock import MagicMock, patch
 from src.services.health_service import HealthService
 from src.models.llm import ValidationRecord
 
@@ -20,6 +20,7 @@ async def test_get_health_summary_aggregation():
     
     # Dashboard 1: Old FAIL, New PASS
     rec1_old = ValidationRecord(
+        id="rec-old",
         dashboard_id="dash_1",
         environment_id="env_1",
         status="FAIL",
@@ -28,6 +29,7 @@ async def test_get_health_summary_aggregation():
         issues=[]
     )
     rec1_new = ValidationRecord(
+        id="rec-new",
         dashboard_id="dash_1",
         environment_id="env_1",
         status="PASS",
@@ -38,6 +40,7 @@ async def test_get_health_summary_aggregation():
     
     # Dashboard 2: Single WARN
     rec2 = ValidationRecord(
+        id="rec-warn",
         dashboard_id="dash_2",
         environment_id="env_1",
         status="WARN",
@@ -69,6 +72,8 @@ async def test_get_health_summary_aggregation():
     dash_1_item = next(item for item in summary.items if item.dashboard_id == "dash_1")
     assert dash_1_item.status == "PASS"
     assert dash_1_item.summary == "New pass"
+    assert dash_1_item.record_id == rec1_new.id
+    assert dash_1_item.dashboard_slug == "dash_1"
 
 @pytest.mark.asyncio
 async def test_get_health_summary_empty():
@@ -84,4 +89,160 @@ async def test_get_health_summary_empty():
     assert summary.pass_count == 0
     assert len(summary.items) == 0
 
