Merge remote-tracking branch 'origin/feat/ui-onboarding-rewrite' into feat/agent-v2

# Conflicts:
#	web/app/(commonLayout)/datasets/(datasetDetailLayout)/[datasetId]/layout-main.tsx
#	web/app/(commonLayout)/layout.tsx
#	web/app/components/full-screen-loading.tsx
#	web/app/components/header/header-wrapper.tsx

.agents/skills/frontend-code-review/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: frontend-code-review
-description: "Trigger when the user requests a review of frontend files (e.g., `.tsx`, `.ts`, `.js`). Support both pending-change reviews and focused file reviews while applying the checklist rules."
+description: "Trigger when the user requests a review of frontend files (e.g., `.tsx`, `.ts`, `.js`). Support pending-change and focused file reviews while applying checklist rules, shared component reuse checks, and React component structure guidance from how-to-write-component."
 ---
 
 # Frontend Code Review
@@ -16,10 +16,12 @@ Stick to the checklist below for every applicable file and mode.
 ## Checklist
 See [references/code-quality.md](references/code-quality.md), [references/performance.md](references/performance.md), [references/business-logic.md](references/business-logic.md) for the living checklist split by category—treat it as the canonical set of rules to follow.
 
+When reviewing React/TypeScript components, also apply the repo-local `how-to-write-component` skill as the component architecture checklist. In particular, check ownership boundaries, props and API types, query/mutation usage, navigation choices, effect usage, unnecessary wrappers, and unnecessary memoization.
+
 Flag each rule violation with urgency metadata so future reviewers can prioritize fixes.
 
 ## Review Process
-1. Open the relevant component/module. Gather lines that relate to class names, React Flow hooks, prop memoization, and styling.
+1. Open the relevant component/module. Gather lines that relate to shared base/dify-ui component reuse, class names, styling/CSS imports, file size and component boundaries, i18n keys, behavior-sensitive UI interactions, React Flow hooks, and prop memoization.
 2. For each rule in the review point, note where the code deviates and capture a representative snippet.
 3. Compose the review section per the template below. Group violations first by **Urgent** flag, then by category order (Code Quality, Performance, Business Logic).
 
@@ -70,4 +72,3 @@ If you use Template A (i.e., there are issues to fix) and at least one issue req
 ## Code review
 No issues found.
 ```
-

.agents/skills/frontend-code-review/references/business-logic.md

@@ -13,3 +13,29 @@ Node components are also used when creating a RAG Pipe from a template, but in t
 ### Suggested Fix
 
 Use `import { useNodes } from 'reactflow'` instead of `import useNodes from '@/app/components/workflow/store/workflow/use-nodes'`.
+
+## Locale keys must be complete
+
+IsUrgent: True
+Category: Business Logic
+
+### Description
+
+When adding or changing user-facing i18n keys, ensure every supported locale file has the same key set as `web/i18n/en-US/`. Do not add only English keys or only a partial subset of locales; `pnpm i18n:check --file <name>` should pass for the touched translation file.
+
+### Suggested Fix
+
+Add matching keys to every existing supported locale file for the touched translation namespace, keeping key paths aligned with the English entry.
+
+## Preserve behavior-sensitive interactions
+
+IsUrgent: True
+Category: Business Logic
+
+### Description
+
+When changing existing navigation, sidebar, dropdown, webapp list, or app-switching UI, compare behavior against the existing implementation before approving the change. Watch for regressions in expand/collapse arrows, hover persistence, pin/delete controls, routing, keyboard/focus handling, and open-state ownership.
+
+### Suggested Fix
+
+Reuse or extend the existing component when it already owns the interaction logic. If a refactor is needed, preserve the old interaction contract and add or update focused tests for the changed behavior.

.agents/skills/frontend-code-review/references/code-quality.md

@@ -7,12 +7,12 @@ Category: Code Quality
 
 ### Description
 
-Ensure conditional CSS is handled via the shared `classNames` instead of custom ternaries, string concatenation, or template strings. Centralizing class logic keeps components consistent and easier to maintain.
+Ensure conditional CSS and multi-line class composition are handled via the shared `cn` helper instead of custom ternaries, string concatenation, array `.join(' ')`, or template strings. Centralizing class logic keeps components consistent and easier to maintain.
 
 ### Suggested Fix
 
 ```ts
-import { cn } from '@/utils/classnames'
+import { cn } from '@langgenius/dify-ui/cn'
 const classNames = cn(isActive ? 'text-primary-600' : 'text-gray-500')
 ```
 
@@ -25,7 +25,34 @@ Category: Code Quality
 
 Favor Tailwind CSS utility classes instead of adding new `.module.css` files unless a Tailwind combination cannot achieve the required styling. Keeping styles in Tailwind improves consistency and reduces maintenance overhead.
 
-Update this file when adding, editing, or removing Code Quality rules so the catalog remains accurate.
+## CSS files must be scoped
+
+IsUrgent: True
+Category: Code Quality
+
+### Description
+
+When CSS is truly necessary, use component-scoped `*.module.css`. Do not add component-level CSS through plain `.css` files, and do not import component CSS from `globals.css`; both patterns risk style leakage across the app.
+
+## Split oversized components cautiously
+
+Category: Code Quality
+
+### Description
+
+When a frontend file grows large or mixes multiple responsibilities, suggest splitting it into focused components, hooks, or utilities. Prefer shallow local structure that matches existing repo patterns, such as a sibling `components/` folder, and avoid deep folder hierarchies unless the surrounding code already uses them.
+
+## Reuse base and dify-ui components before hand-rolling UI
+
+Category: Code Quality
+
+### Description
+
+Before approving new or modified frontend UI, check whether the code manually recreates behavior or styling already owned by `@langgenius/dify-ui/*` or `web/app/components/base/*`. Common examples include `Button`, `Input`, `ToggleGroup`, `Popover`, `DropdownMenu`, `AlertDialog`, `Switch`, `Avatar`, `ScrollArea`, `toast`, and existing feature components. Prefer composing existing primitives instead of duplicating borders, focus states, disabled states, segmented controls, inputs, overlays, or buttons.
+
+### Suggested Fix
+
+Replace hand-written UI chrome with the nearest shared primitive, keeping feature-specific layout, state ownership, labels, and workflow behavior local.
 
 ## Classname ordering for easy overrides
 
@@ -36,9 +63,11 @@ When writing components, always place the incoming `className` prop after the co
 Example:
 
 ```tsx
-import { cn } from '@/utils/classnames'
+import { cn } from '@langgenius/dify-ui/cn'
 
 const Button = ({ className }) => {
   return <div className={cn('bg-primary-600', className)}></div>
 }
 ```
+
+Update this file when adding, editing, or removing Code Quality rules so the catalog remains accurate.

.agents/skills/frontend-code-review/references/performance.md

@@ -43,3 +43,14 @@ const config = useMemo(() => ({
     config={config}
 />
 ```
+
+## Custom SVG icon generation
+
+IsUrgent: False
+Category: Performance
+
+### Description
+
+New custom SVG icons should be added to `packages/iconify-collections/assets/...`, generated with `pnpm --filter @dify/iconify-collections generate`, checked with `pnpm --filter @dify/iconify-collections check:dimensions`, and consumed through Tailwind `i-custom-*` classes. Do not add new generated React icon components or JSON files under `web/app/components/base/icons/src/...` for new custom SVG icons.
+
+When reviewing generated `packages/iconify-collections/custom-*/icons.json` diffs, verify unrelated existing icons did not lose or change intrinsic `width` / `height`.

.agents/skills/karpathy-guidelines/SKILL.md

@@ -0,0 +1,33 @@
+---
+name: karpathy-guidelines
+description: Lightweight coding guardrails for making focused, simple, and verifiable changes in this repo. Use for all coding work.
+---
+
+# Karpathy Guidelines
+
+Use this skill whenever you touch code in this repository.
+
+## Principles
+
+- Keep the change small and directly tied to the user request.
+- Prefer the simplest implementation that fits the existing codebase.
+- Read the nearby code first, then match its patterns.
+- Avoid unrelated refactors, broad rewrites, or style churn.
+- Preserve existing behavior unless the user explicitly asked to change it.
+- Treat regressions as a signal to narrow the change, not to add workaround layers.
+
+## Workflow
+
+1. Inspect the current implementation and tests around the change.
+2. Make the smallest coherent edit.
+3. Add or update focused tests when the behavior changes or the risk is non-trivial.
+4. Run the narrowest relevant verification first.
+5. Report exactly what was verified and anything left unverified.
+
+## Review Checklist
+
+- Does this change solve the stated problem without expanding scope?
+- Did it preserve existing route/component/data-flow semantics?
+- Are new abstractions justified by real complexity?
+- Are tests focused on the behavior that could regress?
+- Are unrelated files and generated artifacts left alone?

.github/CODEOWNERS

@@ -166,6 +166,7 @@
 
 # Frontend - App - API Documentation
 /web/app/components/develop/ @JzoNgKVO @iamjoel
+/web/app/components/develop/template/*.mdx @JzoNgKVO @iamjoel @RiskeyL
 
 # Frontend - App - Logs and Annotations
 /web/app/components/app/workflow-log/ @JzoNgKVO @iamjoel

.github/workflows/cli-tests.yml

@@ -15,8 +15,12 @@ concurrency:
 
 jobs:
   test:
-    name: CLI Tests
-    runs-on: depot-ubuntu-24.04
+    name: CLI Tests (${{ matrix.os }})
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        os: [depot-ubuntu-24.04, windows-latest, macos-latest]
     env:
       CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}
     defaults:
@@ -37,7 +41,7 @@ jobs:
         run: pnpm ci
 
       - name: Report coverage
-        if: ${{ env.CODECOV_TOKEN != '' }}
+        if: ${{ env.CODECOV_TOKEN != '' && matrix.os == 'depot-ubuntu-24.04' }}
         uses: codecov/codecov-action@57e3a136b779b570ffcdbf80b3bdc90e7fab3de2 # v6.0.0
         with:
           directory: cli/coverage

.github/workflows/deploy-saas.yml

@@ -1,4 +1,4 @@
-name: Deploy Agent Dev
+name: Deploy SaaS
 
 permissions:
   contents: read
@@ -7,7 +7,7 @@ on:
   workflow_run:
     workflows: ["Build and Push API & Web"]
     branches:
-      - "deploy/agent-dev"
+      - "deploy/saas"
     types:
       - completed
 
@@ -16,13 +16,13 @@ jobs:
     runs-on: depot-ubuntu-24.04
     if: |
       github.event.workflow_run.conclusion == 'success' &&
-      github.event.workflow_run.head_branch == 'deploy/agent-dev'
+      github.event.workflow_run.head_branch == 'deploy/saas'
     steps:
       - name: Deploy to server
         uses: appleboy/ssh-action@0ff4204d59e8e51228ff73bce53f80d53301dee2 # v1.2.5
         with:
-          host: ${{ secrets.AGENT_DEV_SSH_HOST }}
+          host: ${{ secrets.SAAS_DEV_SSH_HOST }}
           username: ${{ secrets.SSH_USER }}
           key: ${{ secrets.SSH_PRIVATE_KEY }}
           script: |
-            ${{ vars.SSH_SCRIPT || secrets.SSH_SCRIPT }}
+            ${{ vars.SSH_SCRIPT_SAAS_DEV || secrets.SSH_SCRIPT_SAAS_DEV }}

.github/workflows/style.yml

@@ -95,6 +95,51 @@ jobs:
         if: steps.changed-files.outputs.any_changed == 'true'
         uses: ./.github/actions/setup-web
 
+      - name: Web tsslint
+        if: steps.changed-files.outputs.any_changed == 'true'
+        env:
+          NODE_OPTIONS: --max-old-space-size=4096
+        run: vp run lint:tss
+
+      - name: Web dead code check
+        if: steps.changed-files.outputs.any_changed == 'true'
+        run: vp run knip
+
+  ts-common-style:
+    name: TS Common
+    runs-on: depot-ubuntu-24.04
+    permissions:
+      checks: write
+      pull-requests: read
+
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+        with:
+          persist-credentials: false
+
+      - name: Check changed files
+        id: changed-files
+        uses: tj-actions/changed-files@9426d40962ed5378910ee2e21d5f8c6fcbf2dd96 # v47.0.6
+        with:
+          files: |
+            web/**
+            cli/**
+            e2e/**
+            sdks/nodejs-client/**
+            packages/**
+            package.json
+            pnpm-lock.yaml
+            pnpm-workspace.yaml
+            .nvmrc
+            eslint.config.mjs
+            .github/workflows/style.yml
+            .github/actions/setup-web/**
+
+      - name: Setup web environment
+        if: steps.changed-files.outputs.any_changed == 'true'
+        uses: ./.github/actions/setup-web
+
       - name: Restore ESLint cache
         if: steps.changed-files.outputs.any_changed == 'true'
         id: eslint-cache-restore
@@ -105,28 +150,14 @@ jobs:
           restore-keys: |
             ${{ runner.os }}-eslint-${{ hashFiles('pnpm-lock.yaml', 'eslint.config.mjs', 'web/eslint.config.mjs', 'web/eslint.constants.mjs', 'web/plugins/eslint/**') }}-
 
-      - name: Web style check
+      - name: Style check
         if: steps.changed-files.outputs.any_changed == 'true'
-        working-directory: .
         run: vp run lint:ci
 
-      - name: Web tsslint
+      - name: Type check
         if: steps.changed-files.outputs.any_changed == 'true'
-        working-directory: ./web
-        env:
-          NODE_OPTIONS: --max-old-space-size=4096
-        run: vp run lint:tss
-
-      - name: Web type check
-        if: steps.changed-files.outputs.any_changed == 'true'
-        working-directory: .
         run: vp run type-check
 
-      - name: Web dead code check
-        if: steps.changed-files.outputs.any_changed == 'true'
-        working-directory: ./web
-        run: vp run knip
-
       - name: Save ESLint cache
         if: steps.changed-files.outputs.any_changed == 'true' && success() && steps.eslint-cache-restore.outputs.cache-hit != 'true'
         uses: actions/cache/save@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5

MOCKS_TO_REMOVE_BEFORE_RELEASE.md

@@ -0,0 +1,4 @@
+# Mocks to Remove Before Release
+
+- `emptyAppList=true`: frontend URL preview flag for forcing the `/apps` page into the first-empty state. Remove the parser and rendering override before release.
+- `emptyDataList=true`: frontend URL preview flag for forcing the `/datasets` page into the first-empty state. Remove the parser and rendering override before release.

api/Dockerfile

@@ -27,7 +27,7 @@ COPY api/providers ./providers
 COPY dify-agent/pyproject.toml dify-agent/README.md /app/dify-agent/
 COPY dify-agent/src /app/dify-agent/src
 # Trust the checked-in lock during image builds; local path sources are copied from the repository context.
-RUN uv sync --frozen --no-dev
+RUN uv sync --frozen --no-dev --no-editable
 
 # production stage
 FROM base AS production

api/app_factory.py

@@ -223,10 +223,11 @@ def initialize_extensions(app: DifyApp):
 
 def create_migrations_app() -> DifyApp:
     app = create_flask_app_with_configs()
-    from extensions import ext_database, ext_migrate
+    from extensions import ext_commands, ext_database, ext_migrate
 
     # Initialize only required extensions
     ext_database.init_app(app)
     ext_migrate.init_app(app)
+    ext_commands.init_app(app)
 
     return app

api/clients/agent_backend/__init__.py

@@ -31,18 +31,22 @@ from clients.agent_backend.fake_client import FakeAgentBackendRunClient, FakeAge
 from clients.agent_backend.request_builder import (
     AGENT_SOUL_PROMPT_LAYER_ID,
     DIFY_EXECUTION_CONTEXT_LAYER_ID,
+    DIFY_PLUGIN_TOOLS_LAYER_ID,
     WORKFLOW_NODE_JOB_PROMPT_LAYER_ID,
     WORKFLOW_USER_PROMPT_LAYER_ID,
     AgentBackendModelConfig,
     AgentBackendOutputConfig,
     AgentBackendRunRequestBuilder,
     AgentBackendWorkflowNodeRunInput,
+    CleanupLayerSpec,
+    extract_cleanup_layer_specs,
     redact_for_agent_backend_log,
 )
 
 __all__ = [
     "AGENT_SOUL_PROMPT_LAYER_ID",
     "DIFY_EXECUTION_CONTEXT_LAYER_ID",
+    "DIFY_PLUGIN_TOOLS_LAYER_ID",
     "WORKFLOW_NODE_JOB_PROMPT_LAYER_ID",
     "WORKFLOW_USER_PROMPT_LAYER_ID",
     "AgentBackendError",
@@ -66,9 +70,11 @@ __all__ = [
     "AgentBackendTransportError",
     "AgentBackendValidationError",
     "AgentBackendWorkflowNodeRunInput",
+    "CleanupLayerSpec",
     "DifyAgentBackendRunClient",
     "FakeAgentBackendRunClient",
     "FakeAgentBackendScenario",
     "create_agent_backend_run_client",
+    "extract_cleanup_layer_specs",
     "redact_for_agent_backend_log",
 ]

api/clients/agent_backend/fake_client.py

@@ -20,6 +20,8 @@ from dify_agent.protocol import (
     RunEvent,
     RunFailedEvent,
     RunFailedEventData,
+    RunPausedEvent,
+    RunPausedEventData,
     RunStartedEvent,
     RunStatusResponse,
     RunSucceededEvent,
@@ -34,6 +36,7 @@ class FakeAgentBackendScenario(StrEnum):
 
     SUCCESS = "success"
     FAILED = "failed"
+    PAUSED = "paused"
 
 
 class FakeAgentBackendRunClient:
@@ -89,6 +92,13 @@ class FakeAgentBackendRunClient:
                     updated_at=_FIXED_TIME,
                     error="fake failure",
                 )
+            case FakeAgentBackendScenario.PAUSED:
+                return RunStatusResponse(
+                    run_id=run_id,
+                    status="paused",
+                    created_at=_FIXED_TIME,
+                    updated_at=_FIXED_TIME,
+                )
 
     def _events(self, run_id: str) -> tuple[RunEvent, ...]:
         match self.scenario:
@@ -115,3 +125,17 @@ class FakeAgentBackendRunClient:
                         data=RunFailedEventData(error="fake failure", reason="unit_test"),
                     ),
                 )
+            case FakeAgentBackendScenario.PAUSED:
+                return (
+                    RunStartedEvent(id="1-0", run_id=run_id, created_at=_FIXED_TIME),
+                    RunPausedEvent(
+                        id="2-0",
+                        run_id=run_id,
+                        created_at=_FIXED_TIME,
+                        data=RunPausedEventData(
+                            reason="human_input_required",
+                            message="Agent requested human input.",
+                            session_snapshot=CompositorSessionSnapshot(layers=[]),
+                        ),
+                    ),
+                )

api/clients/agent_backend/request_builder.py

@@ -11,15 +11,19 @@ composition-driven.
 
 from __future__ import annotations
 
-from typing import ClassVar
+from typing import ClassVar, cast
 
 from agenton.compositor import CompositorSessionSnapshot
+from agenton.compositor.schemas import LayerSessionSnapshot
 from agenton.layers import ExitIntent
 from agenton_collections.layers.plain import PLAIN_PROMPT_LAYER_TYPE_ID, PromptLayerConfig
+from agenton_collections.layers.pydantic_ai import PYDANTIC_AI_HISTORY_LAYER_TYPE_ID
 from dify_agent.layers.dify_plugin import (
     DIFY_PLUGIN_LLM_LAYER_TYPE_ID,
+    DIFY_PLUGIN_TOOLS_LAYER_TYPE_ID,
     DifyPluginCredentialValue,
     DifyPluginLLMLayerConfig,
+    DifyPluginToolsLayerConfig,
 )
 from dify_agent.layers.execution_context import (
     DIFY_EXECUTION_CONTEXT_LAYER_TYPE_ID,
@@ -27,6 +31,7 @@ from dify_agent.layers.execution_context import (
 )
 from dify_agent.layers.output import DIFY_OUTPUT_LAYER_TYPE_ID, DifyOutputLayerConfig
 from dify_agent.protocol import (
+    DIFY_AGENT_HISTORY_LAYER_ID,
     DIFY_AGENT_MODEL_LAYER_ID,
     DIFY_AGENT_OUTPUT_LAYER_ID,
     CreateRunRequest,
@@ -41,6 +46,85 @@ AGENT_SOUL_PROMPT_LAYER_ID = "agent_soul_prompt"
 WORKFLOW_NODE_JOB_PROMPT_LAYER_ID = "workflow_node_job_prompt"
 WORKFLOW_USER_PROMPT_LAYER_ID = "workflow_user_prompt"
 DIFY_EXECUTION_CONTEXT_LAYER_ID = "execution_context"
+DIFY_PLUGIN_TOOLS_LAYER_ID = "tools"
+
+# Layer types that hold credentials in their per-run config. These are excluded
+# from the cleanup-replay composition (and from the snapshot that is sent with
+# the cleanup request) because we deliberately do not persist plaintext
+# credentials between runs.
+_CLEANUP_EXCLUDED_LAYER_TYPES: tuple[str, ...] = (
+    DIFY_PLUGIN_LLM_LAYER_TYPE_ID,
+    DIFY_PLUGIN_TOOLS_LAYER_TYPE_ID,
+)
+
+
+class CleanupLayerSpec(BaseModel):
+    """One layer node replayed by an Agent backend cleanup-only run.
+
+    Cleanup composition cannot include credential-bearing plugin layers, so we
+    persist only the non-plugin layer specs together with the original config.
+    Storing the config (rather than just ``name``/``type``) means cleanup does
+    not depend on the original build-time inputs being re-derivable.
+    """
+
+    name: str
+    type: str
+    deps: dict[str, str] = Field(default_factory=dict)
+    metadata: dict[str, JsonValue] = Field(default_factory=dict)
+    config: JsonValue = None
+
+    model_config: ClassVar[ConfigDict] = ConfigDict(extra="forbid")
+
+
+def extract_cleanup_layer_specs(composition: RunComposition) -> list[CleanupLayerSpec]:
+    """Project the in-flight composition into the persistable cleanup spec list.
+
+    Plugin layers are intentionally dropped (their configs hold credentials and
+    the lifecycle contract says "do not include an LLM layer" during cleanup).
+    The filtered names must later drive snapshot filtering so the agenton
+    compositor's name-order check still passes for the cleanup run.
+    """
+    excluded = set(_CLEANUP_EXCLUDED_LAYER_TYPES)
+    specs: list[CleanupLayerSpec] = []
+    for layer in composition.layers:
+        if layer.type in excluded:
+            continue
+        config_value: JsonValue = None
+        if isinstance(layer.config, BaseModel):
+            config_value = layer.config.model_dump(mode="json", warnings=False)
+        else:
+            # ``RunLayerSpec.config`` is typed as ``LayerConfigInput`` which
+            # includes ``Mapping[str, object] | bytes``. In the cleanup-replay
+            # pipeline our builder only emits BaseModel-derived configs or
+            # ``None``, so the wider input alias narrows safely here.
+            config_value = cast(JsonValue, layer.config)
+        specs.append(
+            CleanupLayerSpec(
+                name=layer.name,
+                type=layer.type,
+                deps=dict(layer.deps),
+                metadata=dict(layer.metadata),
+                config=config_value,
+            )
+        )
+    return specs
+
+
+def _filter_snapshot_to_specs(
+    snapshot: CompositorSessionSnapshot,
+    specs: list[CleanupLayerSpec],
+) -> CompositorSessionSnapshot:
+    """Keep only snapshot layers whose names appear in the cleanup spec list.
+
+    The agenton compositor rejects a snapshot whose layer-name sequence does
+    not match the active composition exactly. Cleanup-replay drops plugin
+    layers, so we must drop the matching snapshot entries here.
+    """
+    kept_names = {spec.name for spec in specs}
+    filtered_layers: list[LayerSessionSnapshot] = [layer for layer in snapshot.layers if layer.name in kept_names]
+    if len(filtered_layers) == len(snapshot.layers):
+        return snapshot
+    return CompositorSessionSnapshot(schema_version=snapshot.schema_version, layers=filtered_layers)
 
 
 class AgentBackendModelConfig(BaseModel):
@@ -81,8 +165,10 @@ class AgentBackendWorkflowNodeRunInput(BaseModel):
     purpose: RunPurpose = "workflow_node"
     idempotency_key: str | None = None
     output: AgentBackendOutputConfig | None = None
+    tools: DifyPluginToolsLayerConfig | None = None
     session_snapshot: CompositorSessionSnapshot | None = None
-    suspend_on_exit: bool = False
+    include_history: bool = True
+    suspend_on_exit: bool = True
     metadata: dict[str, JsonValue] = Field(default_factory=dict)
 
     model_config: ClassVar[ConfigDict] = ConfigDict(extra="forbid", arbitrary_types_allowed=True)
@@ -98,6 +184,50 @@ class AgentBackendWorkflowNodeRunInput(BaseModel):
 class AgentBackendRunRequestBuilder:
     """Converts API product state into the public ``dify-agent`` run protocol."""
 