-# [/DEF:test_health_service:Module]
+
+@pytest.mark.asyncio
+async def test_get_health_summary_resolves_slug_and_title_from_superset():
+    db = MagicMock()
+    config_manager = MagicMock()
+    config_manager.get_environments.return_value = [MagicMock(id="env_1")]
+
+    record = ValidationRecord(
+        id="rec-1",
+        dashboard_id="42",
+        environment_id="env_1",
+        status="PASS",
+        timestamp=datetime.utcnow(),
+        summary="Healthy",
+        issues=[],
+    )
+    db.query.return_value.join.return_value.all.return_value = [record]
+
+    with patch("src.services.health_service.SupersetClient") as mock_client_cls:
+        mock_client = MagicMock()
+        mock_client.get_dashboards_summary.return_value = [
+            {"id": 42, "slug": "ops-overview", "title": "Ops Overview"}
+        ]
+        mock_client_cls.return_value = mock_client
+
+        service = HealthService(db, config_manager=config_manager)
+        summary = await service.get_health_summary(environment_id="env_1")
+
+    assert summary.items[0].dashboard_slug == "ops-overview"
+    assert summary.items[0].dashboard_title == "Ops Overview"
+    mock_client.get_dashboards_summary.assert_called_once_with()
+
+
+def test_delete_validation_report_deletes_dashboard_scope_and_linked_tasks():
+    db = MagicMock()
+    config_manager = MagicMock()
+    task_manager = MagicMock()
+    task_manager.tasks = {"task-1": object(), "task-2": object(), "task-3": object()}
+
+    target_record = ValidationRecord(
+        id="rec-1",
+        task_id="task-1",
+        dashboard_id="42",
+        environment_id="env_1",
+        status="PASS",
+        timestamp=datetime.utcnow(),
+        summary="Healthy",
+        issues=[],
+        screenshot_path=None,
+    )
+    older_peer = ValidationRecord(
+        id="rec-2",
+        task_id="task-2",
+        dashboard_id="42",
+        environment_id="env_1",
+        status="FAIL",
+        timestamp=datetime.utcnow() - timedelta(hours=1),
+        summary="Older",
+        issues=[],
+        screenshot_path=None,
+    )
+    other_environment = ValidationRecord(
+        id="rec-3",
+        task_id="task-3",
+        dashboard_id="42",
+        environment_id="env_2",
+        status="WARN",
+        timestamp=datetime.utcnow(),
+        summary="Other environment",
+        issues=[],
+        screenshot_path=None,
+    )
+
+    first_query = MagicMock()
+    first_query.first.return_value = target_record
+
+    peer_query = MagicMock()
+    peer_query.filter.return_value = peer_query
+    peer_query.all.return_value = [target_record, older_peer]
+
+    db.query.side_effect = [first_query, peer_query]
+
+    with patch("src.services.health_service.TaskCleanupService") as cleanup_cls:
+        cleanup_instance = MagicMock()
+        cleanup_cls.return_value = cleanup_instance
+
+        service = HealthService(db, config_manager=config_manager)
+        deleted = service.delete_validation_report("rec-1", task_manager=task_manager)
+
+    assert deleted is True
+    assert db.delete.call_count == 2
+    db.delete.assert_any_call(target_record)
+    db.delete.assert_any_call(older_peer)
+    db.commit.assert_called_once()
+    cleanup_instance.delete_task_with_logs.assert_any_call("task-1")
+    cleanup_instance.delete_task_with_logs.assert_any_call("task-2")
+    cleanup_instance.delete_task_with_logs.call_count == 2
+    assert "task-1" not in task_manager.tasks
+    assert "task-2" not in task_manager.tasks
+    assert "task-3" in task_manager.tasks
+
+
+def test_delete_validation_report_returns_false_for_unknown_record():
+    db = MagicMock()
+    db.query.return_value.filter.return_value.first.return_value = None
+
+    service = HealthService(db, config_manager=MagicMock())
+
+    assert service.delete_validation_report("missing") is False
+
+
+def test_delete_validation_report_swallows_linked_task_cleanup_failure():
+    db = MagicMock()
+    config_manager = MagicMock()
+    task_manager = MagicMock()
+    task_manager.tasks = {"task-1": object()}
+
+    record = ValidationRecord(
+        id="rec-1",
+        task_id="task-1",
+        dashboard_id="42",
+        environment_id="env_1",
+        status="PASS",
+        timestamp=datetime.utcnow(),
+        summary="Healthy",
+        issues=[],
+        screenshot_path=None,
+    )
+
+    first_query = MagicMock()
+    first_query.first.return_value = record
+
+    peer_query = MagicMock()
+    peer_query.filter.return_value = peer_query
+    peer_query.all.return_value = [record]
+
+    db.query.side_effect = [first_query, peer_query]
+
+    with patch("src.services.health_service.TaskCleanupService") as cleanup_cls, patch(
+        "src.services.health_service.logger"
+    ) as mock_logger:
+        cleanup_instance = MagicMock()
+        cleanup_instance.delete_task_with_logs.side_effect = RuntimeError("cleanup exploded")
+        cleanup_cls.return_value = cleanup_instance
+
+        service = HealthService(db, config_manager=config_manager)
+        deleted = service.delete_validation_report("rec-1", task_manager=task_manager)
+
+    assert deleted is True
+    db.delete.assert_called_once_with(record)
+    db.commit.assert_called_once()
+    cleanup_instance.delete_task_with_logs.assert_called_once_with("task-1")
+    mock_logger.warning.assert_called_once()
+    assert "task-1" not in task_manager.tasks
+
+
+# [/DEF:test_health_service:Module]

backend/src/services/__tests__/test_llm_provider.py

@@ -7,9 +7,15 @@ import pytest
 import os
 from unittest.mock import MagicMock
 from sqlalchemy.orm import Session
+from cryptography.fernet import Fernet
 from src.services.llm_provider import EncryptionManager, LLMProviderService
 from src.models.llm import LLMProvider
-from src.plugins.llm_analysis.models import LLMProviderConfig, ProviderType
+from src.plugins.llm_analysis.models import LLMProviderConfig, LLMProviderType
+
+# [DEF:_test_encryption_key_fixture:Global]
+# @PURPOSE: Ensure encryption-dependent provider tests run with a valid Fernet key.
+os.environ.setdefault("ENCRYPTION_KEY", Fernet.generate_key().decode())
+# [/DEF:_test_encryption_key_fixture:Global]
 
 # @TEST_CONTRACT: EncryptionManagerModel -> Invariants
 # @TEST_INVARIANT: symmetric_encryption
@@ -50,7 +56,7 @@ def test_get_all_providers(service, mock_db):
 
 def test_create_provider(service, mock_db):
     config = LLMProviderConfig(
-        provider_type=ProviderType.OPENAI,
+        provider_type=LLMProviderType.OPENAI,
         name="Test OpenAI",
         base_url="https://api.openai.com",
         api_key="sk-test",
@@ -79,3 +85,32 @@ def test_get_decrypted_api_key(service, mock_db):
 def test_get_decrypted_api_key_not_found(service, mock_db):
     mock_db.query().filter().first.return_value = None
     assert service.get_decrypted_api_key("missing") is None
+
+def test_update_provider_ignores_masked_placeholder_api_key(service, mock_db):
+    existing_encrypted = EncryptionManager().encrypt("secret-value")
+    mock_provider = LLMProvider(
+        id="p1",
+        provider_type="openai",
+        name="Existing",
+        base_url="https://api.openai.com/v1",
+        api_key=existing_encrypted,
+        default_model="gpt-4o",
+        is_active=True,
+    )
+    mock_db.query().filter().first.return_value = mock_provider
+    config = LLMProviderConfig(
+        id="p1",
+        provider_type=LLMProviderType.OPENAI,
+        name="Existing",
+        base_url="https://api.openai.com/v1",
+        api_key="********",
+        default_model="gpt-4o",
+        is_active=False,
+    )
+
+    updated = service.update_provider("p1", config)
+
+    assert updated is mock_provider
+    assert updated.api_key == existing_encrypted
+    assert EncryptionManager().decrypt(updated.api_key) == "secret-value"
+    assert updated.is_active is False

backend/src/services/health_service.py

@@ -4,21 +4,138 @@
 # @PURPOSE: Business logic for aggregating dashboard health status from validation records.
 # @LAYER: Domain/Service
 # @RELATION: DEPENDS_ON -> ValidationRecord
+# @RELATION: DEPENDS_ON -> backend.src.core.superset_client.SupersetClient
+# @RELATION: DEPENDS_ON -> backend.src.core.task_manager.cleanup.TaskCleanupService
 
-from typing import List, Dict, Any
+from typing import List, Dict, Any, Optional, Tuple
 from sqlalchemy.orm import Session
 from sqlalchemy import func, desc
+import os
 from ..models.llm import ValidationRecord
 from ..schemas.health import DashboardHealthItem, HealthSummaryResponse
 from ..core.logger import logger
+from ..core.superset_client import SupersetClient
+from ..core.task_manager.cleanup import TaskCleanupService
+from ..core.task_manager import TaskManager
 
+# [DEF:HealthService:Class]
+# @TIER: STANDARD
+# @PURPOSE: Aggregate latest dashboard validation state and manage persisted health report lifecycle.
+# @RELATION: CALLS -> backend.src.core.superset_client.SupersetClient
+# @RELATION: CALLS -> backend.src.core.task_manager.cleanup.TaskCleanupService
 class HealthService:
     """
     @PURPOSE: Service for managing and querying dashboard health data.
     """
-    def __init__(self, db: Session):
+    # [DEF:HealthService.__init__:Function]
+    # @PURPOSE: Initialize health service with DB session and optional config access for dashboard metadata resolution.
+    # @PRE: db is a valid SQLAlchemy session.
+    # @POST: Service is ready to aggregate summaries and delete health reports.
+    def __init__(self, db: Session, config_manager = None):
         self.db = db
+        self.config_manager = config_manager
+        self._dashboard_meta_cache: Dict[Tuple[str, str], Dict[str, Optional[str]]] = {}
+    # [/DEF:HealthService.__init__:Function]
 