+    def build_cleanup_request(
+        self,
+        *,
+        session_snapshot: CompositorSessionSnapshot,
+        composition_layer_specs: list[CleanupLayerSpec],
+        idempotency_key: str | None = None,
+        metadata: dict[str, JsonValue] | None = None,
+    ) -> CreateRunRequest:
+        """Build a lifecycle-only cleanup request that replays the prior layers.
+
+        The agenton compositor enforces that the session snapshot's layer names
+        match the active composition in order, so cleanup must replay the same
+        non-plugin layer graph that produced the snapshot. Plugin layers
+        (``dify.plugin.llm``, ``dify.plugin.tools``) are excluded from both the
+        composition and the snapshot before submission because their configs
+        require credentials that are not persisted between runs.
+        """
+        if not composition_layer_specs:
+            raise ValueError(
+                "build_cleanup_request requires composition_layer_specs; an empty "
+                "composition would fail the agent backend's snapshot validation."
+            )
+        request_metadata = dict(metadata or {})
+        request_metadata["agent_backend_lifecycle"] = "session_cleanup"
+        layers = [
+            RunLayerSpec(
+                name=spec.name,
+                type=spec.type,
+                deps=dict(spec.deps),
+                metadata=dict(spec.metadata),
+                config=spec.config,
+            )
+            for spec in composition_layer_specs
+        ]
+        filtered_snapshot = _filter_snapshot_to_specs(session_snapshot, composition_layer_specs)
+        return CreateRunRequest(
+            composition=RunComposition(layers=layers),
+            purpose="workflow_node",
+            idempotency_key=idempotency_key,
+            metadata=request_metadata,
+            session_snapshot=filtered_snapshot,
+            on_exit=LayerExitSignals(default=ExitIntent.DELETE),
+        )
+
     def build_for_workflow_node(self, run_input: AgentBackendWorkflowNodeRunInput) -> CreateRunRequest:
         """Build a workflow Agent Node run request without defining another wire schema."""
         layers: list[RunLayerSpec] = []
@@ -131,6 +261,20 @@ class AgentBackendRunRequestBuilder:
                     metadata=run_input.metadata,
                     config=run_input.execution_context,
                 ),
+            ]
+        )
+
+        if run_input.include_history:
+            layers.append(
+                RunLayerSpec(
+                    name=DIFY_AGENT_HISTORY_LAYER_ID,
+                    type=PYDANTIC_AI_HISTORY_LAYER_TYPE_ID,
+                    metadata={**run_input.metadata, "origin": "agent_session_history"},
+                )
+            )
+
+        layers.extend(
+            [
                 RunLayerSpec(
                     name=DIFY_AGENT_MODEL_LAYER_ID,
                     type=DIFY_PLUGIN_LLM_LAYER_TYPE_ID,
@@ -147,6 +291,17 @@ class AgentBackendRunRequestBuilder:
             ]
         )
 
+        if run_input.tools is not None and run_input.tools.tools:
+            layers.append(
+                RunLayerSpec(
+                    name=DIFY_PLUGIN_TOOLS_LAYER_ID,
+                    type=DIFY_PLUGIN_TOOLS_LAYER_TYPE_ID,
+                    deps={"execution_context": DIFY_EXECUTION_CONTEXT_LAYER_ID},
+                    metadata=run_input.metadata,
+                    config=run_input.tools,
+                )
+            )
+
         if run_input.output is not None:
             layers.append(
                 RunLayerSpec(

api/commands/__init__.py

@@ -3,7 +3,15 @@ CLI command modules extracted from `commands.py`.
 """
 
 from .account import create_tenant, reset_email, reset_password
+from .data_migrate import data_migrate, legacy_model_types
+from .data_migration import (
+    export_migration_data,
+    export_migration_data_template,
+    import_migration_data,
+    migration_data_wizard,
+)
 from .plugin import (
+    backfill_plugin_auto_upgrade,
     extract_plugins,
     extract_unique_plugins,
     install_plugins,
@@ -25,7 +33,12 @@ from .retention import (
     restore_workflow_runs,
 )
 from .storage import clear_orphaned_file_records, file_usage, migrate_oss, remove_orphaned_files_on_storage
-from .system import convert_to_agent_apps, fix_app_site_missing, reset_encrypt_key_pair, upgrade_db
+from .system import (
+    convert_to_agent_apps,
+    fix_app_site_missing,
+    reset_encrypt_key_pair,
+    upgrade_db,
+)
 from .vector import (
     add_qdrant_index,
     migrate_annotation_vector_database,
@@ -37,6 +50,7 @@ from .vector import (
 __all__ = [
     "add_qdrant_index",
     "archive_workflow_runs",
+    "backfill_plugin_auto_upgrade",
     "clean_expired_messages",
     "clean_workflow_runs",
     "cleanup_orphaned_draft_variables",
@@ -44,18 +58,24 @@ __all__ = [
     "clear_orphaned_file_records",
     "convert_to_agent_apps",
     "create_tenant",
+    "data_migrate",
     "delete_archived_workflow_runs",
     "export_app_messages",
+    "export_migration_data",
+    "export_migration_data_template",
     "extract_plugins",
     "extract_unique_plugins",
     "file_usage",
     "fix_app_site_missing",
+    "import_migration_data",
     "install_plugins",
     "install_rag_pipeline_plugins",
+    "legacy_model_types",
     "migrate_annotation_vector_database",
     "migrate_data_for_plugin",
     "migrate_knowledge_vector_database",
     "migrate_oss",
+    "migration_data_wizard",
     "old_metadata_migration",
     "remove_orphaned_files_on_storage",
     "reset_email",

api/commands/data_migrate.py

@@ -0,0 +1,179 @@
+import io
+import os
+import sys
+from contextlib import AbstractContextManager, nullcontext
+from pathlib import Path
+from typing import cast
+
+import click
+
+from extensions.ext_database import db
+from graphon.model_runtime.entities.model_entities import ModelType
+from services.legacy_model_type_migration import (
+    VALID_TABLE_NAMES,
+    LegacyModelTypeMigrationService,
+    load_tenant_ids_from_file,
+)
+
+_SUPPORTED_MODEL_TYPE_CHOICES = (
+    ModelType.LLM.value,
+    ModelType.TEXT_EMBEDDING.value,
+    ModelType.RERANK.value,
+)
+_DEFAULT_CONCURRENCY = os.cpu_count() or 1
+
+
+def _normalize_multi_value_option(
+    values: tuple[str, ...],
+    *,
+    valid_values: tuple[str, ...],
+    option_name: str,
+) -> tuple[str, ...]:
+    normalized_values: list[str] = []
+    seen_values: set[str] = set()
+
+    for value in values:
+        for item in value.split(","):
+            normalized_item = item.strip()
+            if not normalized_item:
+                continue
+            if normalized_item not in valid_values:
+                raise click.BadParameter(
+                    f"invalid value '{normalized_item}'. valid values: {', '.join(valid_values)}",
+                    param_hint=option_name,
+                )
+            if normalized_item in seen_values:
+                continue
+            seen_values.add(normalized_item)
+            normalized_values.append(normalized_item)
+
+    return tuple(normalized_values)
+
+
+@click.group(
+    "data-migrate",
+    help="Online data migration commands.",
+)
+def data_migrate() -> None:
+    """Namespace for production data migration commands."""
+
+
+@click.command(
+    "legacy-model-types",
+    help=(
+        "Migrate legacy provider model_type values to canonical values. "
+        "Default is dry-run and emits JSON lines only. "
+        "If --tables includes provider_model_credentials, the command may also update "
+        "provider_models and load_balancing_model_configs references so merged credentials stay reachable."
+    ),
+)
+@click.option(
+    "--apply",
+    is_flag=True,
+    default=False,
+    help="Apply the migration. Default is dry-run.",
+)
+@click.option(
+    "--tables",
+    "tables",
+    multiple=True,
+    type=str,
+    help=(
+        "Limit migration to specific tables. Accepts comma-separated values or repeated flags.\n"
+        "\n"
+        "Options: load_balancing_model_configs, provider_model_credentials, "
+        "provider_model_settings, provider_models, tenant_default_models.\n\n"
+        "When provider_model_credentials is selected, provider_models and "
+        "load_balancing_model_configs may also be updated for credential reference rewrites.\n"
+        "\n"
+        "If unspecified, all relevant tables are migrated."
+    ),
+)
+@click.option(
+    "--model-types",
+    "model_types",
+    multiple=True,
+    type=str,
+    help=(
+        "Canonical model types to migrate. Accepts comma-separated values or repeated flags.\n"
+        "\n"
+        "Options: llm,text-embedding,rerank\n"
+        "\n"
+        "If unspecified, all relevant legacy model types are migrated."
+    ),
+)
+@click.option(
+    "--tenant-id-file",
+    type=click.Path(exists=True, dir_okay=False, readable=True, resolve_path=True),
+    help="Optional file containing tenant ids, one per line.",
+)
+@click.option(
+    "--output",
+    type=click.Path(dir_okay=False, resolve_path=True, path_type=Path),
+    help=(
+        "Optional file path for JSON lines event logs. Defaults to stdout.\n"
+        "It's highly recommended to save the event logs to a file and preserve it for a period of time."
+    ),
+)
+@click.option(
+    "--concurrency",
+    type=click.IntRange(min=1),
+    default=_DEFAULT_CONCURRENCY,
+    show_default=True,
+    help="Number of tenant-level worker threads to run in parallel.",
+)
+def legacy_model_types(
+    apply: bool,
+    tables: tuple[str, ...],
+    model_types: tuple[str, ...],
+    tenant_id_file: str | None,
+    output: Path | None,
+    concurrency: int = _DEFAULT_CONCURRENCY,
+) -> None:
+    """
+    Migrate legacy provider-related model_type values and emit JSON lines events.
+    """
+
+    normalized_tables = _normalize_multi_value_option(
+        tables,
+        valid_values=VALID_TABLE_NAMES,
+        option_name="--tables",
+    )
+    normalized_model_types = _normalize_multi_value_option(
+        model_types,
+        valid_values=_SUPPORTED_MODEL_TYPE_CHOICES,
+        option_name="--model-types",
+    )
+    selected_model_types = (
+        tuple(ModelType.value_of(model_type) for model_type in normalized_model_types)
+        if normalized_model_types
+        else (
+            ModelType.LLM,
+            ModelType.TEXT_EMBEDDING,
+            ModelType.RERANK,
+        )
+    )
+    tenant_ids = load_tenant_ids_from_file(tenant_id_file) if tenant_id_file else None
+
+    output_context: AbstractContextManager[io.TextIOBase]
+    if output is None:
+        output_context = nullcontext(cast(io.TextIOBase, sys.stdout))
+    else:
+        try:
+            output_context = output.open("w", encoding="utf-8")
+        except OSError as exc:
+            raise click.ClickException(f"failed to open output file '{output}': {exc.strerror or exc}") from exc
+
+    with output_context as output_stream:
+        LegacyModelTypeMigrationService(
+            engine=db.engine,
+            apply=apply,
+            concurrency=concurrency,
+            output=cast(io.TextIOBase, output_stream),
+            tables=normalized_tables or None,
+            model_types=selected_model_types,
+            tenant_ids=tenant_ids,
+        ).migrate()
+
+
+data_migrate.add_command(legacy_model_types)

api/commands/data_migration.py

@@ -0,0 +1,754 @@
+from __future__ import annotations
+
+import json
+from datetime import datetime
+from pathlib import Path
+from typing import Any, cast
+from uuid import UUID
+
+import click
+import sqlalchemy as sa
+import yaml
+
+from extensions.ext_database import db
+from models import Tenant
+from models.model import App
+from models.tools import ApiToolProvider, MCPToolProvider, WorkflowToolProvider
+from services.app_dsl_service import AppDslService
+from services.data_migration.dependency_discovery_service import DependencyDiscoveryService
+from services.data_migration.entities import (
+    DependencyKind,
+    ImportOptions,
+    MigrationDataError,
+    ReportContext,
+    ResourceReportItem,
+)
+from services.data_migration.export_service import ExportConfigParser, MigrationExportService
+from services.data_migration.import_service import ImportRequest, MigrationImportService
+from services.data_migration.package_service import MigrationPackageService
+from services.data_migration.report_service import MigrationReportService
+
+ID_STRATEGY_CHOICES = ["preserve-id", "generate-new-id"]
+CONFLICT_STRATEGY_CHOICES = ["fail", "skip", "update"]
+SUPPORTED_WIZARD_APP_MODES = ["workflow", "advanced-chat"]
+WizardToolMap = dict[str, dict[str, str | None]]
+WizardToolSelection = dict[str, list[str]]
+
+
+def _scripted_export_template() -> dict[str, Any]:
+    return {
+        "source_tenant": {
+            "mode": "single",
+            "id": "",
+            "name": "admin's Workspace",
+        },
+        "apps": {
+            "modes": ["workflow", "advanced-chat"],
+            "ids": [],
+            "all": True,
+        },
+        "include_referenced_tools": True,
+        "additional_tools": {
+            "api_tools": [],
+            "workflow_tools": [],
+            "mcp_tools": [],
+        },
+        "include_secrets": False,
+        "import_options": {
+            "create_app_api_token_on_import": False,
+            "id_strategy": "preserve-id",
+            "conflict_strategy": "fail",
+        },
+    }
+
+
+@click.command("app-migration-template", help="Print or write a scripted export config JSON template.")
+@click.option(
+    "--output",
+    "output_file",
+    required=False,
+    type=click.Path(dir_okay=False),
+    help="Path to write the export config JSON template. Prints to stdout when omitted.",
+)
+@click.option("--overwrite", is_flag=True, default=False, help="Overwrite output if it already exists.")
+def export_migration_data_template(output_file: str | None, overwrite: bool) -> None:
+    template_json = json.dumps(_scripted_export_template(), indent=2, ensure_ascii=False) + "\n"
+    if output_file is None:
+        click.echo(template_json, nl=False)
+        return
+    path = Path(output_file)
+    if path.exists() and not overwrite:
+        raise click.ClickException(f"Output file already exists: {output_file}")
+    path.write_text(template_json)
+    click.echo(click.style(f"Output written to {output_file}", fg="green"))
+
+
+@click.command("export-app-migration", help="Export workflow migration data to a versioned JSON package.")
+@click.option(
+    "--input",
+    "input_file",
+    required=False,
+    type=click.Path(exists=True, dir_okay=False),
+    help="Path to export config JSON.",
+)
+@click.option(
+    "--output",
+    "output_file",
+    required=False,
+    type=click.Path(dir_okay=False),
+    help="Path to migration package JSON.",
+)
+@click.option("--overwrite", is_flag=True, default=False, help="Overwrite output if it already exists.")
+def export_migration_data(input_file: str | None, output_file: str | None, overwrite: bool) -> None:
+    try:
+        _require_options(("--input", input_file), ("--output", output_file))
+        assert input_file is not None
+        assert output_file is not None
+        raw_config = _load_json_object(input_file, "Export config")
+        selection = ExportConfigParser().parse(raw_config)
+        result = MigrationExportService().export(selection)
+        MigrationPackageService().save_package(result.package, output_file, overwrite=overwrite)
+        click.echo(click.style(f"Output written to {output_file}", fg="green"))
+        _render_report(result.report_items, context=_with_output_path(result.report_context, output_file))
+    except MigrationDataError as exc:
+        raise click.ClickException(str(exc)) from exc
+
+
+@click.command("import-app-migration", help="Import a versioned migration data package.")
+@click.option(
+    "--input",
+    "input_file",
+    required=False,
+    type=click.Path(exists=True, dir_okay=False),
+    help="Path to migration package JSON.",
+)
+@click.option("--target-tenant", default=None, help="Target tenant/workspace name. Overrides package metadata.")
+@click.option("--operator-email", default=None, help="Operator account email in the target tenant.")
+@click.option(
+    "--id-strategy",
+    default=None,
+    type=click.Choice(ID_STRATEGY_CHOICES),
+    help="Override package ID strategy.",
+)
+@click.option(
+    "--conflict-strategy",
+    default=None,
+    type=click.Choice(CONFLICT_STRATEGY_CHOICES),
+    help="Override package conflict strategy.",
+)
+@click.option(
+    "--create-app-api-token-on-import/--no-create-app-api-token-on-import",
+    default=None,
+    help="Override package app API token creation behavior.",
+)
+def import_migration_data(
+    input_file: str | None,
+    target_tenant: str | None,
+    operator_email: str | None,
+    id_strategy: str | None,
+    conflict_strategy: str | None,
+    create_app_api_token_on_import: bool | None,
+) -> None:
+    try:
+        _require_options(("--input", input_file))
+        assert input_file is not None
+        package = MigrationPackageService().load_package(input_file)
+        result = MigrationImportService().import_package(
+            ImportRequest(
+                package=package,
+                cli_target_tenant=target_tenant,
+                operator_email=operator_email,
+                options_override=_build_options_override(
+                    package.metadata.import_options,
+                    id_strategy=id_strategy,
+                    conflict_strategy=conflict_strategy,
+                    create_app_api_token_on_import=create_app_api_token_on_import,
+                ),
+            )
+        )
+        _render_report(result.report_items, context=result.report_context)
+    except MigrationDataError as exc:
+        raise click.ClickException(str(exc)) from exc
+
+
+def parse_index_selection(raw: str, values: list[str]) -> list[str]:
+    normalized = raw.strip().lower()
+    if normalized == "all":
+        return values
+
+    selected: list[str] = []
+    for part in raw.split(","):
+        stripped = part.strip()
+        if not stripped:
+            continue
+        try:
+            index = int(stripped)
+        except ValueError as exc:
+            raise click.ClickException(f"Selection must be 'all' or comma-separated numbers: {raw}") from exc
+        if index < 1 or index > len(values):
+            raise click.ClickException(f"Selection index out of range: {index}")
+        selected.append(values[index - 1])
+    return list(dict.fromkeys(selected))
+
+
+def _print_wizard_step(title: str) -> None:
+    click.echo("")
+    click.echo(f"==== {title} ====")
+
+
+def _print_wizard_substep(title: str) -> None:
+    click.echo("")
+    click.echo(f"-- {title} --")
+
+
+@click.command("app-migration-wizard", help="Interactively export workflow migration data.")
+def migration_data_wizard() -> None:
+    try:
+        tenant = _prompt_source_tenant()
+        apps = _eligible_apps_for_tenant(tenant.id)
+        app_ids = _prompt_app_ids(apps)
+        _print_wizard_step("Referenced Tools")
+        include_referenced_tools = click.confirm(
+            "Automatically export tools referenced by selected apps? [y/n, default: y]",
+            default=True,
+            show_default=False,
+        )
+        auto_tools = _discover_auto_tools([app for app in apps if app.id in set(app_ids)], include_referenced_tools)
+        auto_tools = _resolve_auto_tool_names(tenant.id, auto_tools)
+        _print_auto_tools(auto_tools)
+        additional_tools = _prompt_additional_tools(tenant.id, auto_tools)
+        include_secrets, create_tokens, id_strategy, conflict_strategy = _prompt_import_options()
+        _print_wizard_step("Output")
+        output_file, overwrite = _prompt_output_file()
+
+        selection = ExportConfigParser().parse(
+            {
+                "source_tenant": {"mode": "single", "id": tenant.id, "name": tenant.name},
+                "apps": {"ids": app_ids, "all": False},
+                "include_referenced_tools": include_referenced_tools,
+                "additional_tools": additional_tools,
+                "include_secrets": include_secrets,
+                "import_options": {
+                    "create_app_api_token_on_import": create_tokens,
+                    "id_strategy": id_strategy,
+                    "conflict_strategy": conflict_strategy,
+                },
+            }
+        )
+        _confirm_wizard_summary(
+            tenant_name=tenant.name,
+            app_names=[app.name for app in apps if app.id in set(app_ids)],
+            auto_tools=auto_tools,
+            additional_tools=additional_tools,
+            manual_labels=_selected_tool_labels_for_tenant(tenant.id, additional_tools),
+            include_referenced_tools=include_referenced_tools,
+            include_secrets=include_secrets,
+            create_tokens=create_tokens,
+            id_strategy=id_strategy,
+            conflict_strategy=conflict_strategy,
+            output_file=output_file,
+        )
+        result = MigrationExportService().export(selection)
+        MigrationPackageService().save_package(result.package, output_file, overwrite=overwrite)
+        click.echo(click.style(f"Output written to {output_file}", fg="green"))
+        _print_wizard_step("Report")
+        _render_report(result.report_items, context=_with_output_path(result.report_context, output_file))
+    except MigrationDataError as exc:
+        raise click.ClickException(str(exc)) from exc
+
+
+def _load_json_object(path: str, label: str) -> dict[str, Any]:
+    try:
+        with Path(path).open(encoding="utf-8") as file:
+            raw = json.load(file)
+    except json.JSONDecodeError as exc:
+        raise MigrationDataError(f"{label} JSON is invalid: {exc.msg}") from exc
+    if not isinstance(raw, dict):
+        raise MigrationDataError(f"{label} JSON must be an object.")
+    return raw
+
+
+def _require_options(*options: tuple[str, object | None]) -> None:
+    missing_options = [name for name, value in options if value is None]
+    if missing_options:
+        raise click.UsageError(f"Missing option(s): {', '.join(missing_options)}.")
+
+
+def _build_options_override(
+    package_options: ImportOptions,
+    *,
+    id_strategy: str | None,
+    conflict_strategy: str | None,
+    create_app_api_token_on_import: bool | None,
+) -> ImportOptions | None:
+    if id_strategy is None and conflict_strategy is None and create_app_api_token_on_import is None:
+        return None
+    return ImportOptions.from_mapping(
+        {
+            "id_strategy": id_strategy or package_options.id_strategy,
+            "conflict_strategy": conflict_strategy or package_options.conflict_strategy,
+            "create_app_api_token_on_import": (
+                create_app_api_token_on_import
+                if create_app_api_token_on_import is not None
+                else package_options.create_app_api_token_on_import
+            ),
+        }
+    )
+
+
+def _prompt_source_tenant() -> Tenant:
+    tenants = list(db.session.scalars(sa.select(Tenant).order_by(Tenant.name.asc())).all())
+    if not tenants:
+        raise MigrationDataError("No tenants found.")
+
+    _print_wizard_step("Source Tenant")
+    click.echo("Source tenants:")
+    for index, tenant in enumerate(tenants, 1):
+        click.echo(f"{index}. {tenant.name} ({tenant.id})")
+
+    tenant_index = click.prompt("Select one source tenant by number", type=int, default=1, show_default=True)
+    if tenant_index < 1 or tenant_index > len(tenants):
+        raise click.ClickException(f"Selection index out of range: {tenant_index}")
+    return tenants[tenant_index - 1]
+
+
+def _eligible_apps_for_tenant(tenant_id: str) -> list[App]:
+    return list(
+        db.session.scalars(
+            sa.select(App)
+            .where(App.tenant_id == tenant_id, App.mode.in_(SUPPORTED_WIZARD_APP_MODES))
+            .order_by(App.name.asc())
+        ).all()
+    )
+
+
+def _prompt_app_ids(apps: list[App]) -> list[str]:
+    if not apps:
+        raise MigrationDataError("No workflow or advanced-chat apps found for the selected tenant.")
+
+    _print_wizard_step("App Selection")
+    click.echo("Currently supported app types: workflow and chatflow.")
+    click.echo("Workflow/chatflow apps:")
+    for index, app in enumerate(apps, 1):
+        mode = app.mode.value if hasattr(app.mode, "value") else app.mode
+        click.echo(f"{index}. {app.name} [{mode}] ({app.id})")
+    app_ids = parse_index_selection(
+        click.prompt("Select apps by number, comma-separated numbers, or all", default="all"),
+        [app.id for app in apps],
+    )
+    selected_apps = [app for app in apps if app.id in set(app_ids)]
+    click.echo("Selected apps:")
+    for app in selected_apps:
+        click.echo(f"- {app.name} ({app.id})")
+    return app_ids
+
+
+def _prompt_import_options() -> tuple[bool, bool, str, str]:
+    _print_wizard_step("Import Options")
+    _print_wizard_substep("Secrets")
+    click.echo("Secrets include workflow/app DSL secret values, custom API tool credentials,")
+    click.echo("and full MCP provider connection data such as server URL, headers, authentication, and tool list.")
+    click.echo("If you choose no, credentials are omitted or masked,")
+    click.echo("and MCP providers are exported as dependency metadata only.")
+    click.echo("Treat the output JSON as sensitive if you choose yes.")
+    include_secrets = click.confirm(
+        "Include secrets in output JSON? [y/n, default: n]",
+        default=False,
+        show_default=False,
+    )
+    _print_wizard_substep("App API Tokens")
+    click.echo("When enabled, import will create an app API token if the imported app has none,")
+    click.echo("or reuse an existing app API token if one already exists.")
+    create_tokens = click.confirm(
+        "Create or reuse app API tokens during import? [y/n, default: n]",
+        default=False,
+        show_default=False,
+    )
+    _print_wizard_substep("ID Strategy")
+    click.echo("ID strategy controls whether imported app and tool IDs preserve source IDs")
+    click.echo("or use target-generated IDs.")
+    click.echo("preserve-id: keep source IDs where the target service supports it.")
+    click.echo("generate-new-id: let the target environment generate new IDs and rewrite references via mapping.")
+    id_strategy = click.prompt(
+        "Import ID strategy. Enter one of: preserve-id, generate-new-id",
+        type=click.Choice(ID_STRATEGY_CHOICES),
+        default="preserve-id",
+        show_default=True,
+    )
+    _print_wizard_substep("Conflict Strategy")
+    click.echo("Conflict strategy controls what import does when a target resource already exists.")
+    click.echo("fail: stop at the first conflict; previously committed resources are not rolled back.")
+    click.echo("skip: keep the existing target resource and skip importing that resource.")
+    click.echo("update: update the existing target resource in place.")
+    conflict_strategy = click.prompt(
+        "Import conflict strategy. Enter one of: fail, skip, update",
+        type=click.Choice(CONFLICT_STRATEGY_CHOICES),
+        default="update",
+        show_default=True,
+    )
+    return include_secrets, create_tokens, id_strategy, conflict_strategy
+
+
+def _discover_auto_tools(apps: list[App], include_referenced_tools: bool) -> WizardToolMap:
+    auto_tools: WizardToolMap = {"api_tools": {}, "workflow_tools": {}, "mcp_tools": {}}
+    if not include_referenced_tools:
+        return auto_tools
+    discovery_service = DependencyDiscoveryService()
+    for app in apps:
+        dsl_content = AppDslService.export_dsl(app_model=app, include_secret=False)
+        raw_dsl = yaml.safe_load(dsl_content) if dsl_content else {}
+        dsl = raw_dsl if isinstance(raw_dsl, dict) else {}
+        for dependency in discovery_service.discover_from_dsl(dsl):
+            if dependency.kind == DependencyKind.API_TOOL:
+                auto_tools["api_tools"][dependency.provider_name or dependency.provider_id] = dependency.provider_id
+            elif dependency.kind == DependencyKind.WORKFLOW_TOOL:
+                auto_tools["workflow_tools"][dependency.provider_name or dependency.provider_id] = (
+                    dependency.provider_id
+                )
+            elif dependency.kind == DependencyKind.MCP_TOOL:
+                auto_tools["mcp_tools"][dependency.provider_name or dependency.provider_id] = dependency.provider_id
+    return auto_tools
+
+
+def _resolve_auto_tool_names(tenant_id: str, auto_tools: WizardToolMap) -> WizardToolMap:
+    return {
+        "api_tools": _resolve_api_tool_names(tenant_id, auto_tools["api_tools"]),
+        "workflow_tools": _resolve_workflow_tool_names(tenant_id, auto_tools["workflow_tools"]),
+        "mcp_tools": _resolve_mcp_tool_names(tenant_id, auto_tools["mcp_tools"]),
+    }
+
+
+def _resolve_api_tool_names(tenant_id: str, tools: dict[str, str | None]) -> dict[str, str | None]:
+    resolved: dict[str, str | None] = {}
+    for name, identifier in tools.items():
+        predicates = [ApiToolProvider.name == name]
+        if _is_uuid_string(identifier):
+            predicates.append(ApiToolProvider.id == identifier)
+        provider = db.session.scalar(
+            sa.select(ApiToolProvider).where(
+                ApiToolProvider.tenant_id == tenant_id,
+                sa.or_(*predicates),
+            )
+        )
+        resolved[provider.name if provider else name] = provider.id if provider else identifier
+    return resolved
+
+
+def _resolve_workflow_tool_names(tenant_id: str, tools: dict[str, str | None]) -> dict[str, str | None]:
+    resolved: dict[str, str | None] = {}
+    for name, identifier in tools.items():
+        predicates = [WorkflowToolProvider.name == name]
+        if _is_uuid_string(identifier):
+            predicates.append(WorkflowToolProvider.id == identifier)
+        provider = db.session.scalar(
+            sa.select(WorkflowToolProvider).where(
+                WorkflowToolProvider.tenant_id == tenant_id,
+                sa.or_(*predicates),
+            )
+        )
+        resolved[provider.name if provider else name] = provider.id if provider else identifier
+    return resolved
+
+
+def _resolve_mcp_tool_names(tenant_id: str, tools: dict[str, str | None]) -> dict[str, str | None]:
+    resolved: dict[str, str | None] = {}
+    for name, identifier in tools.items():
+        predicates = [MCPToolProvider.name == name]
+        if identifier:
+            predicates.append(MCPToolProvider.server_identifier == identifier)
+        if _is_uuid_string(identifier):
+            predicates.append(MCPToolProvider.id == identifier)
+        provider = db.session.scalar(
+            sa.select(MCPToolProvider).where(
+                MCPToolProvider.tenant_id == tenant_id,
+                sa.or_(*predicates),
+            )
+        )
+        resolved[provider.name if provider else name] = provider.id if provider else identifier
+    return resolved
+
+
+def _is_uuid_string(value: str | None) -> bool:
+    if not value:
+        return False
+    try:
+        UUID(value)
+    except ValueError:
+        return False
+    return True
+
+
+def _print_auto_tools(auto_tools: WizardToolMap) -> None:
+    _print_wizard_step("Automatically Discovered Tools")
+    click.echo("Automatically discovered tools:")
+    _print_auto_tool_category("Custom API tools", auto_tools["api_tools"])
+    _print_auto_tool_category("Workflow tools", auto_tools["workflow_tools"])
+    _print_auto_tool_category("MCP tools", auto_tools["mcp_tools"])
+
+
+def _print_auto_tool_category(label: str, values: dict[str, str | None]) -> None:
+    click.echo(label)
+    if not values:
+        click.echo("- none")
+        return
+    for name, identifier in sorted(values.items()):
+        click.echo(f"- {_format_tool_name_id(name, identifier)}")
+
+
+def _prompt_additional_tools(tenant_id: str, auto_tools: WizardToolMap) -> WizardToolSelection:
+    selections: WizardToolSelection = {"api_tools": [], "workflow_tools": [], "mcp_tools": []}
+    _print_wizard_step("Additional Tools")
+    if not click.confirm(
+        "Export additional tools manually? [y/n, default: n]",
+        default=False,
+        show_default=False,
+    ):
+        _print_final_tool_selection(auto_tools, selections, {})
+        return selections
+    manual_labels: dict[str, str] = {}
+    api_tool_options = [
+        (tool.name, tool.name, tool.id)
+        for tool in db.session.scalars(
+            sa.select(ApiToolProvider).where(ApiToolProvider.tenant_id == tenant_id).order_by(ApiToolProvider.name)
+        ).all()
+    ]
+    selections["api_tools"] = _prompt_tool_category(
+        "Custom API tools",
+        api_tool_options,
+        auto_tools=auto_tools["api_tools"],
+    )
+    manual_labels.update(_selected_tool_labels(api_tool_options, selections["api_tools"]))
+    workflow_tool_options = [
+        (tool.id, tool.name, tool.id)
+        for tool in db.session.scalars(
+            sa.select(WorkflowToolProvider)
+            .where(WorkflowToolProvider.tenant_id == tenant_id)
+            .order_by(WorkflowToolProvider.name)
+        ).all()
+    ]
+    selections["workflow_tools"] = _prompt_tool_category(
+        "Workflow tools",
+        workflow_tool_options,
+        auto_tools=auto_tools["workflow_tools"],
+    )
+    manual_labels.update(_selected_tool_labels(workflow_tool_options, selections["workflow_tools"]))
+    mcp_tool_options = [
+        (tool.id, tool.name, tool.server_identifier)
+        for tool in db.session.scalars(
+            sa.select(MCPToolProvider).where(MCPToolProvider.tenant_id == tenant_id).order_by(MCPToolProvider.name)
+        ).all()
+    ]
+    selections["mcp_tools"] = _prompt_tool_category(
+        "MCP tools",
+        mcp_tool_options,
+        auto_tools=auto_tools["mcp_tools"],
+    )
+    manual_labels.update(_selected_tool_labels(mcp_tool_options, selections["mcp_tools"]))
+    _print_final_tool_selection(auto_tools, selections, manual_labels)
+    return selections
+
+
+def _selected_tool_labels_for_tenant(tenant_id: str, selected_tools: WizardToolSelection) -> dict[str, str]:
+    labels: dict[str, str] = {}
+    if selected_tools["api_tools"]:
+        labels.update(
+            _selected_tool_labels(
+                [
+                    (tool.name, tool.name, tool.id)
+                    for tool in db.session.scalars(
+                        sa.select(ApiToolProvider)
+                        .where(ApiToolProvider.tenant_id == tenant_id)
+                        .order_by(ApiToolProvider.name)
+                    ).all()
+                ],
+                selected_tools["api_tools"],
+            )
+        )
+    if selected_tools["workflow_tools"]:
+        labels.update(
+            _selected_tool_labels(
+                [
+                    (tool.id, tool.name, tool.id)
+                    for tool in db.session.scalars(
+                        sa.select(WorkflowToolProvider)
+                        .where(WorkflowToolProvider.tenant_id == tenant_id)
+                        .order_by(WorkflowToolProvider.name)
+                    ).all()
+                ],
+                selected_tools["workflow_tools"],
+            )
+        )
+    if selected_tools["mcp_tools"]:
+        labels.update(
+            _selected_tool_labels(
+                [
+                    (tool.id, tool.name, tool.server_identifier)
+                    for tool in db.session.scalars(
+                        sa.select(MCPToolProvider)
+                        .where(MCPToolProvider.tenant_id == tenant_id)
+                        .order_by(MCPToolProvider.name)
+                    ).all()
+                ],
+                selected_tools["mcp_tools"],
+            )
+        )
+    return labels
+
+
+def _selected_tool_labels(options: list[tuple[str, str, str]], selected_values: list[str]) -> dict[str, str]:
+    selected = set(selected_values)
+    return {value: _format_tool_name_id(name, detail) for value, name, detail in options if value in selected}
+
+
+def _prompt_tool_category(
+    label: str,
+    options: list[tuple[str, str, str]],
+    *,
+    auto_tools: dict[str, str | None],
+) -> list[str]:
+    if not options:
+        click.echo(f"{label}: none")
+        return []
+    _print_wizard_step(label)
+    for index, (value, name, detail) in enumerate(options, 1):
+        marker = "[auto]" if _is_auto_tool(value, name, detail, auto_tools) else "[ ]"
+        click.echo(f"{index}. {marker} {name} ({detail})")
+    raw = click.prompt(
+        f"Select {label.lower()} by number, comma-separated numbers, all, or empty",
+        default="",
+        show_default=cast(Any, "empty"),
+    )
+    if not raw.strip():
+        return []
+    return parse_index_selection(raw, [value for value, _, _ in options])
+
+
+def _is_auto_tool(value: str, name: str, detail: str, auto_tools: dict[str, str | None]) -> bool:
+    return name in auto_tools or value in auto_tools or value in auto_tools.values() or detail in auto_tools.values()
+
+
+def _print_final_tool_selection(
+    auto_tools: WizardToolMap,
+    additional_tools: WizardToolSelection,
+    manual_labels: dict[str, str],
+) -> None:
+    _print_wizard_step("Final Tool Selection")
+    _print_tool_selection_body(auto_tools, additional_tools, manual_labels)
+
+
+def _print_tool_selection_body(
+    auto_tools: WizardToolMap,
+    additional_tools: WizardToolSelection,
+    manual_labels: dict[str, str],
+) -> None:
+    click.echo("Final tools to export:")
+    _print_final_tool_category(
+        "Custom API tools",
+        auto_tools["api_tools"],
+        additional_tools["api_tools"],
+        manual_labels,
+    )
+    _print_final_tool_category(
+        "Workflow tools",
+        auto_tools["workflow_tools"],
+        additional_tools["workflow_tools"],
+        manual_labels,
+    )
+    _print_final_tool_category("MCP tools", auto_tools["mcp_tools"], additional_tools["mcp_tools"], manual_labels)
+
+
+def _print_final_tool_category(
+    label: str,
+    auto_tools: dict[str, str | None],
+    manual_values: list[str],
+    manual_labels: dict[str, str],
+) -> None:
+    click.echo(label)
+    lines = [f"- [auto] {_format_tool_name_id(name, identifier)}" for name, identifier in sorted(auto_tools.items())]
+    auto_identifiers = {identifier for identifier in auto_tools.values() if identifier}
+    lines.extend(
+        f"- [manual] {manual_labels.get(value, value)}"
+        for value in manual_values
+        if value not in auto_tools and value not in auto_identifiers
+    )
+    if not lines:
+        click.echo("- none")
+        return
+    for line in lines:
+        click.echo(line)
+
+
+def _format_tool_name_id(name: str, identifier: str | None) -> str:
+    if identifier and identifier != name:
+        return f"{name}: {identifier}"
+    return name
+
+
+def _confirm_wizard_summary(
+    *,
+    tenant_name: str,
+    app_names: list[str],
+    auto_tools: WizardToolMap,
+    additional_tools: WizardToolSelection,
+    manual_labels: dict[str, str],
+    include_referenced_tools: bool,
+    include_secrets: bool,
+    create_tokens: bool,
+    id_strategy: str,
+    conflict_strategy: str,
+    output_file: str,
+) -> None:
+    _print_wizard_step("Summary")
+    click.echo("Migration export summary:")
+    click.echo(f"source tenant: {tenant_name}")
+    click.echo(f"selected apps: {len(app_names)}")
+    for app_name in app_names:
+        click.echo(f"- {app_name}")
+    click.echo(f"auto referenced tools: {str(include_referenced_tools).lower()}")
+    _print_tool_selection_body(auto_tools, additional_tools, manual_labels)
+    click.echo(f"include secrets: {str(include_secrets).lower()}")
+    click.echo(f"create app api token on import: {str(create_tokens).lower()}")
+    click.echo(f"id strategy: {id_strategy}")
+    click.echo(f"conflict strategy: {conflict_strategy}")
+    click.echo(f"output path: {output_file}")
+    if not click.confirm("Write migration package? [y/n, default: y]", default=True, show_default=False):
+        raise click.Abort()
+
+
+def _prompt_output_file() -> tuple[str, bool]:
+    default_output = f"migration-data-{datetime.now().strftime('%Y%m%d-%H%M%S')}.json"
+    output_file = click.prompt("Output path", default=default_output, show_default=True)
+    if output_file.lower() in {"y", "yes", "n", "no"}:
+        raise click.ClickException("Output path must be a file path. Press Enter to use the default path.")
+    overwrite = False
+    if Path(output_file).exists():
+        overwrite = click.confirm(
+            "Output file exists. Overwrite? [y/n, default: n]",
+            default=False,
+            show_default=False,
+        )
+        if not overwrite:
+            raise click.ClickException(f"Output file already exists: {output_file}")
+    return output_file, overwrite
+
+
+def _with_output_path(context: ReportContext | None, output_path: str) -> ReportContext:
+    if context is None:
+        return ReportContext(output_path=output_path)
+    return ReportContext(
+        output_path=output_path,
+        source_scope=context.source_scope,
+        selected_app_count=context.selected_app_count,
+        include_secrets=context.include_secrets,
+        target_tenant=context.target_tenant,
+        operator_email=context.operator_email,
+        app_api_tokens_created=context.app_api_tokens_created,
+        app_api_tokens_reused=context.app_api_tokens_reused,
+        id_mapping_count=context.id_mapping_count,
+        id_mappings=context.id_mappings,
+    )
+
+
+def _render_report(report_items: list[ResourceReportItem], *, context: ReportContext | None = None) -> None:
+    for line in MigrationReportService().render(report_items, context=context):
+        click.echo(line)

api/commands/plugin.py

@@ -1,10 +1,11 @@
 import json
 import logging
+import time
 from typing import Any, cast
 
 import click
 from pydantic import TypeAdapter
-from sqlalchemy import delete, select
+from sqlalchemy import delete, func, select
 from sqlalchemy.engine import CursorResult
 
 from configs import dify_config
@@ -15,11 +16,13 @@ from core.plugin.plugin_service import PluginService
 from core.tools.utils.system_encryption import encrypt_system_params
 from extensions.ext_database import db
 from models import Tenant
+from models.account import TenantPluginAutoUpgradeStrategy
 from models.oauth import DatasourceOauthParamConfig, DatasourceProvider
 from models.provider_ids import DatasourceProviderID, ToolProviderID
 from models.source import DataSourceApiKeyAuthBinding, DataSourceOauthBinding
 from models.tools import ToolOAuthSystemClient
 from services.plugin.data_migration import PluginDataMigration
+from services.plugin.plugin_auto_upgrade_service import PluginAutoUpgradeService
 from services.plugin.plugin_migration import PluginMigration
 
 logger = logging.getLogger(__name__)
@@ -402,6 +405,110 @@ def migrate_data_for_plugin():
     click.echo(click.style("Migrate data for plugin completed.", fg="green"))
 
 
+def _candidate_auto_upgrade_strategy_tenant_ids_stmt(limit: int | None = None):
+    category_count = len(TenantPluginAutoUpgradeStrategy.PluginCategory)
+    stmt = (
+        select(TenantPluginAutoUpgradeStrategy.tenant_id)
+        .group_by(TenantPluginAutoUpgradeStrategy.tenant_id)
+        .having(func.count(func.distinct(TenantPluginAutoUpgradeStrategy.category)) < category_count)
+        .order_by(TenantPluginAutoUpgradeStrategy.tenant_id)
+    )
+
+    if limit is not None:
+        stmt = stmt.limit(limit)
+
+    return stmt
+
+
+def _count_auto_upgrade_strategy_tenant_ids(limit: int | None) -> int:
+    candidate_stmt = _candidate_auto_upgrade_strategy_tenant_ids_stmt(limit).subquery()
+    return db.session.scalar(select(func.count()).select_from(candidate_stmt)) or 0
+
+
+def _iter_auto_upgrade_strategy_tenant_ids(limit: int | None):
+    stmt = _candidate_auto_upgrade_strategy_tenant_ids_stmt(limit).execution_options(yield_per=1000)
+    yield from db.session.scalars(stmt)
+
+
+@click.command(
+    "backfill-plugin-auto-upgrade",
+    help="Backfill category-scoped plugin auto-upgrade strategies and normalize plugin lists.",
+)
+@click.option("--tenant-id", multiple=True, help="Tenant ID to backfill. Can be passed multiple times.")
+@click.option("--limit", type=int, default=None, help="Maximum number of candidate tenants to process.")
+@click.option("--batch-size", type=int, default=500, show_default=True, help="Progress reporting batch size.")
+@click.option("--dry-run", is_flag=True, help="Only print candidate tenant count.")
+def backfill_plugin_auto_upgrade(
+    tenant_id: tuple[str, ...],
+    limit: int | None,
+    batch_size: int,
+    dry_run: bool,
+):
+    """
+    Backfill historical auto-upgrade strategies after the category column exists.
+
+    Missing category rows are created from the tenant's tool/default row. Pure default
+    strategies become latest for model plugins and fix-only for all other categories.
+    Tenants with include/exclude plugin IDs are split
+    by installed plugin category using plugin daemon metadata.
+    """
+    start_at = time.perf_counter()
+    candidate_count = len(tenant_id) if tenant_id else _count_auto_upgrade_strategy_tenant_ids(limit)
+    click.echo(click.style(f"Found {candidate_count} candidate tenants.", fg="yellow"))
+
+    if dry_run:
+        elapsed = time.perf_counter() - start_at
+        click.echo(click.style(f"Dry run completed. elapsed={elapsed:.2f}s", fg="green"))
+        return
+
+    tenant_ids = list(tenant_id) if tenant_id else _iter_auto_upgrade_strategy_tenant_ids(limit)
+
+    backfilled_count = 0
+    created_count = 0
+    normalized_count = 0
+    skipped_count = 0
+    failed_count = 0
+    for index, current_tenant_id in enumerate(tenant_ids, start=1):
+        try:
+            result = PluginAutoUpgradeService.backfill_strategy_categories(
+                current_tenant_id,
+            )
+        except Exception as e:
+            failed_count += 1
+            click.echo(click.style(f"Failed tenant {current_tenant_id}: {str(e)}", fg="red"))
+            continue
+
+        if result.created_count > 0:
+            backfilled_count += 1
+            created_count += result.created_count
+        elif not result.normalized:
+            skipped_count += 1
+        if result.normalized:
+            normalized_count += 1
+
+        if batch_size > 0 and index % batch_size == 0:
+            click.echo(
+                click.style(
+                    f"Processed {index}/{candidate_count} tenants. "
+                    f"backfilled={backfilled_count}, created_rows={created_count}, "
+                    f"normalized={normalized_count}, skipped={skipped_count}, failed={failed_count}, "
+                    f"elapsed={time.perf_counter() - start_at:.2f}s",
+                    fg="yellow",
+                )
+            )
+
+    elapsed = time.perf_counter() - start_at
+    click.echo(
+        click.style(
+            f"Backfill plugin auto-upgrade strategy categories completed. "
+            f"backfilled={backfilled_count}, created_rows={created_count}, "
+            f"normalized={normalized_count}, skipped={skipped_count}, failed={failed_count}, "
+            f"elapsed={elapsed:.2f}s",
+            fg="green",
+        )
+    )
+
+
 @click.command("extract-plugins", help="Extract plugins.")
 @click.option("--output_file", prompt=True, help="The file to store the extracted plugins.", default="plugins.jsonl")
 @click.option("--workers", prompt=True, help="The number of workers to extract plugins.", default=10)