+    # [DEF:HealthService._prime_dashboard_meta_cache:Function]
+    # @PURPOSE: Warm dashboard slug/title cache with one Superset list fetch per environment.
+    # @PRE: records may contain mixed numeric and slug dashboard identifiers.
+    # @POST: Numeric dashboard ids for known environments are cached when discoverable.
+    # @SIDE_EFFECT: May call Superset dashboard list API once per referenced environment.
+    def _prime_dashboard_meta_cache(self, records: List[ValidationRecord]) -> None:
+        if not self.config_manager or not records:
+            return
+
+        numeric_ids_by_env: Dict[str, set[str]] = {}
+        for record in records:
+            environment_id = str(record.environment_id or "").strip()
+            dashboard_id = str(record.dashboard_id or "").strip()
+            if not environment_id or not dashboard_id or not dashboard_id.isdigit():
+                continue
+            cache_key = (environment_id, dashboard_id)
+            if cache_key in self._dashboard_meta_cache:
+                continue
+            numeric_ids_by_env.setdefault(environment_id, set()).add(dashboard_id)
+
+        if not numeric_ids_by_env:
+            return
+
+        environments = {
+            str(getattr(env, "id", "")).strip(): env
+            for env in self.config_manager.get_environments()
+            if str(getattr(env, "id", "")).strip()
+        }
+
+        for environment_id, dashboard_ids in numeric_ids_by_env.items():
+            env = environments.get(environment_id)
+            if not env:
+                for dashboard_id in dashboard_ids:
+                    self._dashboard_meta_cache[(environment_id, dashboard_id)] = {
+                        "slug": None,
+                        "title": None,
+                    }
+                continue
+
+            try:
+                dashboards = SupersetClient(env).get_dashboards_summary()
+                dashboard_meta_map = {
+                    str(item.get("id")): {
+                        "slug": item.get("slug"),
+                        "title": item.get("title"),
+                    }
+                    for item in dashboards
+                    if str(item.get("id") or "").strip()
+                }
+                for dashboard_id in dashboard_ids:
+                    self._dashboard_meta_cache[(environment_id, dashboard_id)] = dashboard_meta_map.get(
+                        dashboard_id,
+                        {"slug": None, "title": None},
+                    )
+            except Exception as exc:
+                logger.warning(
+                    "[HealthService][_prime_dashboard_meta_cache] Failed to preload dashboard metadata for env=%s: %s",
+                    environment_id,
+                    exc,
+                )
+                for dashboard_id in dashboard_ids:
+                    self._dashboard_meta_cache[(environment_id, dashboard_id)] = {
+                        "slug": None,
+                        "title": None,
+                    }
+    # [/DEF:HealthService._prime_dashboard_meta_cache:Function]
+
+    # [DEF:HealthService._resolve_dashboard_meta:Function]
+    # @PURPOSE: Resolve slug/title for a dashboard referenced by persisted validation record.
+    # @PRE: dashboard_id may be numeric or slug-like; environment_id may be empty.
+    # @POST: Returns dict with `slug` and `title` keys, using cache when possible.
+    # @SIDE_EFFECT: May call Superset API through SupersetClient.
+    def _resolve_dashboard_meta(self, dashboard_id: str, environment_id: Optional[str]) -> Dict[str, Optional[str]]:
+        normalized_dashboard_id = str(dashboard_id or "").strip()
+        normalized_environment_id = str(environment_id or "").strip()
+        if not normalized_dashboard_id:
+            return {"slug": None, "title": None}
+
+        if not normalized_dashboard_id.isdigit():
+            return {"slug": normalized_dashboard_id, "title": None}
+
+        if not self.config_manager or not normalized_environment_id:
+            return {"slug": None, "title": None}
+
+        cache_key = (normalized_environment_id, normalized_dashboard_id)
+        cached = self._dashboard_meta_cache.get(cache_key)
+        if cached is not None:
+            return cached
+
+        meta = {"slug": None, "title": None}
+        self._dashboard_meta_cache[cache_key] = meta
+        return meta
+    # [/DEF:HealthService._resolve_dashboard_meta:Function]
+
+    # [DEF:HealthService.get_health_summary:Function]
+    # @PURPOSE: Aggregate latest validation status per dashboard and enrich rows with dashboard slug/title.
+    # @PRE: environment_id may be omitted to aggregate across all environments.
+    # @POST: Returns HealthSummaryResponse with counts and latest record row per dashboard.
+    # @SIDE_EFFECT: May call Superset API to resolve dashboard metadata.
     async def get_health_summary(self, environment_id: str = None) -> HealthSummaryResponse:
         """
         @PURPOSE: Aggregates the latest validation status for all dashboards.
@@ -44,6 +161,8 @@ class HealthService:
         