api/commands/vector.py

@@ -30,7 +30,7 @@ def vdb_migrate(scope: str):
 
 def migrate_annotation_vector_database():
     """
-    Migrate annotation datas to target vector database .
+    Migrate annotation data to target vector database.
     """
     click.echo(click.style("Starting annotation data migration.", fg="green"))
     create_count = 0
@@ -140,7 +140,7 @@ def migrate_annotation_vector_database():
 
 def migrate_knowledge_vector_database():
     """
-    Migrate vector database datas to target vector database .
+    Migrate vector database data to target vector database.
     """
     click.echo(click.style("Starting vector database migration.", fg="green"))
     create_count = 0

api/configs/middleware/vdb/milvus_config.py

@@ -41,3 +41,21 @@ class MilvusConfig(BaseSettings):
         description='Milvus text analyzer parameters, e.g., {"type": "chinese"} for Chinese segmentation support.',
         default=None,
     )
+
+    MILVUS_SECURE: bool = Field(
+        description="Enable TLS for the Milvus connection (one-way TLS). When True, the client uses gRPC over TLS "
+        "and verifies the server certificate. Equivalent to passing secure=True to pymilvus.",
+        default=False,
+    )
+
+    MILVUS_SERVER_PEM_PATH: str | None = Field(
+        description="Filesystem path inside the container to the Milvus server certificate (PEM). Mount this via "
+        "a Kubernetes secret. Used as pymilvus's server_pem_path when MILVUS_SECURE is True.",
+        default=None,
+    )
+
+    MILVUS_SERVER_NAME: str | None = Field(
+        description="Server name (TLS SNI / certificate CN or SAN) to verify against the Milvus server certificate. "
+        "Required when MILVUS_SERVER_PEM_PATH is set.",
+        default=None,
+    )

api/controllers/console/__init__.py

@@ -68,6 +68,7 @@ from .app import (
     workflow_app_log,
     workflow_comment,
     workflow_draft_variable,
+    workflow_node_output_inspector,
     workflow_run,
     workflow_statistic,
     workflow_trigger,
@@ -218,6 +219,7 @@ __all__ = [
     "workflow_app_log",
     "workflow_comment",
     "workflow_draft_variable",
+    "workflow_node_output_inspector",
     "workflow_run",
     "workflow_statistic",
     "workflow_trigger",

api/controllers/console/agent/composer.py

@@ -1,153 +1,220 @@
 from flask_restx import Resource
 
-from controllers.common.schema import register_schema_models
+from controllers.common.schema import register_response_schema_models, register_schema_models
 from controllers.console import console_ns
 from controllers.console.app.wraps import get_app_model
 from controllers.console.wraps import account_initialization_required, edit_permission_required, setup_required
+from fields.agent_fields import (
+    AgentAppComposerResponse,
+    AgentComposerCandidatesResponse,
+    AgentComposerImpactResponse,
+    AgentComposerValidateResponse,
+    WorkflowAgentComposerResponse,
+)
+from libs.helper import dump_response
 from libs.login import current_account_with_tenant, login_required
-from models.model import AppMode
+from models.model import App, AppMode
 from services.agent.composer_service import AgentComposerService
 from services.agent.composer_validator import ComposerConfigValidator
 from services.entities.agent_entities import ComposerSavePayload
 
 register_schema_models(console_ns, ComposerSavePayload)
+register_response_schema_models(
+    console_ns,
+    AgentAppComposerResponse,
+    AgentComposerCandidatesResponse,
+    AgentComposerImpactResponse,
+    AgentComposerValidateResponse,
+    WorkflowAgentComposerResponse,
+)
 
 
 @console_ns.route("/apps/<uuid:app_id>/workflows/draft/nodes/<string:node_id>/agent-composer")
 class WorkflowAgentComposerApi(Resource):
+    @console_ns.response(
+        200, "Workflow agent composer state", console_ns.models[WorkflowAgentComposerResponse.__name__]
+    )
     @setup_required
     @login_required
     @account_initialization_required
     @get_app_model(mode=[AppMode.WORKFLOW, AppMode.ADVANCED_CHAT])
-    def get(self, app_model, node_id: str):
+    def get(self, app_model: App, node_id: str):
         _, tenant_id = current_account_with_tenant()
-        return AgentComposerService.load_workflow_composer(
-            tenant_id=tenant_id,
-            app_id=app_model.id,
-            node_id=node_id,
+        return dump_response(
+            WorkflowAgentComposerResponse,
+            AgentComposerService.load_workflow_composer(
+                tenant_id=tenant_id,
+                app_id=app_model.id,
+                node_id=node_id,
+            ),
         )
 
     @console_ns.expect(console_ns.models[ComposerSavePayload.__name__])
+    @console_ns.response(
+        200, "Workflow agent composer saved", console_ns.models[WorkflowAgentComposerResponse.__name__]
+    )
     @setup_required
     @login_required
     @account_initialization_required
     @edit_permission_required
     @get_app_model(mode=[AppMode.WORKFLOW, AppMode.ADVANCED_CHAT])
-    def put(self, app_model, node_id: str):
+    def put(self, app_model: App, node_id: str):
         account, tenant_id = current_account_with_tenant()
         payload = ComposerSavePayload.model_validate(console_ns.payload or {})
-        return AgentComposerService.save_workflow_composer(
-            tenant_id=tenant_id,
-            app_id=app_model.id,
-            node_id=node_id,
-            account_id=account.id,
-            payload=payload,
+        return dump_response(
+            WorkflowAgentComposerResponse,
+            AgentComposerService.save_workflow_composer(
+                tenant_id=tenant_id,
+                app_id=app_model.id,
+                node_id=node_id,
+                account_id=account.id,
+                payload=payload,
+            ),
         )
 
 
 @console_ns.route("/apps/<uuid:app_id>/workflows/draft/nodes/<string:node_id>/agent-composer/validate")
 class WorkflowAgentComposerValidateApi(Resource):
     @console_ns.expect(console_ns.models[ComposerSavePayload.__name__])
+    @console_ns.response(
+        200, "Workflow agent composer validation result", console_ns.models[AgentComposerValidateResponse.__name__]
+    )
     @setup_required
     @login_required
     @account_initialization_required
     @get_app_model(mode=[AppMode.WORKFLOW, AppMode.ADVANCED_CHAT])
-    def post(self, app_model, node_id: str):
+    def post(self, app_model: App, node_id: str):
         payload = ComposerSavePayload.model_validate(console_ns.payload or {})
         ComposerConfigValidator.validate_save_payload(payload)
-        return {"result": "success", "errors": []}
+        return dump_response(AgentComposerValidateResponse, {"result": "success", "errors": []})
 
 
 @console_ns.route("/apps/<uuid:app_id>/workflows/draft/nodes/<string:node_id>/agent-composer/candidates")
 class WorkflowAgentComposerCandidatesApi(Resource):
+    @console_ns.response(
+        200, "Workflow agent composer candidates", console_ns.models[AgentComposerCandidatesResponse.__name__]
+    )
     @setup_required
     @login_required
     @account_initialization_required
     @get_app_model(mode=[AppMode.WORKFLOW, AppMode.ADVANCED_CHAT])
-    def get(self, app_model, node_id: str):
-        return AgentComposerService.get_workflow_candidates(app_id=app_model.id)
+    def get(self, app_model: App, node_id: str):
+        return dump_response(
+            AgentComposerCandidatesResponse,
+            AgentComposerService.get_workflow_candidates(app_id=app_model.id),
+        )
 
 
 @console_ns.route("/apps/<uuid:app_id>/workflows/draft/nodes/<string:node_id>/agent-composer/impact")
 class WorkflowAgentComposerImpactApi(Resource):
+    @console_ns.expect(console_ns.models[ComposerSavePayload.__name__])
+    @console_ns.response(200, "Workflow agent composer impact", console_ns.models[AgentComposerImpactResponse.__name__])
     @setup_required
     @login_required
     @account_initialization_required
     @get_app_model(mode=[AppMode.WORKFLOW, AppMode.ADVANCED_CHAT])
-    def post(self, app_model, node_id: str):
+    def post(self, app_model: App, node_id: str):
         _, tenant_id = current_account_with_tenant()
         payload = ComposerSavePayload.model_validate(console_ns.payload or {})
         current_snapshot_id = payload.binding.current_snapshot_id if payload.binding else None
         if not current_snapshot_id:
-            return {"current_snapshot_id": None, "workflow_node_count": 0, "bindings": []}
-        return AgentComposerService.calculate_impact(tenant_id=tenant_id, current_snapshot_id=current_snapshot_id)
+            return dump_response(
+                AgentComposerImpactResponse, {"current_snapshot_id": None, "workflow_node_count": 0, "bindings": []}
+            )
+        return dump_response(
+            AgentComposerImpactResponse,
+            AgentComposerService.calculate_impact(tenant_id=tenant_id, current_snapshot_id=current_snapshot_id),
+        )
 
 
 @console_ns.route("/apps/<uuid:app_id>/workflows/draft/nodes/<string:node_id>/agent-composer/save-to-roster")
 class WorkflowAgentComposerSaveToRosterApi(Resource):
     @console_ns.expect(console_ns.models[ComposerSavePayload.__name__])
+    @console_ns.response(
+        200, "Workflow agent composer saved to roster", console_ns.models[WorkflowAgentComposerResponse.__name__]
+    )
     @setup_required
     @login_required
     @account_initialization_required
     @edit_permission_required
     @get_app_model(mode=[AppMode.WORKFLOW, AppMode.ADVANCED_CHAT])
-    def post(self, app_model, node_id: str):
+    def post(self, app_model: App, node_id: str):
         account, tenant_id = current_account_with_tenant()
         payload = ComposerSavePayload.model_validate(console_ns.payload or {})
-        return AgentComposerService.save_workflow_composer(
-            tenant_id=tenant_id,
-            app_id=app_model.id,
-            node_id=node_id,
-            account_id=account.id,
-            payload=payload,
+        return dump_response(
+            WorkflowAgentComposerResponse,
+            AgentComposerService.save_workflow_composer(
+                tenant_id=tenant_id,
+                app_id=app_model.id,
+                node_id=node_id,
+                account_id=account.id,
+                payload=payload,
+            ),
         )
 
 
 @console_ns.route("/apps/<uuid:app_id>/agent-composer")
 class AgentAppComposerApi(Resource):
+    @console_ns.response(200, "Agent app composer state", console_ns.models[AgentAppComposerResponse.__name__])
     @setup_required
     @login_required
     @account_initialization_required
     @get_app_model()
-    def get(self, app_model):
+    def get(self, app_model: App):
         _, tenant_id = current_account_with_tenant()
-        return AgentComposerService.load_agent_app_composer(tenant_id=tenant_id, app_id=app_model.id)
+        return dump_response(
+            AgentAppComposerResponse,
+            AgentComposerService.load_agent_app_composer(tenant_id=tenant_id, app_id=app_model.id),
+        )
 
     @console_ns.expect(console_ns.models[ComposerSavePayload.__name__])
+    @console_ns.response(200, "Agent app composer saved", console_ns.models[AgentAppComposerResponse.__name__])
     @setup_required
     @login_required
     @account_initialization_required
     @edit_permission_required
     @get_app_model()
-    def put(self, app_model):
+    def put(self, app_model: App):
         account, tenant_id = current_account_with_tenant()
         payload = ComposerSavePayload.model_validate(console_ns.payload or {})
-        return AgentComposerService.save_agent_app_composer(
-            tenant_id=tenant_id,
-            app_id=app_model.id,
-            account_id=account.id,
-            payload=payload,
+        return dump_response(
+            AgentAppComposerResponse,
+            AgentComposerService.save_agent_app_composer(
+                tenant_id=tenant_id,
+                app_id=app_model.id,
+                account_id=account.id,
+                payload=payload,
+            ),
         )
 
 
 @console_ns.route("/apps/<uuid:app_id>/agent-composer/validate")
 class AgentAppComposerValidateApi(Resource):
     @console_ns.expect(console_ns.models[ComposerSavePayload.__name__])
+    @console_ns.response(
+        200, "Agent app composer validation result", console_ns.models[AgentComposerValidateResponse.__name__]
+    )
     @setup_required
     @login_required
     @account_initialization_required
     @get_app_model()
-    def post(self, app_model):
+    def post(self, app_model: App):
         payload = ComposerSavePayload.model_validate(console_ns.payload or {})
         ComposerConfigValidator.validate_save_payload(payload)
-        return {"result": "success", "errors": []}
+        return dump_response(AgentComposerValidateResponse, {"result": "success", "errors": []})
 
 
 @console_ns.route("/apps/<uuid:app_id>/agent-composer/candidates")
 class AgentAppComposerCandidatesApi(Resource):
+    @console_ns.response(
+        200, "Agent app composer candidates", console_ns.models[AgentComposerCandidatesResponse.__name__]
+    )
     @setup_required
     @login_required
     @account_initialization_required
     @get_app_model()
-    def get(self, app_model):
-        return AgentComposerService.get_agent_app_candidates(app_id=app_model.id)
+    def get(self, app_model: App):
+        return dump_response(
+            AgentComposerCandidatesResponse,
+            AgentComposerService.get_agent_app_candidates(app_id=app_model.id),
+        )

api/controllers/console/agent/roster.py

@@ -4,10 +4,18 @@ from flask import request
 from flask_restx import Resource
 from pydantic import BaseModel, Field
 
-from controllers.common.schema import register_schema_models
+from controllers.common.schema import query_params_from_model, register_response_schema_models, register_schema_models
 from controllers.console import console_ns
 from controllers.console.wraps import account_initialization_required, edit_permission_required, setup_required
 from extensions.ext_database import db
+from fields.agent_fields import (
+    AgentConfigSnapshotDetailResponse,
+    AgentConfigSnapshotListResponse,
+    AgentInviteOptionsResponse,
+    AgentRosterListResponse,
+    AgentRosterResponse,
+)
+from libs.helper import dump_response
 from libs.login import current_account_with_tenant, login_required
 from services.agent.roster_service import AgentRosterService
 from services.entities.agent_entities import RosterAgentCreatePayload, RosterAgentUpdatePayload, RosterListQuery
@@ -29,6 +37,14 @@ register_schema_models(
     RosterAgentUpdatePayload,
     RosterListQuery,
 )
+register_response_schema_models(
+    console_ns,
+    AgentConfigSnapshotDetailResponse,
+    AgentConfigSnapshotListResponse,
+    AgentInviteOptionsResponse,
+    AgentRosterListResponse,
+    AgentRosterResponse,
+)
 
 
 def _agent_roster_service() -> AgentRosterService:
@@ -37,17 +53,23 @@ def _agent_roster_service() -> AgentRosterService:
 
 @console_ns.route("/agents")
 class AgentRosterListApi(Resource):
+    @console_ns.doc(params=query_params_from_model(RosterListQuery))
+    @console_ns.response(200, "Agent roster list", console_ns.models[AgentRosterListResponse.__name__])
     @setup_required
     @login_required
     @account_initialization_required
     def get(self):
         _, tenant_id = current_account_with_tenant()
         query = RosterListQuery.model_validate(request.args.to_dict(flat=True))
-        return _agent_roster_service().list_roster_agents(
-            tenant_id=tenant_id, page=query.page, limit=query.limit, keyword=query.keyword
+        return dump_response(
+            AgentRosterListResponse,
+            _agent_roster_service().list_roster_agents(
+                tenant_id=tenant_id, page=query.page, limit=query.limit, keyword=query.keyword
+            ),
         )
 
     @console_ns.expect(console_ns.models[RosterAgentCreatePayload.__name__])
+    @console_ns.response(201, "Agent created", console_ns.models[AgentRosterResponse.__name__])
     @setup_required
     @login_required
     @account_initialization_required
@@ -57,36 +79,49 @@ class AgentRosterListApi(Resource):
         payload = RosterAgentCreatePayload.model_validate(console_ns.payload or {})
         service = _agent_roster_service()
         agent = service.create_roster_agent(tenant_id=tenant_id, account_id=account.id, payload=payload)
-        return service.get_roster_agent_detail(tenant_id=tenant_id, agent_id=agent.id), 201
+        return dump_response(
+            AgentRosterResponse,
+            service.get_roster_agent_detail(tenant_id=tenant_id, agent_id=agent.id),
+        ), 201
 
 
 @console_ns.route("/agents/invite-options")
 class AgentInviteOptionsApi(Resource):
+    @console_ns.doc(params=query_params_from_model(AgentInviteOptionsQuery))
+    @console_ns.response(200, "Agent invite options", console_ns.models[AgentInviteOptionsResponse.__name__])
     @setup_required
     @login_required
     @account_initialization_required
     def get(self):
         _, tenant_id = current_account_with_tenant()
         query = AgentInviteOptionsQuery.model_validate(request.args.to_dict(flat=True))
-        return _agent_roster_service().list_invite_options(
-            tenant_id=tenant_id,
-            page=query.page,
-            limit=query.limit,
-            keyword=query.keyword,
-            app_id=query.app_id,
+        return dump_response(
+            AgentInviteOptionsResponse,
+            _agent_roster_service().list_invite_options(
+                tenant_id=tenant_id,
+                page=query.page,
+                limit=query.limit,
+                keyword=query.keyword,
+                app_id=query.app_id,
+            ),
         )
 
 
 @console_ns.route("/agents/<uuid:agent_id>")
 class AgentRosterDetailApi(Resource):
+    @console_ns.response(200, "Agent detail", console_ns.models[AgentRosterResponse.__name__])
     @setup_required
     @login_required
     @account_initialization_required
     def get(self, agent_id: UUID):
         _, tenant_id = current_account_with_tenant()
-        return _agent_roster_service().get_roster_agent_detail(tenant_id=tenant_id, agent_id=str(agent_id))
+        return dump_response(
+            AgentRosterResponse,
+            _agent_roster_service().get_roster_agent_detail(tenant_id=tenant_id, agent_id=str(agent_id)),
+        )
 
     @console_ns.expect(console_ns.models[RosterAgentUpdatePayload.__name__])
+    @console_ns.response(200, "Agent updated", console_ns.models[AgentRosterResponse.__name__])
     @setup_required
     @login_required
     @account_initialization_required
@@ -94,10 +129,14 @@ class AgentRosterDetailApi(Resource):
     def patch(self, agent_id: UUID):
         account, tenant_id = current_account_with_tenant()
         payload = RosterAgentUpdatePayload.model_validate(console_ns.payload or {})
-        return _agent_roster_service().update_roster_agent(
-            tenant_id=tenant_id, agent_id=str(agent_id), account_id=account.id, payload=payload
+        return dump_response(
+            AgentRosterResponse,
+            _agent_roster_service().update_roster_agent(
+                tenant_id=tenant_id, agent_id=str(agent_id), account_id=account.id, payload=payload
+            ),
         )
 
+    @console_ns.response(204, "Agent archived")
     @setup_required
     @login_required
     @account_initialization_required
@@ -110,23 +149,31 @@ class AgentRosterDetailApi(Resource):
 
 @console_ns.route("/agents/<uuid:agent_id>/versions")
 class AgentRosterVersionsApi(Resource):
+    @console_ns.response(200, "Agent versions", console_ns.models[AgentConfigSnapshotListResponse.__name__])
     @setup_required
     @login_required
     @account_initialization_required
     def get(self, agent_id: UUID):
         _, tenant_id = current_account_with_tenant()
-        return {"data": _agent_roster_service().list_agent_versions(tenant_id=tenant_id, agent_id=str(agent_id))}
+        return dump_response(
+            AgentConfigSnapshotListResponse,
+            {"data": _agent_roster_service().list_agent_versions(tenant_id=tenant_id, agent_id=str(agent_id))},
+        )
 
 
 @console_ns.route("/agents/<uuid:agent_id>/versions/<uuid:version_id>")
 class AgentRosterVersionDetailApi(Resource):
+    @console_ns.response(200, "Agent version detail", console_ns.models[AgentConfigSnapshotDetailResponse.__name__])
     @setup_required
     @login_required
     @account_initialization_required
     def get(self, agent_id: UUID, version_id: UUID):
         _, tenant_id = current_account_with_tenant()
-        return _agent_roster_service().get_agent_version_detail(
-            tenant_id=tenant_id,
-            agent_id=str(agent_id),
-            version_id=str(version_id),
+        return dump_response(
+            AgentConfigSnapshotDetailResponse,
+            _agent_roster_service().get_agent_version_detail(
+                tenant_id=tenant_id,
+                agent_id=str(agent_id),
+                version_id=str(version_id),
+            ),
         )

api/controllers/console/apikey.py

@@ -9,18 +9,25 @@ from sqlalchemy import delete, func, select
 from sqlalchemy.orm import sessionmaker
 from werkzeug.exceptions import Forbidden
 
-from controllers.common.schema import register_schema_models
+from controllers.common.schema import register_response_schema_models
 from extensions.ext_database import db
 from fields.base import ResponseModel
-from libs.helper import to_timestamp
-from libs.login import current_account_with_tenant, login_required
+from libs.helper import dump_response, to_timestamp
+from libs.login import login_required
+from models import Account
 from models.dataset import Dataset
 from models.enums import ApiTokenType
 from models.model import ApiToken, App
 from services.api_token_service import ApiTokenCache
 
 from . import console_ns
-from .wraps import account_initialization_required, edit_permission_required, setup_required
+from .wraps import (
+    account_initialization_required,
+    edit_permission_required,
+    setup_required,
+    with_current_tenant_id,
+    with_current_user,
+)
 
 
 class ApiKeyItem(ResponseModel):
@@ -40,7 +47,7 @@ class ApiKeyList(ResponseModel):
     data: list[ApiKeyItem]
 
 
-register_schema_models(console_ns, ApiKeyItem, ApiKeyList)
+register_response_schema_models(console_ns, ApiKeyItem, ApiKeyList)
 
 
 def _get_resource(resource_id, tenant_id, resource_model):
@@ -64,10 +71,11 @@ class BaseApiKeyListResource(Resource):
     token_prefix: str | None = None
     max_keys = 10
 
-    def get(self, resource_id):
+    def get(self, resource_id: str, current_tenant_id: str) -> dict[str, object]:
+        return dump_response(ApiKeyList, self._get_api_key_list(resource_id, current_tenant_id))
+
+    def _get_api_key_list(self, resource_id: str, current_tenant_id: str) -> ApiKeyList:
         assert self.resource_id_field is not None, "resource_id_field must be set"
-        resource_id = str(resource_id)
-        _, current_tenant_id = current_account_with_tenant()
 
         _get_resource(resource_id, current_tenant_id, self.resource_model)
         keys = db.session.scalars(
@@ -75,13 +83,14 @@ class BaseApiKeyListResource(Resource):
                 ApiToken.type == self.resource_type, getattr(ApiToken, self.resource_id_field) == resource_id
             )
         ).all()
-        return ApiKeyList.model_validate({"data": keys}, from_attributes=True).model_dump(mode="json")
+        return ApiKeyList.model_validate({"data": keys}, from_attributes=True)
 
     @edit_permission_required
-    def post(self, resource_id):
+    def post(self, resource_id: str, current_tenant_id: str) -> tuple[dict[str, object], int]:
+        return dump_response(ApiKeyItem, self._create_api_key(resource_id, current_tenant_id)), 201
+
+    def _create_api_key(self, resource_id: str, current_tenant_id: str) -> ApiToken:
         assert self.resource_id_field is not None, "resource_id_field must be set"
-        resource_id = str(resource_id)
-        _, current_tenant_id = current_account_with_tenant()
         _get_resource(resource_id, current_tenant_id, self.resource_model)
         current_key_count: int = (
             db.session.scalar(
@@ -108,7 +117,7 @@ class BaseApiKeyListResource(Resource):
         api_token.type = self.resource_type
         db.session.add(api_token)
         db.session.commit()
-        return ApiKeyItem.model_validate(api_token, from_attributes=True).model_dump(mode="json"), 201
+        return api_token
 
 
 class BaseApiKeyResource(Resource):
@@ -118,9 +127,20 @@ class BaseApiKeyResource(Resource):
     resource_model: type | None = None
     resource_id_field: str | None = None
 
-    def delete(self, resource_id: str, api_key_id: str):
+    def delete(
+        self, resource_id: str, api_key_id: str, current_tenant_id: str, current_user: Account
+    ) -> tuple[str, int]:
+        self._delete_api_key(resource_id, api_key_id, current_tenant_id, current_user)
+        return "", 204
+
+    def _delete_api_key(
+        self,
+        resource_id: str,
+        api_key_id: str,
+        current_tenant_id: str,
+        current_user: Account,
+    ) -> None:
         assert self.resource_id_field is not None, "resource_id_field must be set"
-        current_user, current_tenant_id = current_account_with_tenant()
         _get_resource(resource_id, current_tenant_id, self.resource_model)
 
         if not current_user.is_admin_or_owner:
@@ -147,8 +167,6 @@ class BaseApiKeyResource(Resource):
         db.session.execute(delete(ApiToken).where(ApiToken.id == api_key_id))
         db.session.commit()
 
-        return "", 204
-
 
 @console_ns.route("/apps/<uuid:resource_id>/api-keys")
 class AppApiKeyListResource(BaseApiKeyListResource):
@@ -156,18 +174,21 @@ class AppApiKeyListResource(BaseApiKeyListResource):
     @console_ns.doc(description="Get all API keys for an app")
     @console_ns.doc(params={"resource_id": "App ID"})
     @console_ns.response(200, "API keys retrieved successfully", console_ns.models[ApiKeyList.__name__])
-    def get(self, resource_id: UUID):
+    @with_current_tenant_id
+    def get(self, current_tenant_id: str, resource_id: UUID) -> dict[str, object]:
         """Get all API keys for an app"""
-        return super().get(resource_id)
+        return dump_response(ApiKeyList, self._get_api_key_list(str(resource_id), current_tenant_id))
 
     @console_ns.doc("create_app_api_key")
     @console_ns.doc(description="Create a new API key for an app")
     @console_ns.doc(params={"resource_id": "App ID"})
     @console_ns.response(201, "API key created successfully", console_ns.models[ApiKeyItem.__name__])
     @console_ns.response(400, "Maximum keys exceeded")
-    def post(self, resource_id: UUID):
+    @with_current_tenant_id
+    @edit_permission_required
+    def post(self, current_tenant_id: str, resource_id: UUID) -> tuple[dict[str, object], int]:
         """Create a new API key for an app"""
-        return super().post(resource_id)
+        return dump_response(ApiKeyItem, self._create_api_key(str(resource_id), current_tenant_id)), 201
 
     resource_type = ApiTokenType.APP
     resource_model = App
@@ -181,9 +202,14 @@ class AppApiKeyResource(BaseApiKeyResource):
     @console_ns.doc(description="Delete an API key for an app")
     @console_ns.doc(params={"resource_id": "App ID", "api_key_id": "API key ID"})
     @console_ns.response(204, "API key deleted successfully")
-    def delete(self, resource_id: UUID, api_key_id: UUID):
+    @with_current_user
+    @with_current_tenant_id
+    def delete(
+        self, current_tenant_id: str, current_user: Account, resource_id: UUID, api_key_id: UUID
+    ) -> tuple[str, int]:
         """Delete an API key for an app"""
-        return super().delete(str(resource_id), str(api_key_id))
+        self._delete_api_key(str(resource_id), str(api_key_id), current_tenant_id, current_user)
+        return "", 204
 
     resource_type = ApiTokenType.APP
     resource_model = App
@@ -196,18 +222,21 @@ class DatasetApiKeyListResource(BaseApiKeyListResource):
     @console_ns.doc(description="Get all API keys for a dataset")
     @console_ns.doc(params={"resource_id": "Dataset ID"})
     @console_ns.response(200, "API keys retrieved successfully", console_ns.models[ApiKeyList.__name__])
-    def get(self, resource_id: UUID):
+    @with_current_tenant_id
+    def get(self, current_tenant_id: str, resource_id: UUID) -> dict[str, object]:
         """Get all API keys for a dataset"""
-        return super().get(resource_id)
+        return dump_response(ApiKeyList, self._get_api_key_list(str(resource_id), current_tenant_id))
 
     @console_ns.doc("create_dataset_api_key")
     @console_ns.doc(description="Create a new API key for a dataset")
     @console_ns.doc(params={"resource_id": "Dataset ID"})
     @console_ns.response(201, "API key created successfully", console_ns.models[ApiKeyItem.__name__])
     @console_ns.response(400, "Maximum keys exceeded")
-    def post(self, resource_id: UUID):
+    @with_current_tenant_id
+    @edit_permission_required
+    def post(self, current_tenant_id: str, resource_id: UUID) -> tuple[dict[str, object], int]:
         """Create a new API key for a dataset"""
-        return super().post(resource_id)
+        return dump_response(ApiKeyItem, self._create_api_key(str(resource_id), current_tenant_id)), 201
 
     resource_type = ApiTokenType.DATASET
     resource_model = Dataset
@@ -221,9 +250,14 @@ class DatasetApiKeyResource(BaseApiKeyResource):
     @console_ns.doc(description="Delete an API key for a dataset")
     @console_ns.doc(params={"resource_id": "Dataset ID", "api_key_id": "API key ID"})
     @console_ns.response(204, "API key deleted successfully")
-    def delete(self, resource_id: UUID, api_key_id: UUID):
+    @with_current_user
+    @with_current_tenant_id
+    def delete(
+        self, current_tenant_id: str, current_user: Account, resource_id: UUID, api_key_id: UUID
+    ) -> tuple[str, int]:
         """Delete an API key for a dataset"""
-        return super().delete(str(resource_id), str(api_key_id))
+        self._delete_api_key(str(resource_id), str(api_key_id), current_tenant_id, current_user)
+        return "", 204
 
     resource_type = ApiTokenType.DATASET
     resource_model = Dataset

api/controllers/console/app/agent.py

@@ -8,7 +8,7 @@ from controllers.console.app.wraps import get_app_model
 from controllers.console.wraps import account_initialization_required, setup_required
 from libs.helper import uuid_value
 from libs.login import login_required
-from models.model import AppMode
+from models.model import App, AppMode
 from services.agent_service import AgentService
 
 
@@ -39,7 +39,7 @@ class AgentLogApi(Resource):
     @login_required
     @account_initialization_required
     @get_app_model(mode=[AppMode.AGENT_CHAT])
-    def get(self, app_model):
+    def get(self, app_model: App):
         """Get agent logs"""
         args = AgentLogQuery.model_validate(request.args.to_dict(flat=True))
 

api/controllers/console/app/app.py

@@ -573,7 +573,7 @@ class AppApi(Resource):
     @account_initialization_required
     @enterprise_license_required
     @get_app_model(mode=None)
-    def get(self, app_model):
+    def get(self, app_model: App):
         """Get app detail"""
         app_service = AppService()
 
@@ -581,7 +581,7 @@ class AppApi(Resource):
 
         if FeatureService.get_system_features().webapp_auth.enabled:
             app_setting = EnterpriseService.WebAppAuth.get_app_access_mode_by_id(app_id=str(app_model.id))
-            app_model.access_mode = app_setting.access_mode
+            app_model.access_mode = app_setting.access_mode  # type: ignore[attr-defined]
 
         response_model = AppDetailWithSite.model_validate(app_model, from_attributes=True)
         return response_model.model_dump(mode="json")
@@ -598,7 +598,7 @@ class AppApi(Resource):
     @account_initialization_required
     @get_app_model(mode=None)
     @edit_permission_required
-    def put(self, app_model):
+    def put(self, app_model: App):
         """Update app"""
         args = UpdateAppPayload.model_validate(console_ns.payload)
 
@@ -627,7 +627,7 @@ class AppApi(Resource):
     @login_required
     @account_initialization_required
     @edit_permission_required
-    def delete(self, app_model):
+    def delete(self, app_model: App):
         """Delete app"""
         app_service = AppService()
         app_service.delete_app(app_model)
@@ -648,7 +648,7 @@ class AppCopyApi(Resource):
     @account_initialization_required
     @get_app_model(mode=None)
     @edit_permission_required
-    def post(self, app_model):
+    def post(self, app_model: App):
         """Copy app"""
         # The role of the current user in the ta table must be admin, owner, or editor
         current_user, _ = current_account_with_tenant()
@@ -709,7 +709,7 @@ class AppExportApi(Resource):
     @login_required
     @account_initialization_required
     @edit_permission_required
-    def get(self, app_model):
+    def get(self, app_model: App):
         """Export app"""
         args = AppExportQuery.model_validate(request.args.to_dict(flat=True))
 
@@ -731,7 +731,7 @@ class AppPublishToCreatorsPlatformApi(Resource):
     @account_initialization_required
     @get_app_model(mode=None)
     @edit_permission_required
-    def post(self, app_model):
+    def post(self, app_model: App):
         """Publish app to Creators Platform"""
         from configs import dify_config
         from core.helper.creators import get_redirect_url, upload_dsl
@@ -762,7 +762,7 @@ class AppNameApi(Resource):
     @account_initialization_required
     @get_app_model(mode=None)
     @edit_permission_required
-    def post(self, app_model):
+    def post(self, app_model: App):
         args = AppNamePayload.model_validate(console_ns.payload)
 
         app_service = AppService()
@@ -784,7 +784,7 @@ class AppIconApi(Resource):
     @account_initialization_required
     @get_app_model(mode=None)
     @edit_permission_required
-    def post(self, app_model):
+    def post(self, app_model: App):
         args = AppIconPayload.model_validate(console_ns.payload or {})
 
         app_service = AppService()
@@ -811,7 +811,7 @@ class AppSiteStatus(Resource):
     @account_initialization_required
     @get_app_model(mode=None)
     @edit_permission_required
-    def post(self, app_model):
+    def post(self, app_model: App):
         args = AppSiteStatusPayload.model_validate(console_ns.payload)
 
         app_service = AppService()
@@ -833,7 +833,7 @@ class AppApiStatus(Resource):
     @is_admin_or_owner_required
     @account_initialization_required
     @get_app_model(mode=None)
-    def post(self, app_model):
+    def post(self, app_model: App):
         args = AppApiStatusPayload.model_validate(console_ns.payload)
 
         app_service = AppService()
@@ -874,7 +874,7 @@ class AppTraceApi(Resource):
     @account_initialization_required
     @edit_permission_required
     @get_app_model
-    def post(self, app_model):
+    def post(self, app_model: App):
         # add app trace
         args = AppTracePayload.model_validate(console_ns.payload)
 

api/controllers/console/app/audio.py

@@ -70,7 +70,7 @@ class ChatMessageAudioApi(Resource):
     @login_required
     @account_initialization_required
     @get_app_model(mode=[AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT])
-    def post(self, app_model):
+    def post(self, app_model: App):
         file = request.files["file"]
 
         try:
@@ -171,7 +171,7 @@ class TextModesApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def get(self, app_model):
+    def get(self, app_model: App):
         try:
             args = TextToSpeechVoiceQuery.model_validate(request.args.to_dict(flat=True))
 

api/controllers/console/app/completion.py

@@ -33,7 +33,7 @@ from libs import helper
 from libs.helper import uuid_value
 from libs.login import current_user, login_required
 from models import Account
-from models.model import AppMode
+from models.model import App, AppMode
 from services.app_generate_service import AppGenerateService
 from services.app_task_service import AppTaskService
 from services.errors.llm import InvokeRateLimitError
@@ -84,7 +84,7 @@ class CompletionMessageApi(Resource):
     @login_required
     @account_initialization_required
     @get_app_model(mode=AppMode.COMPLETION)
-    def post(self, app_model):
+    def post(self, app_model: App):
         args_model = CompletionMessagePayload.model_validate(console_ns.payload)
         args = args_model.model_dump(exclude_none=True, by_alias=True)
 
@@ -131,7 +131,7 @@ class CompletionMessageStopApi(Resource):
     @login_required
     @account_initialization_required
     @get_app_model(mode=AppMode.COMPLETION)
-    def post(self, app_model, task_id: str):
+    def post(self, app_model: App, task_id: str):
         if not isinstance(current_user, Account):
             raise ValueError("current_user must be an Account instance")
 
@@ -159,7 +159,7 @@ class ChatMessageApi(Resource):
     @account_initialization_required
     @get_app_model(mode=[AppMode.CHAT, AppMode.AGENT_CHAT])
     @edit_permission_required
-    def post(self, app_model):
+    def post(self, app_model: App):
         args_model = ChatMessagePayload.model_validate(console_ns.payload)
         args = args_model.model_dump(exclude_none=True, by_alias=True)
 
@@ -212,7 +212,7 @@ class ChatMessageStopApi(Resource):
     @login_required
     @account_initialization_required
     @get_app_model(mode=[AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT])
-    def post(self, app_model, task_id: str):
+    def post(self, app_model: App, task_id: str):
         if not isinstance(current_user, Account):
             raise ValueError("current_user must be an Account instance")
 

api/controllers/console/app/conversation.py

@@ -33,7 +33,7 @@ from fields.conversation_fields import (
 from libs.datetime_utils import naive_utc_now, parse_time_range
 from libs.login import current_account_with_tenant, login_required
 from models import Conversation, EndUser, Message, MessageAnnotation
-from models.model import AppMode
+from models.model import App, AppMode
 from services.conversation_service import ConversationService
 from services.errors.conversation import ConversationNotExistsError
 
@@ -93,7 +93,7 @@ class CompletionConversationApi(Resource):
     @account_initialization_required
     @get_app_model(mode=AppMode.COMPLETION)
     @edit_permission_required
-    def get(self, app_model):
+    def get(self, app_model: App):
         current_user, _ = current_account_with_tenant()
         args = CompletionConversationQuery.model_validate(request.args.to_dict(flat=True))
 
@@ -165,7 +165,7 @@ class CompletionConversationDetailApi(Resource):
     @account_initialization_required
     @get_app_model(mode=AppMode.COMPLETION)
     @edit_permission_required
-    def get(self, app_model, conversation_id: UUID):
+    def get(self, app_model: App, conversation_id: UUID):
         conversation_id_str = str(conversation_id)
         return ConversationMessageDetailResponse.model_validate(
             _get_conversation(app_model, conversation_id_str), from_attributes=True
@@ -182,7 +182,7 @@ class CompletionConversationDetailApi(Resource):
     @account_initialization_required
     @get_app_model(mode=AppMode.COMPLETION)
     @edit_permission_required
-    def delete(self, app_model, conversation_id: UUID):
+    def delete(self, app_model: App, conversation_id: UUID):
         current_user, _ = current_account_with_tenant()
         conversation_id_str = str(conversation_id)
 
@@ -207,7 +207,7 @@ class ChatConversationApi(Resource):
     @account_initialization_required
     @get_app_model(mode=[AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT])
     @edit_permission_required
-    def get(self, app_model):
+    def get(self, app_model: App):
         current_user, _ = current_account_with_tenant()
         args = ChatConversationQuery.model_validate(request.args.to_dict(flat=True))
 
@@ -318,7 +318,7 @@ class ChatConversationDetailApi(Resource):
     @account_initialization_required
     @get_app_model(mode=[AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT])
     @edit_permission_required
-    def get(self, app_model, conversation_id: UUID):
+    def get(self, app_model: App, conversation_id: UUID):
         conversation_id_str = str(conversation_id)
         return ConversationDetailResponse.model_validate(
             _get_conversation(app_model, conversation_id_str), from_attributes=True
@@ -335,7 +335,7 @@ class ChatConversationDetailApi(Resource):
     @get_app_model(mode=[AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT])
     @account_initialization_required
     @edit_permission_required
-    def delete(self, app_model, conversation_id: UUID):
+    def delete(self, app_model: App, conversation_id: UUID):
         current_user, _ = current_account_with_tenant()
         conversation_id_str = str(conversation_id)
 

api/controllers/console/app/conversation_variables.py

@@ -19,7 +19,7 @@ from fields.base import ResponseModel
 from libs.helper import to_timestamp
 from libs.login import login_required
 from models import ConversationVariable
-from models.model import AppMode
+from models.model import App, AppMode
 
 
 class ConversationVariablesQuery(BaseModel):
@@ -94,7 +94,7 @@ class ConversationVariablesApi(Resource):
     @login_required
     @account_initialization_required
     @get_app_model(mode=AppMode.ADVANCED_CHAT)
-    def get(self, app_model):
+    def get(self, app_model: App):
         args = ConversationVariablesQuery.model_validate(request.args.to_dict(flat=True))
 
         stmt = (

api/controllers/console/app/generator.py

@@ -11,7 +11,7 @@ from controllers.console.app.error import (
     ProviderNotInitializeError,
     ProviderQuotaExceededError,
 )
-from controllers.console.wraps import account_initialization_required, setup_required
+from controllers.console.wraps import account_initialization_required, setup_required, with_current_tenant_id
 from core.app.app_config.entities import ModelConfig
 from core.errors.error import ModelCurrentlyNotSupportError, ProviderTokenNotInitError, QuotaExceededError
 from core.helper.code_executor.code_node_provider import CodeNodeProvider
@@ -22,7 +22,7 @@ from core.llm_generator.llm_generator import LLMGenerator
 from extensions.ext_database import db
 from graphon.model_runtime.entities.llm_entities import LLMMode
 from graphon.model_runtime.errors.invoke import InvokeError
-from libs.login import current_account_with_tenant, login_required
+from libs.login import login_required
 from models import App
 from services.workflow_service import WorkflowService
 
@@ -64,9 +64,9 @@ class RuleGenerateApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def post(self):
+    @with_current_tenant_id
+    def post(self, current_tenant_id: str):
         args = RuleGeneratePayload.model_validate(console_ns.payload)
-        _, current_tenant_id = current_account_with_tenant()
 
         try:
             rules = LLMGenerator.generate_rule_config(tenant_id=current_tenant_id, args=args)
@@ -93,9 +93,9 @@ class RuleCodeGenerateApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def post(self):
+    @with_current_tenant_id
+    def post(self, current_tenant_id: str):
         args = RuleCodeGeneratePayload.model_validate(console_ns.payload)
-        _, current_tenant_id = current_account_with_tenant()
 
         try:
             code_result = LLMGenerator.generate_code(
@@ -125,9 +125,9 @@ class RuleStructuredOutputGenerateApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def post(self):
+    @with_current_tenant_id
+    def post(self, current_tenant_id: str):
         args = RuleStructuredOutputPayload.model_validate(console_ns.payload)
-        _, current_tenant_id = current_account_with_tenant()
 
         try:
             structured_output = LLMGenerator.generate_structured_output(
@@ -157,9 +157,9 @@ class InstructionGenerateApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def post(self):
+    @with_current_tenant_id
+    def post(self, current_tenant_id: str):
         args = InstructionGeneratePayload.model_validate(console_ns.payload)
-        _, current_tenant_id = current_account_with_tenant()
         providers: list[type[CodeNodeProvider]] = [Python3CodeProvider, JavascriptCodeProvider]
         code_provider: type[CodeNodeProvider] | None = next(
             (p for p in providers if p.is_accept_language(args.language)), None

api/controllers/console/app/mcp_server.py

@@ -11,13 +11,18 @@ from werkzeug.exceptions import NotFound
 from controllers.common.schema import register_schema_models
 from controllers.console import console_ns
 from controllers.console.app.wraps import get_app_model
-from controllers.console.wraps import account_initialization_required, edit_permission_required, setup_required
+from controllers.console.wraps import (
+    account_initialization_required,
+    edit_permission_required,
+    setup_required,
+    with_current_tenant_id,
+)
 from extensions.ext_database import db
 from fields.base import ResponseModel
 from libs.helper import to_timestamp
-from libs.login import current_account_with_tenant, login_required
+from libs.login import login_required
 from models.enums import AppMCPServerStatus
-from models.model import AppMCPServer
+from models.model import App, AppMCPServer
 
 
 class MCPServerCreatePayload(BaseModel):
@@ -73,7 +78,7 @@ class AppMCPServerController(Resource):
     @account_initialization_required
     @setup_required
     @get_app_model
-    def get(self, app_model):
+    def get(self, app_model: App):
         server = db.session.scalar(select(AppMCPServer).where(AppMCPServer.app_id == app_model.id).limit(1))
         if server is None:
             return {}
@@ -92,8 +97,8 @@ class AppMCPServerController(Resource):
     @login_required
     @setup_required
     @edit_permission_required
-    def post(self, app_model):
-        _, current_tenant_id = current_account_with_tenant()
+    @with_current_tenant_id
+    def post(self, current_tenant_id: str, app_model: App):
         payload = MCPServerCreatePayload.model_validate(console_ns.payload or {})
 
         description = payload.description
@@ -127,7 +132,7 @@ class AppMCPServerController(Resource):
     @setup_required
     @account_initialization_required
     @edit_permission_required
-    def put(self, app_model):
+    def put(self, app_model: App):
         payload = MCPServerUpdatePayload.model_validate(console_ns.payload or {})
         server = db.session.get(AppMCPServer, payload.id)
         if not server:
@@ -163,8 +168,8 @@ class AppMCPServerRefreshController(Resource):
     @login_required
     @account_initialization_required
     @edit_permission_required
-    def get(self, server_id: UUID):
-        _, current_tenant_id = current_account_with_tenant()
+    @with_current_tenant_id
+    def get(self, current_tenant_id: str, server_id: UUID):
         server = db.session.scalar(
             select(AppMCPServer)
             .where(AppMCPServer.id == server_id, AppMCPServer.tenant_id == current_tenant_id)

api/controllers/console/app/message.py

@@ -45,7 +45,7 @@ from libs.helper import to_timestamp, uuid_value
 from libs.infinite_scroll_pagination import InfiniteScrollPagination
 from libs.login import current_account_with_tenant, login_required
 from models.enums import FeedbackFromSource, FeedbackRating
-from models.model import AppMode, Conversation, Message, MessageAnnotation, MessageFeedback
+from models.model import App, AppMode, Conversation, Message, MessageAnnotation, MessageFeedback
 from services.errors.conversation import ConversationNotExistsError
 from services.errors.message import MessageNotExistsError, SuggestedQuestionsAfterAnswerDisabledError
 from services.message_service import MessageService, attach_message_extra_contents
@@ -180,7 +180,7 @@ class ChatMessageListApi(Resource):
     @setup_required
     @get_app_model(mode=[AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT])
     @edit_permission_required
-    def get(self, app_model):
+    def get(self, app_model: App):
         args = ChatMessagesQuery.model_validate(request.args.to_dict())
 
         conversation = db.session.scalar(
@@ -257,7 +257,7 @@ class MessageFeedbackApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def post(self, app_model):
+    def post(self, app_model: App):
         current_user, _ = current_account_with_tenant()
 
         args = MessageFeedbackPayload.model_validate(console_ns.payload)
@@ -314,7 +314,7 @@ class MessageAnnotationCountApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def get(self, app_model):
+    def get(self, app_model: App):
         count = db.session.scalar(
             select(func.count(MessageAnnotation.id)).where(MessageAnnotation.app_id == app_model.id)
         )
@@ -337,7 +337,7 @@ class MessageSuggestedQuestionApi(Resource):
     @login_required
     @account_initialization_required
     @get_app_model(mode=[AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT])
-    def get(self, app_model, message_id: UUID):
+    def get(self, app_model: App, message_id: UUID):
         current_user, _ = current_account_with_tenant()
         message_id_str = str(message_id)
 