         records = query.all()
         
+        self._prime_dashboard_meta_cache(records)
+
         items = []
         pass_count = 0
         warn_count = 0
@@ -62,8 +181,12 @@ class HealthService:
                 unknown_count += 1
                 status = "UNKNOWN"
 
+            meta = self._resolve_dashboard_meta(rec.dashboard_id, rec.environment_id)
             items.append(DashboardHealthItem(
+                record_id=rec.id,
                 dashboard_id=rec.dashboard_id,
+                dashboard_slug=meta.get("slug"),
+                dashboard_title=meta.get("title"),
                 environment_id=rec.environment_id or "unknown",
                 status=status,
                 last_check=rec.timestamp,
@@ -80,5 +203,82 @@ class HealthService:
             fail_count=fail_count,
             unknown_count=unknown_count
         )
+    # [/DEF:HealthService.get_health_summary:Function]
 
-# [/DEF:health_service:Module]
+    # [DEF:HealthService.delete_validation_report:Function]
+    # @PURPOSE: Delete one persisted health report and optionally clean linked task/log artifacts.
+    # @PRE: record_id is a validation record identifier.
+    # @POST: Returns True only when a matching record was deleted.
+    # @SIDE_EFFECT: Deletes DB row, optional screenshot file, and optional task/log persistence.
+    def delete_validation_report(self, record_id: str, task_manager: Optional[TaskManager] = None) -> bool:
+        record = self.db.query(ValidationRecord).filter(ValidationRecord.id == record_id).first()
+        if not record:
+            return False
+
+        peer_query = self.db.query(ValidationRecord).filter(
+            ValidationRecord.dashboard_id == record.dashboard_id
+        )
+        if record.environment_id is None:
+            peer_query = peer_query.filter(ValidationRecord.environment_id.is_(None))
+        else:
+            peer_query = peer_query.filter(ValidationRecord.environment_id == record.environment_id)
+
+        records_to_delete = peer_query.all()
+        screenshot_paths = [
+            str(item.screenshot_path or "").strip()
+            for item in records_to_delete
+            if str(item.screenshot_path or "").strip()
+        ]
+        task_ids = {
+            str(item.task_id or "").strip()
+            for item in records_to_delete
+            if str(item.task_id or "").strip()
+        }
+
+        logger.info(
+            "[HealthService][delete_validation_report] Removing %s validation record(s) for dashboard=%s environment=%s triggered_by_record=%s",
+            len(records_to_delete),
+            record.dashboard_id,
+            record.environment_id,
+            record_id,
+        )
+
+        for item in records_to_delete:
+            self.db.delete(item)
+        self.db.commit()
+
+        for screenshot_path in screenshot_paths:
+            try:
+                if os.path.exists(screenshot_path):
+                    os.remove(screenshot_path)
+            except OSError as exc:
+                logger.warning(
+                    "[HealthService][delete_validation_report] Failed to remove screenshot %s: %s",
+                    screenshot_path,
+                    exc,
+                )
+
+        if task_ids and task_manager and self.config_manager:
+            try:
+                cleanup_service = TaskCleanupService(
+                    task_manager.persistence_service,
+                    task_manager.log_persistence_service,
+                    self.config_manager,
+                )
+                for task_id in task_ids:
+                    task_manager.tasks.pop(task_id, None)
+                    cleanup_service.delete_task_with_logs(task_id)
+            except Exception as exc:
+                logger.warning(
+                    "[HealthService][delete_validation_report] Failed to cleanup linked task/logs for dashboard=%s environment=%s: %s",
+                    record.dashboard_id,
+                    record.environment_id,
+                    exc,
+                )
+
+        return True
+    # [/DEF:HealthService.delete_validation_report:Function]
+    
+# [/DEF:HealthService:Class]
+
+# [/DEF:health_service:Module]

backend/src/services/llm_provider.py

@@ -16,6 +16,8 @@ import os
 if TYPE_CHECKING:
     from ..plugins.llm_analysis.models import LLMProviderConfig
 
+MASKED_API_KEY_PLACEHOLDER = "********"
+
 # [DEF:_require_fernet_key:Function]
 # @TIER: CRITICAL
 # @PURPOSE: Load and validate the Fernet key used for secret encryption.
@@ -145,8 +147,12 @@ class LLMProviderService:
             db_provider.provider_type = config.provider_type.value
             db_provider.name = config.name
             db_provider.base_url = config.base_url
-            # Only update API key if provided (not None and not empty)
-            if config.api_key is not None and config.api_key != "":
+            # Ignore masked placeholder values; they are display-only and must not overwrite secrets.
+            if (
+                config.api_key is not None
+                and config.api_key != ""
+                and config.api_key != MASKED_API_KEY_PLACEHOLDER
+            ):
                 db_provider.api_key = self.encryption.encrypt(config.api_key)
             db_provider.default_model = config.default_model
             db_provider.is_active = config.is_active

backend/src/services/profile_service.py

@@ -117,6 +117,33 @@ class ProfileService:
             )
     # [/DEF:get_my_preference:Function]
 
+    # [DEF:get_dashboard_filter_binding:Function]
+    # @PURPOSE: Return only dashboard-filter fields required by dashboards listing hot path.
+    # @PRE: current_user is authenticated.
+    # @POST: Returns normalized username and profile-default filter toggles without security summary expansion.
+    def get_dashboard_filter_binding(self, current_user: User) -> dict:
+        with belief_scope("ProfileService.get_dashboard_filter_binding", f"user_id={current_user.id}"):
+            preference = self._get_preference_row(current_user.id)
+            if preference is None:
+                return {
+                    "superset_username": None,
+                    "superset_username_normalized": None,
+                    "show_only_my_dashboards": False,
+                    "show_only_slug_dashboards": True,
+                }
+
+            return {
+                "superset_username": self._sanitize_username(preference.superset_username),
+                "superset_username_normalized": self._normalize_username(preference.superset_username),
+                "show_only_my_dashboards": bool(preference.show_only_my_dashboards),
+                "show_only_slug_dashboards": bool(
+                    preference.show_only_slug_dashboards
+                    if preference.show_only_slug_dashboards is not None
+                    else True
+                ),
+            }
+    # [/DEF:get_dashboard_filter_binding:Function]
+
     # [DEF:update_my_preference:Function]
     # @PURPOSE: Validate and persist current user's profile preference in self-scoped mode.
     # @PRE: current_user is authenticated and payload is provided.

backend/src/services/rbac_permission_catalog.py

@@ -10,6 +10,7 @@
 
 # [SECTION: IMPORTS]
 import re
+from functools import lru_cache
 from pathlib import Path
 from typing import Iterable, Set, Tuple
 
@@ -80,6 +81,17 @@ def _discover_route_permissions() -> Set[Tuple[str, str]]:
 # [/DEF:_discover_route_permissions:Function]
 