@@ -379,7 +379,7 @@ class MessageFeedbackExportApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def get(self, app_model):
+    def get(self, app_model: App):
         args = FeedbackExportQuery.model_validate(request.args.to_dict())
 
         # Import the service function
@@ -417,7 +417,7 @@ class MessageApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def get(self, app_model, message_id: UUID):
+    def get(self, app_model: App, message_id: UUID):
         message_id_str = str(message_id)
 
         message = db.session.scalar(

api/controllers/console/app/model_config.py

@@ -16,7 +16,7 @@ from events.app_event import app_model_config_was_updated
 from extensions.ext_database import db
 from libs.datetime_utils import naive_utc_now
 from libs.login import current_account_with_tenant, login_required
-from models.model import AppMode, AppModelConfig
+from models.model import App, AppMode, AppModelConfig
 from services.app_model_config_service import AppModelConfigService
 
 
@@ -52,7 +52,7 @@ class ModelConfigResource(Resource):
     @edit_permission_required
     @account_initialization_required
     @get_app_model(mode=[AppMode.AGENT_CHAT, AppMode.CHAT, AppMode.COMPLETION])
-    def post(self, app_model):
+    def post(self, app_model: App):
         """Modify app model config"""
         current_user, current_tenant_id = current_account_with_tenant()
         # validate config

api/controllers/console/app/site.py

@@ -20,6 +20,7 @@ from fields.base import ResponseModel
 from libs.datetime_utils import naive_utc_now
 from libs.login import current_account_with_tenant, login_required
 from models import Site
+from models.model import App
 
 
 class AppSiteUpdatePayload(BaseModel):
@@ -84,7 +85,7 @@ class AppSite(Resource):
     @edit_permission_required
     @account_initialization_required
     @get_app_model
-    def post(self, app_model):
+    def post(self, app_model: App):
         args = AppSiteUpdatePayload.model_validate(console_ns.payload or {})
         current_user, _ = current_account_with_tenant()
         site = db.session.scalar(select(Site).where(Site.app_id == app_model.id).limit(1))
@@ -133,7 +134,7 @@ class AppSiteAccessTokenReset(Resource):
     @is_admin_or_owner_required
     @account_initialization_required
     @get_app_model
-    def post(self, app_model):
+    def post(self, app_model: App):
         current_user, _ = current_account_with_tenant()
         site = db.session.scalar(select(Site).where(Site.app_id == app_model.id).limit(1))
 

api/controllers/console/app/statistic.py

@@ -15,6 +15,7 @@ from libs.datetime_utils import parse_time_range
 from libs.helper import convert_datetime_to_date
 from libs.login import current_account_with_tenant, login_required
 from models import AppMode
+from models.model import App
 
 
 class StatisticTimeRangeQuery(BaseModel):
@@ -47,7 +48,7 @@ class DailyMessageStatistic(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def get(self, app_model):
+    def get(self, app_model: App):
         account, _ = current_account_with_tenant()
 
         args = StatisticTimeRangeQuery.model_validate(request.args.to_dict(flat=True))
@@ -61,8 +62,12 @@ FROM
 WHERE
     app_id = :app_id
     AND invoke_from != :invoke_from"""
-        arg_dict = {"tz": account.timezone, "app_id": app_model.id, "invoke_from": InvokeFrom.DEBUGGER}
         assert account.timezone is not None
+        arg_dict: dict[str, object] = {
+            "tz": account.timezone,
+            "app_id": app_model.id,
+            "invoke_from": InvokeFrom.DEBUGGER,
+        }
 
         try:
             start_datetime_utc, end_datetime_utc = parse_time_range(args.start, args.end, account.timezone)
@@ -104,7 +109,7 @@ class DailyConversationStatistic(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def get(self, app_model):
+    def get(self, app_model: App):
         account, _ = current_account_with_tenant()
 
         args = StatisticTimeRangeQuery.model_validate(request.args.to_dict(flat=True))
@@ -118,8 +123,12 @@ FROM
 WHERE
     app_id = :app_id
     AND invoke_from != :invoke_from"""
-        arg_dict = {"tz": account.timezone, "app_id": app_model.id, "invoke_from": InvokeFrom.DEBUGGER}
         assert account.timezone is not None
+        arg_dict: dict[str, object] = {
+            "tz": account.timezone,
+            "app_id": app_model.id,
+            "invoke_from": InvokeFrom.DEBUGGER,
+        }
 
         try:
             start_datetime_utc, end_datetime_utc = parse_time_range(args.start, args.end, account.timezone)
@@ -160,7 +169,7 @@ class DailyTerminalsStatistic(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def get(self, app_model):
+    def get(self, app_model: App):
         account, _ = current_account_with_tenant()
 
         args = StatisticTimeRangeQuery.model_validate(request.args.to_dict(flat=True))
@@ -174,8 +183,12 @@ FROM
 WHERE
     app_id = :app_id
     AND invoke_from != :invoke_from"""
-        arg_dict = {"tz": account.timezone, "app_id": app_model.id, "invoke_from": InvokeFrom.DEBUGGER}
         assert account.timezone is not None
+        arg_dict: dict[str, object] = {
+            "tz": account.timezone,
+            "app_id": app_model.id,
+            "invoke_from": InvokeFrom.DEBUGGER,
+        }
 
         try:
             start_datetime_utc, end_datetime_utc = parse_time_range(args.start, args.end, account.timezone)
@@ -217,7 +230,7 @@ class DailyTokenCostStatistic(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def get(self, app_model):
+    def get(self, app_model: App):
         account, _ = current_account_with_tenant()
 
         args = StatisticTimeRangeQuery.model_validate(request.args.to_dict(flat=True))
@@ -232,8 +245,12 @@ FROM
 WHERE
     app_id = :app_id
     AND invoke_from != :invoke_from"""
-        arg_dict = {"tz": account.timezone, "app_id": app_model.id, "invoke_from": InvokeFrom.DEBUGGER}
         assert account.timezone is not None
+        arg_dict: dict[str, object] = {
+            "tz": account.timezone,
+            "app_id": app_model.id,
+            "invoke_from": InvokeFrom.DEBUGGER,
+        }
 
         try:
             start_datetime_utc, end_datetime_utc = parse_time_range(args.start, args.end, account.timezone)
@@ -277,7 +294,7 @@ class AverageSessionInteractionStatistic(Resource):
     @login_required
     @account_initialization_required
     @get_app_model(mode=[AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT])
-    def get(self, app_model):
+    def get(self, app_model: App):
         account, _ = current_account_with_tenant()
 
         args = StatisticTimeRangeQuery.model_validate(request.args.to_dict(flat=True))
@@ -299,8 +316,12 @@ FROM
         WHERE
             c.app_id = :app_id
             AND m.invoke_from != :invoke_from"""
-        arg_dict = {"tz": account.timezone, "app_id": app_model.id, "invoke_from": InvokeFrom.DEBUGGER}
         assert account.timezone is not None
+        arg_dict: dict[str, object] = {
+            "tz": account.timezone,
+            "app_id": app_model.id,
+            "invoke_from": InvokeFrom.DEBUGGER,
+        }
 
         try:
             start_datetime_utc, end_datetime_utc = parse_time_range(args.start, args.end, account.timezone)
@@ -353,7 +374,7 @@ class UserSatisfactionRateStatistic(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def get(self, app_model):
+    def get(self, app_model: App):
         account, _ = current_account_with_tenant()
 
         args = StatisticTimeRangeQuery.model_validate(request.args.to_dict(flat=True))
@@ -371,8 +392,12 @@ LEFT JOIN
 WHERE
     m.app_id = :app_id
     AND m.invoke_from != :invoke_from"""
-        arg_dict = {"tz": account.timezone, "app_id": app_model.id, "invoke_from": InvokeFrom.DEBUGGER}
         assert account.timezone is not None
+        arg_dict: dict[str, object] = {
+            "tz": account.timezone,
+            "app_id": app_model.id,
+            "invoke_from": InvokeFrom.DEBUGGER,
+        }
 
         try:
             start_datetime_utc, end_datetime_utc = parse_time_range(args.start, args.end, account.timezone)
@@ -419,7 +444,7 @@ class AverageResponseTimeStatistic(Resource):
     @login_required
     @account_initialization_required
     @get_app_model(mode=AppMode.COMPLETION)
-    def get(self, app_model):
+    def get(self, app_model: App):
         account, _ = current_account_with_tenant()
 
         args = StatisticTimeRangeQuery.model_validate(request.args.to_dict(flat=True))
@@ -433,8 +458,12 @@ FROM
 WHERE
     app_id = :app_id
     AND invoke_from != :invoke_from"""
-        arg_dict = {"tz": account.timezone, "app_id": app_model.id, "invoke_from": InvokeFrom.DEBUGGER}
         assert account.timezone is not None
+        arg_dict: dict[str, object] = {
+            "tz": account.timezone,
+            "app_id": app_model.id,
+            "invoke_from": InvokeFrom.DEBUGGER,
+        }
 
         try:
             start_datetime_utc, end_datetime_utc = parse_time_range(args.start, args.end, account.timezone)
@@ -476,7 +505,7 @@ class TokensPerSecondStatistic(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def get(self, app_model):
+    def get(self, app_model: App):
         account, _ = current_account_with_tenant()
         args = StatisticTimeRangeQuery.model_validate(request.args.to_dict(flat=True))
 
@@ -492,8 +521,12 @@ FROM
 WHERE
     app_id = :app_id
     AND invoke_from != :invoke_from"""
-        arg_dict = {"tz": account.timezone, "app_id": app_model.id, "invoke_from": InvokeFrom.DEBUGGER}
         assert account.timezone is not None
+        arg_dict: dict[str, object] = {
+            "tz": account.timezone,
+            "app_id": app_model.id,
+            "invoke_from": InvokeFrom.DEBUGGER,
+        }
 
         try:
             start_datetime_utc, end_datetime_utc = parse_time_range(args.start, args.end, account.timezone)

api/controllers/console/app/workflow_draft_variable.py

@@ -83,13 +83,14 @@ def _serialize_var_value(variable: WorkflowDraftVariable):
     # create a copy of the value to avoid affecting the model cache.
     value = value.model_copy(deep=True)
     # Refresh the url signature before returning it to client.
-    if isinstance(value, FileSegment):
-        file = value.value
-        file.remote_url = file.generate_url()
-    elif isinstance(value, ArrayFileSegment):
-        files = value.value
-        for file in files:
+    match value:
+        case FileSegment():
+            file = value.value
             file.remote_url = file.generate_url()
+        case ArrayFileSegment():
+            files = value.value
+            for file in files:
+                file.remote_url = file.generate_url()
     return _convert_values_to_json_serializable_object(value)
 
 

api/controllers/console/app/workflow_node_output_inspector.py

@@ -0,0 +1,415 @@
+"""Console REST endpoints for the Node Output Inspector (Stage 4 §8 / §10.3).
+
+PRD §Node Output Inspector replaces the consumer-organized Variable Inspector
+with a producer-organized view of each node's declared outputs and their
+per-run status. This module exposes two parallel sets of three read-only
+endpoints — one for ``/workflows/draft/runs/...`` (Composer test runs) and one
+for ``/workflows/published/runs/...`` (real App API / webapp / webhook /
+schedule / plugin triggers). Both sets share the same service code, the same
+response shapes, and the same error codes; the URL is the *only* difference,
+so the frontend can pick the right prefix based on which run-detail page the
+user is on.
+
+Decision D-1 (published Inspector deferred) was lifted 2026-05-26 — the
+``published_run_inspector_not_implemented`` 404 code is therefore no longer
+produced.
+
+URLs follow the design doc and reuse the existing
+``/apps/<uuid:app_id>/workflows/draft/...`` prefix from
+:mod:`controllers.console.app.workflow_draft_variable`. The
+``published`` prefix mirrors it shape-for-shape.
+"""
+
+from __future__ import annotations
+
+import json
+import logging
+from collections.abc import Iterator
+from uuid import UUID
+
+from flask import Response
+from flask_restx import Resource
+
+from controllers.console import console_ns
+from controllers.console.app.wraps import get_app_model
+from controllers.console.wraps import account_initialization_required, setup_required
+from libs.exception import BaseHTTPException
+from libs.login import login_required
+from models import App, AppMode
+from services.workflow import inspector_events
+from services.workflow.node_output_inspector_service import (
+    NodeOutputInspectorError,
+    NodeOutputInspectorService,
+)
+
+logger = logging.getLogger(__name__)
+
+
+# Heartbeat cadence — every N empty subscribe ticks emit a SSE comment so
+# intervening proxies (nginx, ingress) don't reap the idle connection.
+# ``inspector_events.subscribe`` ticks at 1s, so 15 → 15s heartbeat.
+_HEARTBEAT_EVERY_TICKS = 15
+# Hard ceiling on a single stream — if we never see a terminal workflow
+# event (engine crashed, redis dropped the message), force-close after this
+# many ticks (= seconds).
+_STREAM_HARD_TIMEOUT_TICKS = 1800  # 30 min
+
+
+def _service() -> NodeOutputInspectorService:
+    """One-line factory so tests can monkeypatch a stub if needed."""
+    return NodeOutputInspectorService()
+
+
+def _serve_snapshot(app_model: App, run_id: UUID) -> dict:
+    """Resource-body shared by draft + published snapshot endpoints.
+
+    Pulled out so the 6 REST routes don't duplicate the same 6-line try/except
+    + ``model_dump`` ritual — the routes shrink to one-liners and the actual
+    behaviour lives here, where unit tests can hit it without spinning up
+    Flask request context.
+    """
+    try:
+        snapshot = _service().snapshot_workflow_run(app_model=app_model, workflow_run_id=str(run_id))
+    except NodeOutputInspectorError as error:
+        raise _InspectorNotFound(error) from error
+    return snapshot.model_dump(mode="json")
+
+
+def _serve_node_detail(app_model: App, run_id: UUID, node_id: str) -> dict:
+    """Resource-body shared by draft + published node-detail endpoints."""
+    try:
+        view = _service().node_detail(
+            app_model=app_model,
+            workflow_run_id=str(run_id),
+            node_id=node_id,
+        )
+    except NodeOutputInspectorError as error:
+        raise _InspectorNotFound(error) from error
+    return view.model_dump(mode="json")
+
+
+def _serve_output_preview(app_model: App, run_id: UUID, node_id: str, output_name: str) -> dict:
+    """Resource-body shared by draft + published output-preview endpoints."""
+    try:
+        preview = _service().output_preview(
+            app_model=app_model,
+            workflow_run_id=str(run_id),
+            node_id=node_id,
+            output_name=output_name,
+        )
+    except NodeOutputInspectorError as error:
+        raise _InspectorNotFound(error) from error
+    return preview.model_dump(mode="json")
+
+
+class _InspectorNotFound(BaseHTTPException):
+    """404 that preserves the inspector's specific error code.
+
+    Without this the response body collapses to a generic ``not_found`` code
+    and clients lose the ability to distinguish, e.g.,
+    ``workflow_run_not_found`` from ``published_run_inspector_not_implemented``.
+    """
+
+    code = 404
+
+    def __init__(self, error: NodeOutputInspectorError) -> None:
+        self.error_code = error.code
+        super().__init__(description=str(error))
+
+
+@console_ns.route("/apps/<uuid:app_id>/workflows/draft/runs/<uuid:run_id>/node-outputs")
+class WorkflowDraftRunNodeOutputsApi(Resource):
+    """Whole-run snapshot organized by producer node."""
+
+    @console_ns.doc("get_workflow_draft_run_node_outputs")
+    @console_ns.doc(description="Snapshot of every node's declared outputs for a draft workflow run.")
+    @console_ns.doc(params={"app_id": "Application ID", "run_id": "Workflow run ID"})
+    @console_ns.response(404, "Workflow run not found")
+    @setup_required
+    @login_required
+    @account_initialization_required
+    @get_app_model(mode=[AppMode.ADVANCED_CHAT, AppMode.WORKFLOW])
+    def get(self, app_model: App, run_id: UUID):
+        return _serve_snapshot(app_model, run_id)
+
+
+@console_ns.route("/apps/<uuid:app_id>/workflows/draft/runs/<uuid:run_id>/node-outputs/<string:node_id>")
+class WorkflowDraftRunNodeOutputDetailApi(Resource):
+    """One node's declared outputs + per-output status."""
+
+    @console_ns.doc("get_workflow_draft_run_node_output_detail")
+    @console_ns.doc(description="One node's declared outputs for a draft workflow run.")
+    @console_ns.doc(
+        params={
+            "app_id": "Application ID",
+            "run_id": "Workflow run ID",
+            "node_id": "Node ID inside the workflow graph",
+        }
+    )
+    @console_ns.response(404, "Workflow run / node not found")
+    @setup_required
+    @login_required
+    @account_initialization_required
+    @get_app_model(mode=[AppMode.ADVANCED_CHAT, AppMode.WORKFLOW])
+    def get(self, app_model: App, run_id: UUID, node_id: str):
+        return _serve_node_detail(app_model, run_id, node_id)
+
+
+@console_ns.route(
+    "/apps/<uuid:app_id>/workflows/draft/runs/<uuid:run_id>/node-outputs/<string:node_id>/<string:output_name>/preview"
+)
+class WorkflowDraftRunNodeOutputPreviewApi(Resource):
+    """Full value for one declared output (with signed URL for file refs)."""
+
+    @console_ns.doc("get_workflow_draft_run_node_output_preview")
+    @console_ns.doc(description="Full value for one declared output, including signed download URL for files.")
+    @console_ns.doc(
+        params={
+            "app_id": "Application ID",
+            "run_id": "Workflow run ID",
+            "node_id": "Node ID inside the workflow graph",
+            "output_name": "Declared output name as exposed by Composer",
+        }
+    )
+    @console_ns.response(404, "Workflow run / node / output not found")
+    @setup_required
+    @login_required
+    @account_initialization_required
+    @get_app_model(mode=[AppMode.ADVANCED_CHAT, AppMode.WORKFLOW])
+    def get(self, app_model: App, run_id: UUID, node_id: str, output_name: str):
+        return _serve_output_preview(app_model, run_id, node_id, output_name)
+
+
+# ──────────────────────────────────────────────────────────────────────────────
+# SSE event stream — shared generator used by draft + published variants
+# ──────────────────────────────────────────────────────────────────────────────
+
+
+def _sse_envelope(event: str, data: dict | str, event_id: int) -> str:
+    """Format one SSE record per D-5 ``{event, data, id}`` envelope.
+
+    ``data`` is JSON-serialized when given as a dict; raw strings are
+    forwarded unchanged so we can also emit ``:keepalive`` comment lines.
+    """
+    payload = data if isinstance(data, str) else json.dumps(data, ensure_ascii=False)
+    return f"event: {event}\nid: {event_id}\ndata: {payload}\n\n"
+
+
+def _stream_inspector_events(app_model: App, run_id: UUID) -> Iterator[str]:
+    """Yield SSE-framed strings for one workflow run.
+
+    The stream begins with a full ``snapshot`` event so the client has a
+    starting state without needing a separate REST GET. Then for every
+    ``node_changed`` message from the pub/sub channel we re-read that node
+    from DB and push a fresh ``node_changed`` event. When the workflow run
+    reaches a terminal state we push one final ``workflow_run_completed``
+    event and close the stream.
+
+    Failures inside the loop are caught and surfaced as ``error`` events so
+    the frontend can show a banner rather than seeing the connection drop
+    silently. The Inspector never raises across the SSE boundary.
+    """
+    service = _service()
+    run_id_str = str(run_id)
+
+    # Initial snapshot — also flushes a 404 back at the client right away
+    # if the run is gone (raised before yielding any bytes, so Flask turns it
+    # into the normal HTTP 404 path).
+    try:
+        snapshot = service.snapshot_workflow_run(app_model=app_model, workflow_run_id=run_id_str)
+    except NodeOutputInspectorError as error:
+        raise _InspectorNotFound(error) from error
+
+    event_id = 0
+    yield _sse_envelope("snapshot", snapshot.model_dump(mode="json"), event_id)
+
+    # If the run already finished by the time the client connected, emit
+    # the terminal envelope synchronously and close — no point subscribing.
+    # The enum value for partial success is the hyphenated ``partial-succeeded``
+    # (graphon.enums.WorkflowExecutionStatus), not ``partial_succeeded``.
+    if snapshot.workflow_run_status.value in {"succeeded", "failed", "stopped", "partial-succeeded"}:
+        event_id += 1
+        yield _sse_envelope(
+            "workflow_run_completed",
+            {"workflow_run_id": run_id_str, "workflow_run_status": snapshot.workflow_run_status.value},
+            event_id,
+        )
+        return
+
+    # Live subscription
+    ticks_since_heartbeat = 0
+    total_ticks = 0
+    for message in inspector_events.subscribe(run_id_str, timeout_seconds=1.0):
+        total_ticks += 1
+        if total_ticks > _STREAM_HARD_TIMEOUT_TICKS:
+            logger.warning(
+                "Inspector SSE: forcing close after %ds without terminal event for run %s",
+                _STREAM_HARD_TIMEOUT_TICKS,
+                run_id_str,
+            )
+            return
+
+        # Heartbeat sentinel — ``inspector_events.subscribe`` synthesizes a
+        # ``node_changed`` message with both fields ``None`` on every redis
+        # timeout. Real ``workflow_completed`` messages keep their kind even
+        # when status couldn't be resolved (publisher race), so checking kind
+        # first makes the heartbeat branch safe.
+        if message.kind == "node_changed" and message.node_id is None and message.status is None:
+            ticks_since_heartbeat += 1
+            if ticks_since_heartbeat >= _HEARTBEAT_EVERY_TICKS:
+                yield ":keepalive\n\n"
+                ticks_since_heartbeat = 0
+            continue
+        ticks_since_heartbeat = 0
+
+        if message.kind == "workflow_completed":
+            event_id += 1
+            yield _sse_envelope(
+                "workflow_run_completed",
+                {"workflow_run_id": run_id_str, "workflow_run_status": message.status or "unknown"},
+                event_id,
+            )
+            return
+
+        # node_changed: recompute the node slice from DB
+        if not message.node_id:
+            continue
+        try:
+            node_view = service.node_detail(
+                app_model=app_model,
+                workflow_run_id=run_id_str,
+                node_id=message.node_id,
+            )
+        except NodeOutputInspectorError:
+            # Node may not appear in the graph yet (race with persistence); skip.
+            continue
+        except Exception:
+            logger.warning(
+                "Inspector SSE: node_detail failed for run %s node %s",
+                run_id_str,
+                message.node_id,
+                exc_info=True,
+            )
+            event_id += 1
+            yield _sse_envelope(
+                "error",
+                {"node_id": message.node_id, "message": "failed to refresh node detail"},
+                event_id,
+            )
+            continue
+
+        event_id += 1
+        yield _sse_envelope("node_changed", node_view.model_dump(mode="json"), event_id)
+
+
+@console_ns.route("/apps/<uuid:app_id>/workflows/draft/runs/<uuid:run_id>/node-outputs/events")
+class WorkflowDraftRunNodeOutputEventsApi(Resource):
+    """SSE stream of inspector deltas for a draft run."""
+
+    @console_ns.doc("stream_workflow_draft_run_node_output_events")
+    @console_ns.doc(description="Server-Sent Events stream of inspector deltas for a draft workflow run.")
+    @console_ns.doc(params={"app_id": "Application ID", "run_id": "Workflow run ID"})
+    @console_ns.response(404, "Workflow run not found")
+    @setup_required
+    @login_required
+    @account_initialization_required
+    @get_app_model(mode=[AppMode.ADVANCED_CHAT, AppMode.WORKFLOW])
+    def get(self, app_model: App, run_id: UUID):
+        return Response(
+            _stream_inspector_events(app_model, run_id),
+            mimetype="text/event-stream",
+            headers={"Cache-Control": "no-cache", "Connection": "keep-alive"},
+        )
+
+
+# ──────────────────────────────────────────────────────────────────────────────
+# Published-run endpoints — symmetric to the draft trio above
+# ──────────────────────────────────────────────────────────────────────────────
+
+
+@console_ns.route("/apps/<uuid:app_id>/workflows/published/runs/<uuid:run_id>/node-outputs")
+class WorkflowPublishedRunNodeOutputsApi(Resource):
+    """Whole-run snapshot for a *published* workflow run.
+
+    Same response shape as the ``/draft/`` variant — frontend can multiplex
+    based on which page (Composer test-run vs. Run History) is mounted.
+    """
+
+    @console_ns.doc("get_workflow_published_run_node_outputs")
+    @console_ns.doc(description="Snapshot of every node's declared outputs for a published workflow run.")
+    @console_ns.doc(params={"app_id": "Application ID", "run_id": "Workflow run ID"})
+    @console_ns.response(404, "Workflow run not found")
+    @setup_required
+    @login_required
+    @account_initialization_required
+    @get_app_model(mode=[AppMode.ADVANCED_CHAT, AppMode.WORKFLOW])
+    def get(self, app_model: App, run_id: UUID):
+        return _serve_snapshot(app_model, run_id)
+
+
+@console_ns.route("/apps/<uuid:app_id>/workflows/published/runs/<uuid:run_id>/node-outputs/<string:node_id>")
+class WorkflowPublishedRunNodeOutputDetailApi(Resource):
+    """One node's declared outputs + per-output status (published run)."""
+
+    @console_ns.doc("get_workflow_published_run_node_output_detail")
+    @console_ns.doc(description="One node's declared outputs for a published workflow run.")
+    @console_ns.doc(
+        params={
+            "app_id": "Application ID",
+            "run_id": "Workflow run ID",
+            "node_id": "Node ID inside the workflow graph",
+        }
+    )
+    @console_ns.response(404, "Workflow run / node not found")
+    @setup_required
+    @login_required
+    @account_initialization_required
+    @get_app_model(mode=[AppMode.ADVANCED_CHAT, AppMode.WORKFLOW])
+    def get(self, app_model: App, run_id: UUID, node_id: str):
+        return _serve_node_detail(app_model, run_id, node_id)
+
+
+@console_ns.route(
+    "/apps/<uuid:app_id>/workflows/published/runs/<uuid:run_id>"
+    "/node-outputs/<string:node_id>/<string:output_name>/preview"
+)
+class WorkflowPublishedRunNodeOutputPreviewApi(Resource):
+    """Full value for one declared output of a published run."""
+
+    @console_ns.doc("get_workflow_published_run_node_output_preview")
+    @console_ns.doc(description="Full value for one declared output of a published run.")
+    @console_ns.doc(
+        params={
+            "app_id": "Application ID",
+            "run_id": "Workflow run ID",
+            "node_id": "Node ID inside the workflow graph",
+            "output_name": "Declared output name as exposed by Composer",
+        }
+    )
+    @console_ns.response(404, "Workflow run / node / output not found")
+    @setup_required
+    @login_required
+    @account_initialization_required
+    @get_app_model(mode=[AppMode.ADVANCED_CHAT, AppMode.WORKFLOW])
+    def get(self, app_model: App, run_id: UUID, node_id: str, output_name: str):
+        return _serve_output_preview(app_model, run_id, node_id, output_name)
+
+
+@console_ns.route("/apps/<uuid:app_id>/workflows/published/runs/<uuid:run_id>/node-outputs/events")
+class WorkflowPublishedRunNodeOutputEventsApi(Resource):
+    """SSE stream of inspector deltas for a published run."""
+
+    @console_ns.doc("stream_workflow_published_run_node_output_events")
+    @console_ns.doc(description="Server-Sent Events stream of inspector deltas for a published workflow run.")
+    @console_ns.doc(params={"app_id": "Application ID", "run_id": "Workflow run ID"})
+    @console_ns.response(404, "Workflow run not found")
+    @setup_required
+    @login_required
+    @account_initialization_required
+    @get_app_model(mode=[AppMode.ADVANCED_CHAT, AppMode.WORKFLOW])
+    def get(self, app_model: App, run_id: UUID):
+        return Response(
+            _stream_inspector_events(app_model, run_id),
+            mimetype="text/event-stream",
+            headers={"Cache-Control": "no-cache", "Connection": "keep-alive"},
+        )

api/controllers/console/app/workflow_statistic.py

@@ -11,7 +11,7 @@ from extensions.ext_database import db
 from libs.datetime_utils import parse_time_range
 from libs.login import current_account_with_tenant, login_required
 from models.enums import WorkflowRunTriggeredFrom
-from models.model import AppMode
+from models.model import App, AppMode
 from repositories.factory import DifyAPIRepositoryFactory
 
 
@@ -46,7 +46,7 @@ class WorkflowDailyRunsStatistic(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def get(self, app_model):
+    def get(self, app_model: App):
         account, _ = current_account_with_tenant()
 
         args = WorkflowStatisticQuery.model_validate(request.args.to_dict(flat=True))
@@ -86,7 +86,7 @@ class WorkflowDailyTerminalsStatistic(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def get(self, app_model):
+    def get(self, app_model: App):
         account, _ = current_account_with_tenant()
 
         args = WorkflowStatisticQuery.model_validate(request.args.to_dict(flat=True))
@@ -126,7 +126,7 @@ class WorkflowDailyTokenCostStatistic(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def get(self, app_model):
+    def get(self, app_model: App):
         account, _ = current_account_with_tenant()
 
         args = WorkflowStatisticQuery.model_validate(request.args.to_dict(flat=True))
@@ -166,7 +166,7 @@ class WorkflowAverageAppInteractionStatistic(Resource):
     @login_required
     @account_initialization_required
     @get_app_model(mode=[AppMode.WORKFLOW])
-    def get(self, app_model):
+    def get(self, app_model: App):
         account, _ = current_account_with_tenant()
 
         args = WorkflowStatisticQuery.model_validate(request.args.to_dict(flat=True))

api/controllers/console/auth/data_source_bearer_auth.py

@@ -5,12 +5,12 @@ from pydantic import BaseModel, Field
 
 from controllers.common.schema import register_response_schema_models, register_schema_models
 from fields.base import ResponseModel
-from libs.login import current_account_with_tenant, login_required
+from libs.login import login_required
 from services.auth.api_key_auth_service import ApiKeyAuthService
 
 from .. import console_ns
 from ..auth.error import ApiKeyAuthFailedError
-from ..wraps import account_initialization_required, is_admin_or_owner_required, setup_required
+from ..wraps import account_initialization_required, is_admin_or_owner_required, setup_required, with_current_tenant_id
 
 
 class ApiKeyAuthBindingPayload(BaseModel):
@@ -42,8 +42,8 @@ class ApiKeyAuthDataSource(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def get(self):
-        _, current_tenant_id = current_account_with_tenant()
+    @with_current_tenant_id
+    def get(self, current_tenant_id: str):
         data_source_api_key_bindings = ApiKeyAuthService.get_provider_auth_list(current_tenant_id)
         if data_source_api_key_bindings:
             return {
@@ -69,9 +69,9 @@ class ApiKeyAuthDataSourceBinding(Resource):
     @account_initialization_required
     @is_admin_or_owner_required
     @console_ns.expect(console_ns.models[ApiKeyAuthBindingPayload.__name__])
-    def post(self):
+    @with_current_tenant_id
+    def post(self, current_tenant_id: str):
         # The role of the current user in the table must be admin or owner
-        _, current_tenant_id = current_account_with_tenant()
         payload = ApiKeyAuthBindingPayload.model_validate(console_ns.payload)
         data = payload.model_dump()
         ApiKeyAuthService.validate_api_key_auth_args(data)
@@ -89,10 +89,9 @@ class ApiKeyAuthDataSourceBindingDelete(Resource):
     @account_initialization_required
     @is_admin_or_owner_required
     @console_ns.response(204, "Binding deleted successfully")
-    def delete(self, binding_id: UUID):
+    @with_current_tenant_id
+    def delete(self, current_tenant_id: str, binding_id: UUID):
         # The role of the current user in the table must be admin or owner
-        _, current_tenant_id = current_account_with_tenant()
-
         ApiKeyAuthService.delete_provider_auth(current_tenant_id, str(binding_id))
 
         return "", 204

api/controllers/console/auth/oauth_server.py

@@ -8,9 +8,9 @@ from flask_restx import Resource
 from pydantic import BaseModel
 from werkzeug.exceptions import BadRequest, NotFound
 
-from controllers.console.wraps import account_initialization_required, setup_required
+from controllers.console.wraps import account_initialization_required, setup_required, with_current_user
 from graphon.model_runtime.utils.encoders import jsonable_encoder
-from libs.login import current_account_with_tenant, login_required
+from libs.login import login_required
 from models import Account
 from models.model import OAuthProviderApp
 from services.oauth_server import OAUTH_ACCESS_TOKEN_EXPIRES_IN, OAuthGrantType, OAuthServerService
@@ -133,12 +133,10 @@ class OAuthServerUserAuthorizeApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
+    @with_current_user
     @oauth_server_client_id_required
-    def post(self, oauth_provider_app: OAuthProviderApp):
-        current_user, _ = current_account_with_tenant()
-        account = current_user
-        user_account_id = account.id
-
+    def post(self, oauth_provider_app: OAuthProviderApp, current_user: Account):
+        user_account_id = current_user.id
         code = OAuthServerService.sign_oauth_authorization_code(oauth_provider_app.client_id, user_account_id)
         return jsonable_encoder(
             {

api/controllers/console/datasets/datasets_segments.py

@@ -5,7 +5,7 @@ from uuid import UUID
 from flask import request
 from flask_restx import Resource, marshal
 from pydantic import BaseModel, Field
-from sqlalchemy import String, cast, func, or_, select
+from sqlalchemy import String, case, cast, func, literal, or_, select
 from sqlalchemy.dialects.postgresql import JSONB
 from werkzeug.exceptions import Forbidden, NotFound
 
@@ -169,9 +169,17 @@ class DatasetDocumentSegmentListApi(Resource):
             # Use database-specific methods for JSON array search
             if dify_config.SQLALCHEMY_DATABASE_URI_SCHEME == "postgresql":
                 # PostgreSQL: Use jsonb_array_elements_text to properly handle Unicode/Chinese text
+                # Feed the set-returning function a JSON array in every row. Filtering in
+                # the subquery is not enough because PostgreSQL can still evaluate the
+                # SRF on scalar JSON before applying the predicate.
+                keywords_jsonb = cast(DocumentSegment.keywords, JSONB)
+                keywords_array = case(
+                    (func.jsonb_typeof(keywords_jsonb) == "array", keywords_jsonb),
+                    else_=cast(literal("[]"), JSONB),
+                )
                 keywords_condition = func.array_to_string(
                     func.array(
-                        select(func.jsonb_array_elements_text(cast(DocumentSegment.keywords, JSONB)))
+                        select(func.jsonb_array_elements_text(keywords_array))
                         .correlate(DocumentSegment)
                         .scalar_subquery()
                     ),

api/controllers/console/datasets/hit_testing.py

@@ -1,15 +1,12 @@
 from __future__ import annotations
 
-from datetime import datetime
-from typing import Any
 from uuid import UUID
 
 from flask_restx import Resource
-from pydantic import Field, field_validator
 
-from controllers.common.schema import register_schema_models
-from fields.base import ResponseModel
-from libs.helper import to_timestamp
+from controllers.common.schema import register_response_schema_models, register_schema_models
+from fields.hit_testing_fields import HitTestingResponse
+from libs.helper import dump_response
 from libs.login import login_required
 
 from .. import console_ns
@@ -20,86 +17,8 @@ from ..wraps import (
     setup_required,
 )
 
-
-class HitTestingDocument(ResponseModel):
-    id: str | None = None
-    data_source_type: str | None = None
-    name: str | None = None
-    doc_type: str | None = None
-    doc_metadata: Any | None = None
-
-
-class HitTestingSegment(ResponseModel):
-    id: str | None = None
-    position: int | None = None
-    document_id: str | None = None
-    content: str | None = None
-    sign_content: str | None = None
-    answer: str | None = None
-    word_count: int | None = None
-    tokens: int | None = None
-    keywords: list[str] = Field(default_factory=list)
-    index_node_id: str | None = None
-    index_node_hash: str | None = None
-    hit_count: int | None = None
-    enabled: bool | None = None
-    disabled_at: int | None = None
-    disabled_by: str | None = None
-    status: str | None = None
-    created_by: str | None = None
-    created_at: int | None = None
-    indexing_at: int | None = None
-    completed_at: int | None = None
-    error: str | None = None
-    stopped_at: int | None = None
-    document: HitTestingDocument | None = None
-
-    @field_validator("disabled_at", "created_at", "indexing_at", "completed_at", "stopped_at", mode="before")
-    @classmethod
-    def _normalize_timestamp(cls, value: datetime | int | None) -> int | None:
-        return to_timestamp(value)
-
-
-class HitTestingChildChunk(ResponseModel):
-    id: str | None = None
-    content: str | None = None
-    position: int | None = None
-    score: float | None = None
-
-
-class HitTestingFile(ResponseModel):
-    id: str | None = None
-    name: str | None = None
-    size: int | None = None
-    extension: str | None = None
-    mime_type: str | None = None
-    source_url: str | None = None
-
-
-class HitTestingRecord(ResponseModel):
-    segment: HitTestingSegment | None = None
-    child_chunks: list[HitTestingChildChunk] = Field(default_factory=list)
-    score: float | None = None
-    tsne_position: Any | None = None
-    files: list[HitTestingFile] = Field(default_factory=list)
-    summary: str | None = None
-
-
-class HitTestingResponse(ResponseModel):
-    query: str
-    records: list[HitTestingRecord] = Field(default_factory=list)
-
-
-register_schema_models(
-    console_ns,
-    HitTestingPayload,
-    HitTestingDocument,
-    HitTestingSegment,
-    HitTestingChildChunk,
-    HitTestingFile,
-    HitTestingRecord,
-    HitTestingResponse,
-)
+register_schema_models(console_ns, HitTestingPayload)
+register_response_schema_models(console_ns, HitTestingResponse)
 
 
 @console_ns.route("/datasets/<uuid:dataset_id>/hit-testing")
@@ -119,12 +38,11 @@ class HitTestingApi(Resource, DatasetsHitTestingBase):
     @login_required
     @account_initialization_required
     @cloud_edition_billing_rate_limit_check("knowledge")
-    def post(self, dataset_id: UUID):
+    def post(self, dataset_id: UUID) -> dict[str, object]:
         dataset_id_str = str(dataset_id)
 
         dataset = self.get_and_validate_dataset(dataset_id_str)
-        payload = HitTestingPayload.model_validate(console_ns.payload or {})
-        args = payload.model_dump(exclude_none=True)
+        args = self.parse_args(console_ns.payload)
         self.hit_testing_args_check(args)
 
-        return HitTestingResponse.model_validate(self.perform_hit_testing(dataset, args)).model_dump(mode="json")
+        return dump_response(HitTestingResponse, self.perform_hit_testing(dataset, args))

api/controllers/console/datasets/hit_testing_base.py

@@ -1,7 +1,6 @@
 import logging
-from typing import Any
+from typing import Any, cast
 
-from flask_restx import marshal
 from pydantic import BaseModel, Field
 from werkzeug.exceptions import Forbidden, InternalServerError, NotFound
 
@@ -19,10 +18,10 @@ from core.errors.error import (
     ProviderTokenNotInitError,
     QuotaExceededError,
 )
-from fields.hit_testing_fields import hit_testing_record_fields
 from graphon.model_runtime.errors.invoke import InvokeError
 from libs.login import current_user
 from models.account import Account
+from models.dataset import Dataset
 from services.dataset_service import DatasetService
 from services.entities.knowledge_entities.knowledge_entities import RetrievalModel
 from services.hit_testing_service import HitTestingService
@@ -38,16 +37,6 @@ class HitTestingPayload(BaseModel):
 
 
 class DatasetsHitTestingBase:
-    @staticmethod
-    def _extract_hit_testing_query(query: Any) -> str:
-        """Return the query string from the service response shape."""
-        if isinstance(query, dict):
-            content = query.get("content")
-            if isinstance(content, str):
-                return content
-
-        raise ValueError("Invalid hit testing query response")
-
     @staticmethod
     def _prepare_hit_testing_records(records: Any) -> list[dict[str, Any]]:
         """Ensure collection fields match the API schema before response validation."""
@@ -63,6 +52,7 @@ class DatasetsHitTestingBase:
             segment = normalized_record.get("segment")
             if isinstance(segment, dict):
                 normalized_segment = dict(segment)
+                normalized_segment.setdefault("sign_content", None)
                 if normalized_segment.get("keywords") is None:
                     normalized_segment["keywords"] = []
                 normalized_record["segment"] = normalized_segment
@@ -73,12 +63,15 @@ class DatasetsHitTestingBase:
             if normalized_record.get("files") is None:
                 normalized_record["files"] = []
 
+            normalized_record.setdefault("tsne_position", None)
+            normalized_record.setdefault("summary", None)
+
             normalized_records.append(normalized_record)
 
         return normalized_records
 
     @staticmethod
-    def get_and_validate_dataset(dataset_id: str):
+    def get_and_validate_dataset(dataset_id: str) -> Dataset:
         assert isinstance(current_user, Account)
         dataset = DatasetService.get_dataset(dataset_id)
         if dataset is None:
@@ -92,33 +85,35 @@ class DatasetsHitTestingBase:
         return dataset
 
     @staticmethod
-    def hit_testing_args_check(args: dict[str, Any]):
+    def hit_testing_args_check(args: dict[str, Any]) -> None:
         HitTestingService.hit_testing_args_check(args)
 
     @staticmethod
-    def parse_args(payload: dict[str, Any]) -> dict[str, Any]:
+    def parse_args(payload: dict[str, Any] | None) -> dict[str, Any]:
         """Validate and return hit-testing arguments from an incoming payload."""
         hit_testing_payload = HitTestingPayload.model_validate(payload or {})
         return hit_testing_payload.model_dump(exclude_none=True)
 
     @staticmethod
-    def perform_hit_testing(dataset, args):
+    def perform_hit_testing(dataset: Dataset, args: dict[str, Any]) -> dict[str, Any]:
         assert isinstance(current_user, Account)
         try:
             response = HitTestingService.retrieve(
                 dataset=dataset,
-                query=args.get("query"),
+                query=cast(str, args.get("query")),
                 account=current_user,
                 retrieval_model=args.get("retrieval_model"),
-                external_retrieval_model=args.get("external_retrieval_model"),
+                external_retrieval_model=cast(dict[str, Any], args.get("external_retrieval_model")),
                 attachment_ids=args.get("attachment_ids"),
                 limit=10,
             )
+            query = response.get("query")
+            if not isinstance(query, dict) or not isinstance(query.get("content"), str):
+                raise ValueError("Invalid hit testing query response")
+
             return {
-                "query": DatasetsHitTestingBase._extract_hit_testing_query(response.get("query")),
-                "records": DatasetsHitTestingBase._prepare_hit_testing_records(
-                    marshal(response.get("records", []), hit_testing_record_fields)
-                ),
+                "query": {"content": query["content"]},
+                "records": DatasetsHitTestingBase._prepare_hit_testing_records(response.get("records", [])),
             }
         except services.errors.index.IndexNotInitializedError:
             raise DatasetNotInitializedError()

api/controllers/console/explore/audio.py

@@ -20,6 +20,7 @@ from controllers.console.app.error import (
 from controllers.console.explore.wraps import InstalledAppResource
 from core.errors.error import ModelCurrentlyNotSupportError, ProviderTokenNotInitError, QuotaExceededError
 from graphon.model_runtime.errors.invoke import InvokeError
+from models.model import InstalledApp
 from services.audio_service import AudioService
 from services.errors.audio import (
     AudioTooLargeServiceError,
@@ -40,8 +41,10 @@ register_schema_model(console_ns, TextToAudioPayload)
     endpoint="installed_app_audio",
 )
 class ChatAudioApi(InstalledAppResource):
-    def post(self, installed_app):
+    def post(self, installed_app: InstalledApp):
         app_model = installed_app.app
+        if app_model is None:
+            raise AppUnavailableError()
 
         file = request.files["file"]
 
@@ -81,8 +84,10 @@ class ChatAudioApi(InstalledAppResource):
 )
 class ChatTextApi(InstalledAppResource):
     @console_ns.expect(console_ns.models[TextToAudioPayload.__name__])
-    def post(self, installed_app):
+    def post(self, installed_app: InstalledApp):
         app_model = installed_app.app
+        if app_model is None:
+            raise AppUnavailableError()
         try:
             payload = TextToAudioPayload.model_validate(console_ns.payload or {})
 

api/controllers/console/explore/completion.py

@@ -31,7 +31,7 @@ from libs import helper
 from libs.datetime_utils import naive_utc_now
 from libs.login import current_user
 from models import Account
-from models.model import AppMode
+from models.model import AppMode, InstalledApp
 from services.app_generate_service import AppGenerateService
 from services.app_task_service import AppTaskService
 from services.errors.llm import InvokeRateLimitError
@@ -83,8 +83,10 @@ register_response_schema_models(console_ns, SimpleResultResponse)
 )
 class CompletionApi(InstalledAppResource):
     @console_ns.expect(console_ns.models[CompletionMessageExplorePayload.__name__])
-    def post(self, installed_app):
+    def post(self, installed_app: InstalledApp):
         app_model = installed_app.app
+        if app_model is None:
+            raise AppUnavailableError()
         if app_model.mode != AppMode.COMPLETION:
             raise NotCompletionAppError()
 
@@ -133,8 +135,10 @@ class CompletionApi(InstalledAppResource):
 )
 class CompletionStopApi(InstalledAppResource):
     @console_ns.response(200, "Success", console_ns.models[SimpleResultResponse.__name__])
-    def post(self, installed_app, task_id: str):
+    def post(self, installed_app: InstalledApp, task_id: str):
         app_model = installed_app.app
+        if app_model is None:
+            raise AppUnavailableError()
         if app_model.mode != AppMode.COMPLETION:
             raise NotCompletionAppError()
 