 
+# [DEF:_discover_route_permissions_cached:Function]
+# @PURPOSE: Cache route permission discovery because route source files are static during normal runtime.
+# @PRE: None.
+# @POST: Returns stable discovered route permission pairs without repeated filesystem scans.
+@lru_cache(maxsize=1)
+def _discover_route_permissions_cached() -> Tuple[Tuple[str, str], ...]:
+    with belief_scope("rbac_permission_catalog._discover_route_permissions_cached"):
+        return tuple(sorted(_discover_route_permissions()))
+# [/DEF:_discover_route_permissions_cached:Function]
+
+
 # [DEF:_discover_plugin_execute_permissions:Function]
 # @PURPOSE: Derives dynamic task permissions of form plugin:{plugin_id}:EXECUTE from plugin registry.
 # @PRE: plugin_loader is optional and may expose get_all_plugin_configs.
@@ -108,6 +120,19 @@ def _discover_plugin_execute_permissions(plugin_loader=None) -> Set[Tuple[str, s
 # [/DEF:_discover_plugin_execute_permissions:Function]
 
 
+# [DEF:_discover_plugin_execute_permissions_cached:Function]
+# @PURPOSE: Cache dynamic plugin EXECUTE permission pairs by normalized plugin id tuple.
+# @PRE: plugin_ids is a deterministic tuple of plugin ids.
+# @POST: Returns stable permission tuple without repeated plugin catalog expansion.
+@lru_cache(maxsize=8)
+def _discover_plugin_execute_permissions_cached(
+    plugin_ids: Tuple[str, ...],
+) -> Tuple[Tuple[str, str], ...]:
+    with belief_scope("rbac_permission_catalog._discover_plugin_execute_permissions_cached"):
+        return tuple((f"plugin:{plugin_id}", "EXECUTE") for plugin_id in plugin_ids)
+# [/DEF:_discover_plugin_execute_permissions_cached:Function]
+
+
 # [DEF:discover_declared_permissions:Function]
 # @PURPOSE: Builds canonical RBAC permission catalog from routes and plugin registry.
 # @PRE: plugin_loader may be provided for dynamic task plugin permission discovery.
@@ -115,8 +140,17 @@ def _discover_plugin_execute_permissions(plugin_loader=None) -> Set[Tuple[str, s
 # @RETURN: Set[Tuple[str, str]] - Complete discovered permission set.
 def discover_declared_permissions(plugin_loader=None) -> Set[Tuple[str, str]]:
     with belief_scope("rbac_permission_catalog.discover_declared_permissions"):
-        permissions = _discover_route_permissions()
-        permissions.update(_discover_plugin_execute_permissions(plugin_loader))
+        permissions = set(_discover_route_permissions_cached())
+        plugin_ids = tuple(
+            sorted(
+                {
+                    str(getattr(plugin_config, "id", "") or "").strip()
+                    for plugin_config in (plugin_loader.get_all_plugin_configs() if plugin_loader else [])
+                    if str(getattr(plugin_config, "id", "") or "").strip()
+                }
+            )
+        )
+        permissions.update(_discover_plugin_execute_permissions_cached(plugin_ids))
         return permissions
 # [/DEF:discover_declared_permissions:Function]
 
@@ -153,4 +187,4 @@ def sync_permission_catalog(
         return len(missing_pairs)
 # [/DEF:sync_permission_catalog:Function]
 
-# [/DEF:backend.src.services.rbac_permission_catalog:Module]
+# [/DEF:backend.src.services.rbac_permission_catalog:Module]

frontend/src/components/llm/ProviderConfig.svelte

@@ -142,10 +142,15 @@
     togglingProviderIds = new Set(togglingProviderIds);
 
     try {
-      await requestApi(`/llm/providers/${provider.id}`, "PUT", {
-        ...provider,
+      const updatePayload = {
+        id: provider.id,
+        name: provider.name,
+        provider_type: provider.provider_type,
+        base_url: provider.base_url,
+        default_model: provider.default_model,
         is_active: provider.is_active,
-      });
+      };
+      await requestApi(`/llm/providers/${provider.id}`, "PUT", updatePayload);
     } catch (err) {
       provider.is_active = previousState;
       providers = [...providers];

frontend/src/components/llm/__tests__/provider_config.integration.test.js

@@ -46,6 +46,15 @@ describe('ProviderConfig edit interaction contract', () => {
     expect(source).toContain('await requestApi(`/llm/providers/${provider.id}`, "DELETE")');
     expect(source).toContain('onclick={() => handleDelete(provider)}');
   });
+
+  it('does not forward masked api_key when toggling provider activation', () => {
+    const source = fs.readFileSync(COMPONENT_PATH, 'utf-8');
+
+    expect(source).toContain('const updatePayload = {');
+    expect(source).toContain('provider_type: provider.provider_type');
+    expect(source).toContain('default_model: provider.default_model');
+    expect(source).not.toContain('await requestApi(`/llm/providers/${provider.id}`, "PUT", {\n        ...provider,');
+  });
 });
 // [/DEF:provider_config_edit_contract_tests:Function]
 // [/DEF:frontend.src.components.llm.__tests__.provider_config_integration:Module]

frontend/src/routes/reports/llm/[taskId]/+page.svelte

@@ -27,7 +27,7 @@
   * @TEST_FIXTURE init_state -> {"taskId": "task-1"}
   * @TEST_INVARIANT correct_fetch -> verifies: [init_state]
    */
-  import { onDestroy, onMount } from "svelte";
+  import { onDestroy, onMount, untrack } from "svelte";
   import { page } from "$app/state";
   import { goto } from "$app/navigation";
   import { api } from "$lib/api.js";
@@ -171,7 +171,10 @@
   );
 
   $effect(() => {
-    void loadScreenshotBlobUrls(screenshotPaths);
+    const paths = screenshotPaths;
+    untrack(() => {
+      void loadScreenshotBlobUrls(paths);
+    });
   });
 </script>
 

frontend/src/routes/reports/llm/[taskId]/report_page.contract.test.js

@@ -0,0 +1,32 @@
+// [DEF:frontend.src.routes.reports.llm.taskid.report_page_contract_test:Module]
+// @TIER: STANDARD
+// @SEMANTICS: llm-report, svelte-effect, screenshot-loading, regression-test
+// @PURPOSE: Protect the LLM report page from self-triggering screenshot load effects.
+// @LAYER: UI Tests
+// @RELATION: VERIFIES -> frontend/src/routes/reports/llm/[taskId]/+page.svelte
+
+import { describe, it, expect } from "vitest";
+import fs from "node:fs";
+import path from "node:path";
+
+const PAGE_PATH = path.resolve(
+  process.cwd(),
+  "src/routes/reports/llm/[taskId]/+page.svelte",
+);
+
+// [DEF:llm_report_screenshot_effect_contract:Function]
+// @TIER: STANDARD
+// @PURPOSE: Ensure screenshot loading stays untracked from blob-url mutation state.
+// @PRE: Report page source exists.
+// @POST: Contract fails if screenshot loading effect can subscribe to screenshotBlobUrls updates.
+describe("LLM report screenshot effect contract", () => {
+  it("uses untrack around screenshot blob loading side effect", () => {
+    const source = fs.readFileSync(PAGE_PATH, "utf-8");
+
+    expect(source).toContain('import { onDestroy, onMount, untrack } from "svelte";');
+    expect(source).toContain("untrack(() => {");
+    expect(source).toContain("void loadScreenshotBlobUrls(paths);");
+  });
+});
+// [/DEF:llm_report_screenshot_effect_contract:Function]
+// [/DEF:frontend.src.routes.reports.llm.taskid.report_page_contract_test:Module]