@@ -157,8 +161,10 @@ class CompletionStopApi(InstalledAppResource):
 )
 class ChatApi(InstalledAppResource):
     @console_ns.expect(console_ns.models[ChatMessagePayload.__name__])
-    def post(self, installed_app):
+    def post(self, installed_app: InstalledApp):
         app_model = installed_app.app
+        if app_model is None:
+            raise AppUnavailableError()
         app_mode = AppMode.value_of(app_model.mode)
         if app_mode not in {AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT}:
             raise NotChatAppError()
@@ -209,8 +215,10 @@ class ChatApi(InstalledAppResource):
 )
 class ChatStopApi(InstalledAppResource):
     @console_ns.response(200, "Success", console_ns.models[SimpleResultResponse.__name__])
-    def post(self, installed_app, task_id: str):
+    def post(self, installed_app: InstalledApp, task_id: str):
         app_model = installed_app.app
+        if app_model is None:
+            raise AppUnavailableError()
         app_mode = AppMode.value_of(app_model.mode)
         if app_mode not in {AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT}:
             raise NotChatAppError()

api/controllers/console/explore/conversation.py

@@ -8,6 +8,7 @@ from werkzeug.exceptions import NotFound
 
 from controllers.common.controller_schemas import ConversationRenamePayload
 from controllers.common.schema import register_response_schema_models, register_schema_models
+from controllers.console.app.error import AppUnavailableError
 from controllers.console.explore.error import NotChatAppError
 from controllers.console.explore.wraps import InstalledAppResource
 from core.app.entities.app_invoke_entities import InvokeFrom
@@ -20,7 +21,7 @@ from fields.conversation_fields import (
 from libs.helper import UUIDStrOrEmpty
 from libs.login import current_user
 from models import Account
-from models.model import AppMode
+from models.model import AppMode, InstalledApp
 from services.conversation_service import ConversationService
 from services.errors.conversation import ConversationNotExistsError, LastConversationNotExistsError
 from services.web_conversation_service import WebConversationService
@@ -44,8 +45,10 @@ register_response_schema_models(console_ns, ResultResponse)
 )
 class ConversationListApi(InstalledAppResource):
     @console_ns.expect(console_ns.models[ConversationListQuery.__name__])
-    def get(self, installed_app):
+    def get(self, installed_app: InstalledApp):
         app_model = installed_app.app
+        if app_model is None:
+            raise AppUnavailableError()
         app_mode = AppMode.value_of(app_model.mode)
         if app_mode not in {AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT}:
             raise NotChatAppError()
@@ -92,8 +95,10 @@ class ConversationListApi(InstalledAppResource):
 )
 class ConversationApi(InstalledAppResource):
     @console_ns.response(204, "Conversation deleted successfully")
-    def delete(self, installed_app, c_id: UUID):
+    def delete(self, installed_app: InstalledApp, c_id: UUID):
         app_model = installed_app.app
+        if app_model is None:
+            raise AppUnavailableError()
         app_mode = AppMode.value_of(app_model.mode)
         if app_mode not in {AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT}:
             raise NotChatAppError()
@@ -115,8 +120,10 @@ class ConversationApi(InstalledAppResource):
 )
 class ConversationRenameApi(InstalledAppResource):
     @console_ns.expect(console_ns.models[ConversationRenamePayload.__name__])
-    def post(self, installed_app, c_id: UUID):
+    def post(self, installed_app: InstalledApp, c_id: UUID):
         app_model = installed_app.app
+        if app_model is None:
+            raise AppUnavailableError()
         app_mode = AppMode.value_of(app_model.mode)
         if app_mode not in {AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT}:
             raise NotChatAppError()
@@ -146,8 +153,10 @@ class ConversationRenameApi(InstalledAppResource):
 )
 class ConversationPinApi(InstalledAppResource):
     @console_ns.response(200, "Success", console_ns.models[ResultResponse.__name__])
-    def patch(self, installed_app, c_id: UUID):
+    def patch(self, installed_app: InstalledApp, c_id: UUID):
         app_model = installed_app.app
+        if app_model is None:
+            raise AppUnavailableError()
         app_mode = AppMode.value_of(app_model.mode)
         if app_mode not in {AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT}:
             raise NotChatAppError()
@@ -170,8 +179,10 @@ class ConversationPinApi(InstalledAppResource):
 )
 class ConversationUnPinApi(InstalledAppResource):
     @console_ns.response(200, "Success", console_ns.models[ResultResponse.__name__])
-    def patch(self, installed_app, c_id: UUID):
+    def patch(self, installed_app: InstalledApp, c_id: UUID):
         app_model = installed_app.app
+        if app_model is None:
+            raise AppUnavailableError()
         app_mode = AppMode.value_of(app_model.mode)
         if app_mode not in {AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT}:
             raise NotChatAppError()

api/controllers/console/explore/installed_app.py

@@ -149,19 +149,28 @@ class InstalledAppsListApi(Resource):
         if current_user.current_tenant is None:
             raise ValueError("current_user.current_tenant must not be None")
         current_user.role = TenantService.get_user_role(current_user, current_user.current_tenant)
-        installed_app_list: list[dict[str, Any]] = [
-            {
-                "id": installed_app.id,
-                "app": installed_app.app,
-                "app_owner_tenant_id": installed_app.app_owner_tenant_id,
-                "is_pinned": installed_app.is_pinned,
-                "last_used_at": installed_app.last_used_at,
-                "editable": current_user.role in {"owner", "admin"},
-                "uninstallable": current_tenant_id == installed_app.app_owner_tenant_id,
-            }
-            for installed_app in installed_apps
-            if installed_app.app is not None
-        ]
+
+        app_ids = [installed_app.app_id for installed_app in installed_apps]
+        apps = db.session.scalars(select(App).where(App.id.in_(app_ids))).all() if app_ids else []
+        apps_by_id = {app.id: app for app in apps}
+
+        installed_app_list: list[dict[str, Any]] = []
+        for installed_app in installed_apps:
+            app_model = apps_by_id.get(installed_app.app_id)
+            if app_model is None:
+                continue
+
+            installed_app_list.append(
+                {
+                    "id": installed_app.id,
+                    "app": app_model,
+                    "app_owner_tenant_id": installed_app.app_owner_tenant_id,
+                    "is_pinned": installed_app.is_pinned,
+                    "last_used_at": installed_app.last_used_at,
+                    "editable": current_user.role in {"owner", "admin"},
+                    "uninstallable": current_tenant_id == installed_app.app_owner_tenant_id,
+                }
+            )
 
         # filter out apps that user doesn't have access to
         if FeatureService.get_system_features().webapp_auth.enabled:
@@ -262,7 +271,7 @@ class InstalledAppApi(InstalledAppResource):
     """
 
     @console_ns.response(204, "App uninstalled successfully")
-    def delete(self, installed_app):
+    def delete(self, installed_app: InstalledApp):
         _, current_tenant_id = current_account_with_tenant()
         if installed_app.app_owner_tenant_id == current_tenant_id:
             raise BadRequest("You can't uninstall an app owned by the current tenant")
@@ -273,7 +282,7 @@ class InstalledAppApi(InstalledAppResource):
         return "", 204
 
     @console_ns.response(200, "Success", console_ns.models[SimpleResultMessageResponse.__name__])
-    def patch(self, installed_app):
+    def patch(self, installed_app: InstalledApp):
         payload = InstalledAppUpdatePayload.model_validate(console_ns.payload or {})
 
         commit_args = False

api/controllers/console/explore/message.py

@@ -10,6 +10,7 @@ from controllers.common.controller_schemas import MessageFeedbackPayload, Messag
 from controllers.common.schema import register_response_schema_models, register_schema_models
 from controllers.console.app.error import (
     AppMoreLikeThisDisabledError,
+    AppUnavailableError,
     CompletionRequestError,
     ProviderModelCurrentlyNotSupportError,
     ProviderNotInitializeError,
@@ -30,7 +31,7 @@ from graphon.model_runtime.errors.invoke import InvokeError
 from libs import helper
 from models import Account
 from models.enums import FeedbackRating
-from models.model import AppMode
+from models.model import AppMode, InstalledApp
 from services.app_generate_service import AppGenerateService
 from services.errors.app import MoreLikeThisDisabledError
 from services.errors.conversation import ConversationNotExistsError
@@ -61,8 +62,10 @@ register_response_schema_models(console_ns, ResultResponse, SuggestedQuestionsRe
 class MessageListApi(InstalledAppResource):
     @console_ns.expect(console_ns.models[MessageListQuery.__name__])
     @with_current_user
-    def get(self, current_user: Account, installed_app):
+    def get(self, current_user: Account, installed_app: InstalledApp):
         app_model = installed_app.app
+        if app_model is None:
+            raise AppUnavailableError()
 
         app_mode = AppMode.value_of(app_model.mode)
         if app_mode not in {AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT}:
@@ -98,8 +101,10 @@ class MessageFeedbackApi(InstalledAppResource):
     @console_ns.expect(console_ns.models[MessageFeedbackPayload.__name__])
     @console_ns.response(200, "Feedback submitted successfully", console_ns.models[ResultResponse.__name__])
     @with_current_user
-    def post(self, current_user: Account, installed_app, message_id: UUID):
+    def post(self, current_user: Account, installed_app: InstalledApp, message_id: UUID):
         app_model = installed_app.app
+        if app_model is None:
+            raise AppUnavailableError()
 
         message_id_str = str(message_id)
 
@@ -126,8 +131,10 @@ class MessageFeedbackApi(InstalledAppResource):
 class MessageMoreLikeThisApi(InstalledAppResource):
     @console_ns.expect(console_ns.models[MoreLikeThisQuery.__name__])
     @with_current_user
-    def get(self, current_user: Account, installed_app, message_id: UUID):
+    def get(self, current_user: Account, installed_app: InstalledApp, message_id: UUID):
         app_model = installed_app.app
+        if app_model is None:
+            raise AppUnavailableError()
         if app_model.mode != "completion":
             raise NotCompletionAppError()
 
@@ -172,8 +179,10 @@ class MessageMoreLikeThisApi(InstalledAppResource):
 class MessageSuggestedQuestionApi(InstalledAppResource):
     @console_ns.response(200, "Success", console_ns.models[SuggestedQuestionsResponse.__name__])
     @with_current_user
-    def get(self, current_user: Account, installed_app, message_id: UUID):
+    def get(self, current_user: Account, installed_app: InstalledApp, message_id: UUID):
         app_model = installed_app.app
+        if app_model is None:
+            raise AppUnavailableError()
         app_mode = AppMode.value_of(app_model.mode)
         if app_mode not in {AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT}:
             raise NotChatAppError()

api/controllers/console/explore/recommended_app.py

@@ -64,15 +64,28 @@ class RecommendedAppListResponse(ResponseModel):
     categories: list[str]
 
 
+class LearnDifyAppListResponse(ResponseModel):
+    recommended_apps: list[RecommendedAppResponse]
+
+
 register_schema_models(
     console_ns,
     RecommendedAppsQuery,
     RecommendedAppInfoResponse,
     RecommendedAppResponse,
     RecommendedAppListResponse,
+    LearnDifyAppListResponse,
 )
 
 
+def _resolve_language(language: str | None) -> str:
+    if language and language in languages:
+        return language
+    if current_user and current_user.interface_language:
+        return current_user.interface_language
+    return languages[0]
+
+
 @console_ns.route("/explore/apps")
 class RecommendedAppListApi(Resource):
     @console_ns.doc(params=query_params_from_model(RecommendedAppsQuery))
@@ -82,13 +95,7 @@ class RecommendedAppListApi(Resource):
     def get(self):
         # language args
         args = RecommendedAppsQuery.model_validate(request.args.to_dict(flat=True))
-        language = args.language
-        if language and language in languages:
-            language_prefix = language
-        elif current_user and current_user.interface_language:
-            language_prefix = current_user.interface_language
-        else:
-            language_prefix = languages[0]
+        language_prefix = _resolve_language(args.language)
 
         return RecommendedAppListResponse.model_validate(
             RecommendedAppService.get_recommended_apps_and_categories(language_prefix),
@@ -96,6 +103,22 @@ class RecommendedAppListApi(Resource):
         ).model_dump(mode="json")
 
 
+@console_ns.route("/explore/apps/learn-dify")
+class LearnDifyAppListApi(Resource):
+    @console_ns.doc(params=query_params_from_model(RecommendedAppsQuery))
+    @console_ns.response(200, "Success", console_ns.models[LearnDifyAppListResponse.__name__])
+    @login_required
+    @account_initialization_required
+    def get(self):
+        args = RecommendedAppsQuery.model_validate(request.args.to_dict(flat=True))
+        language_prefix = _resolve_language(args.language)
+
+        return LearnDifyAppListResponse.model_validate(
+            RecommendedAppService.get_learn_dify_apps(language_prefix),
+            from_attributes=True,
+        ).model_dump(mode="json")
+
+
 @console_ns.route("/explore/apps/<uuid:app_id>")
 class RecommendedAppApi(Resource):
     @login_required

api/controllers/console/explore/saved_message.py

@@ -7,12 +7,14 @@ from werkzeug.exceptions import NotFound
 from controllers.common.controller_schemas import SavedMessageCreatePayload, SavedMessageListQuery
 from controllers.common.schema import register_response_schema_models, register_schema_models
 from controllers.console import console_ns
+from controllers.console.app.error import AppUnavailableError
 from controllers.console.explore.error import NotCompletionAppError
 from controllers.console.explore.wraps import InstalledAppResource
 from controllers.console.wraps import with_current_user
 from fields.conversation_fields import ResultResponse
 from fields.message_fields import SavedMessageInfiniteScrollPagination, SavedMessageItem
 from models import Account
+from models.model import InstalledApp
 from services.errors.message import MessageNotExistsError
 from services.saved_message_service import SavedMessageService
 
@@ -24,8 +26,10 @@ register_response_schema_models(console_ns, ResultResponse)
 class SavedMessageListApi(InstalledAppResource):
     @console_ns.expect(console_ns.models[SavedMessageListQuery.__name__])
     @with_current_user
-    def get(self, current_user: Account, installed_app):
+    def get(self, current_user: Account, installed_app: InstalledApp):
         app_model = installed_app.app
+        if app_model is None:
+            raise AppUnavailableError()
         if app_model.mode != "completion":
             raise NotCompletionAppError()
 
@@ -48,8 +52,10 @@ class SavedMessageListApi(InstalledAppResource):
     @console_ns.expect(console_ns.models[SavedMessageCreatePayload.__name__])
     @console_ns.response(200, "Success", console_ns.models[ResultResponse.__name__])
     @with_current_user
-    def post(self, current_user: Account, installed_app):
+    def post(self, current_user: Account, installed_app: InstalledApp):
         app_model = installed_app.app
+        if app_model is None:
+            raise AppUnavailableError()
         if app_model.mode != "completion":
             raise NotCompletionAppError()
 
@@ -69,8 +75,10 @@ class SavedMessageListApi(InstalledAppResource):
 class SavedMessageApi(InstalledAppResource):
     @console_ns.response(204, "Saved message deleted successfully")
     @with_current_user
-    def delete(self, current_user: Account, installed_app, message_id: UUID):
+    def delete(self, current_user: Account, installed_app: InstalledApp, message_id: UUID):
         app_model = installed_app.app
+        if app_model is None:
+            raise AppUnavailableError()
 
         message_id_str = str(message_id)
 

api/controllers/console/files.py

@@ -28,7 +28,7 @@ from controllers.console.wraps import (
 from extensions.ext_database import db
 from fields.file_fields import FileResponse, UploadConfig
 from libs.login import login_required
-from models.account import Account
+from models import Account
 from services.file_service import FileService
 
 from . import console_ns

api/controllers/console/remote_files.py

@@ -18,7 +18,7 @@ from extensions.ext_database import db
 from fields.file_fields import FileWithSignedUrl, RemoteFileInfo
 from graphon.file import helpers as file_helpers
 from libs.login import login_required
-from models.account import Account
+from models import Account
 from services.file_service import FileService
 
 

api/controllers/console/tag/tags.py

@@ -9,9 +9,16 @@ from werkzeug.exceptions import Forbidden
 from controllers.common.fields import SimpleResultResponse
 from controllers.common.schema import register_response_schema_models, register_schema_models
 from controllers.console import console_ns
-from controllers.console.wraps import account_initialization_required, edit_permission_required, setup_required
+from controllers.console.wraps import (
+    account_initialization_required,
+    edit_permission_required,
+    setup_required,
+    with_current_tenant_id,
+    with_current_user,
+)
 from fields.base import ResponseModel
-from libs.login import current_account_with_tenant, login_required
+from libs.login import login_required
+from models import Account
 from models.enums import TagType
 from services.tag_service import (
     SaveTagPayload,
@@ -92,8 +99,8 @@ class TagListApi(Resource):
         params={"type": 'Tag type filter. Can be "knowledge" or "app".', "keyword": "Search keyword for tag name."}
     )
     @console_ns.doc(responses={200: ("Success", [console_ns.models[TagResponse.__name__]])})
-    def get(self):
-        _, current_tenant_id = current_account_with_tenant()
+    @with_current_tenant_id
+    def get(self, current_tenant_id: str):
         raw_args = request.args.to_dict()
         param = TagListQueryParam.model_validate(raw_args)
         tags = TagService.get_tags(param.type, current_tenant_id, param.keyword)
@@ -109,9 +116,9 @@ class TagListApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def post(self):
-        current_user, _ = current_account_with_tenant()
-        # The role of the current user in the ta table must be admin, owner, or editor
+    @with_current_user
+    def post(self, current_user: Account):
+        # Allow users with edit permission, or dataset editors (including dataset operators).
         if not (current_user.has_edit_permission or current_user.is_dataset_editor):
             raise Forbidden()
 
@@ -132,8 +139,8 @@ class TagUpdateDeleteApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def patch(self, tag_id: UUID):
-        current_user, _ = current_account_with_tenant()
+    @with_current_user
+    def patch(self, current_user: Account, tag_id: UUID):
         tag_id_str = str(tag_id)
         # The role of the current user in the ta table must be admin, owner, or editor
         if not (current_user.has_edit_permission or current_user.is_dataset_editor):
@@ -163,20 +170,19 @@ class TagUpdateDeleteApi(Resource):
         return "", 204
 
 
-def _require_tag_binding_edit_permission() -> None:
+def _require_tag_binding_edit_permission(current_user: Account) -> None:
     """
     Ensure the current account can edit tag bindings.
 
     Tag binding operations are allowed for users who can edit resources (app/dataset) within the current tenant.
     """
-    current_user, _ = current_account_with_tenant()
     # The role of the current user in the ta table must be admin, owner, editor, or dataset_operator
     if not (current_user.has_edit_permission or current_user.is_dataset_editor):
         raise Forbidden()
 
 
-def _create_tag_bindings() -> tuple[dict[str, str], int]:
-    _require_tag_binding_edit_permission()
+def _create_tag_bindings(current_user: Account) -> tuple[dict[str, str], int]:
+    _require_tag_binding_edit_permission(current_user)
 
     payload = TagBindingPayload.model_validate(console_ns.payload or {})
     TagService.save_tag_binding(
@@ -189,8 +195,8 @@ def _create_tag_bindings() -> tuple[dict[str, str], int]:
     return {"result": "success"}, 200
 
 
-def _remove_tag_bindings() -> tuple[dict[str, str], int]:
-    _require_tag_binding_edit_permission()
+def _remove_tag_bindings(current_user: Account) -> tuple[dict[str, str], int]:
+    _require_tag_binding_edit_permission(current_user)
 
     payload = TagBindingRemovePayload.model_validate(console_ns.payload or {})
     TagService.delete_tag_binding(
@@ -213,8 +219,9 @@ class TagBindingCollectionApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def post(self):
-        return _create_tag_bindings()
+    @with_current_user
+    def post(self, current_user: Account):
+        return _create_tag_bindings(current_user)
 
 
 @console_ns.route("/tag-bindings/remove")
@@ -228,5 +235,6 @@ class TagBindingRemoveApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def post(self):
-        return _remove_tag_bindings()
+    @with_current_user
+    def post(self, current_user: Account):
+        return _remove_tag_bindings(current_user)

api/controllers/console/workspace/members.py

@@ -77,7 +77,7 @@ register_response_schema_models(console_ns, SimpleResultDataResponse, Verificati
 def _is_role_enabled(role: TenantAccountRole | str, tenant_id: str) -> bool:
     if role != TenantAccountRole.DATASET_OPERATOR:
         return True
-    return FeatureService.get_features(tenant_id=tenant_id).dataset_operator_enabled
+    return FeatureService.get_features(tenant_id=tenant_id, exclude_vector_space=True).dataset_operator_enabled
 
 
 def _normalize_invitee_emails(emails: list[str]) -> list[str]:
@@ -113,7 +113,7 @@ def _check_member_invite_limits(tenant_id: str, new_member_count: int) -> None:
     if new_member_count <= 0:
         return
 
-    features = FeatureService.get_features(tenant_id=tenant_id)
+    features = FeatureService.get_features(tenant_id=tenant_id, exclude_vector_space=True)
 
     if dify_config.ENTERPRISE_ENABLED:
         workspace_members = features.workspace_members

api/controllers/console/workspace/models.py

@@ -8,12 +8,17 @@ from pydantic import BaseModel, Field, field_validator
 from controllers.common.fields import SimpleResultResponse
 from controllers.common.schema import register_enum_models, register_response_schema_models, register_schema_models
 from controllers.console import console_ns
-from controllers.console.wraps import account_initialization_required, is_admin_or_owner_required, setup_required
+from controllers.console.wraps import (
+    account_initialization_required,
+    is_admin_or_owner_required,
+    setup_required,
+    with_current_tenant_id,
+)
 from graphon.model_runtime.entities.model_entities import ModelType
 from graphon.model_runtime.errors.validate import CredentialsValidateFailedError
 from graphon.model_runtime.utils.encoders import jsonable_encoder
 from libs.helper import uuid_value
-from libs.login import current_account_with_tenant, login_required
+from libs.login import login_required
 from services.model_load_balancing_service import ModelLoadBalancingService
 from services.model_provider_service import ModelProviderService
 
@@ -138,9 +143,8 @@ class DefaultModelApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def get(self):
-        _, tenant_id = current_account_with_tenant()
-
+    @with_current_tenant_id
+    def get(self, tenant_id: str):
         args = ParserGetDefault.model_validate(request.args.to_dict(flat=True))
 
         model_provider_service = ModelProviderService()
@@ -156,9 +160,8 @@ class DefaultModelApi(Resource):
     @login_required
     @is_admin_or_owner_required
     @account_initialization_required
-    def post(self):
-        _, tenant_id = current_account_with_tenant()
-
+    @with_current_tenant_id
+    def post(self, tenant_id: str):
         args = ParserPostDefault.model_validate(console_ns.payload)
         model_provider_service = ModelProviderService()
         model_settings = args.model_settings
@@ -189,9 +192,8 @@ class ModelProviderModelApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def get(self, provider):
-        _, tenant_id = current_account_with_tenant()
-
+    @with_current_tenant_id
+    def get(self, tenant_id: str, provider):
         model_provider_service = ModelProviderService()
         models = model_provider_service.get_models_by_provider(tenant_id=tenant_id, provider=provider)
 
@@ -202,9 +204,9 @@ class ModelProviderModelApi(Resource):
     @login_required
     @is_admin_or_owner_required
     @account_initialization_required
-    def post(self, provider: str):
+    @with_current_tenant_id
+    def post(self, tenant_id: str, provider: str):
         # To save the model's load balance configs
-        _, tenant_id = current_account_with_tenant()
         args = ParserPostModels.model_validate(console_ns.payload)
 
         if args.config_from == "custom-model":
@@ -249,9 +251,8 @@ class ModelProviderModelApi(Resource):
     @login_required
     @is_admin_or_owner_required
     @account_initialization_required
-    def delete(self, provider: str):
-        _, tenant_id = current_account_with_tenant()
-
+    @with_current_tenant_id
+    def delete(self, tenant_id: str, provider: str):
         args = ParserDeleteModels.model_validate(console_ns.payload)
 
         model_provider_service = ModelProviderService()
@@ -268,9 +269,8 @@ class ModelProviderModelCredentialApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def get(self, provider: str):
-        _, tenant_id = current_account_with_tenant()
-
+    @with_current_tenant_id
+    def get(self, tenant_id: str, provider: str):
         args = ParserGetCredentials.model_validate(request.args.to_dict(flat=True))
 
         model_provider_service = ModelProviderService()
@@ -323,9 +323,8 @@ class ModelProviderModelCredentialApi(Resource):
     @login_required
     @is_admin_or_owner_required
     @account_initialization_required
-    def post(self, provider: str):
-        _, tenant_id = current_account_with_tenant()
-
+    @with_current_tenant_id
+    def post(self, tenant_id: str, provider: str):
         args = ParserCreateCredential.model_validate(console_ns.payload)
 
         model_provider_service = ModelProviderService()
@@ -355,8 +354,8 @@ class ModelProviderModelCredentialApi(Resource):
     @login_required
     @is_admin_or_owner_required
     @account_initialization_required
-    def put(self, provider: str):
-        _, current_tenant_id = current_account_with_tenant()
+    @with_current_tenant_id
+    def put(self, current_tenant_id: str, provider: str):
         args = ParserUpdateCredential.model_validate(console_ns.payload)
 
         model_provider_service = ModelProviderService()
@@ -382,8 +381,8 @@ class ModelProviderModelCredentialApi(Resource):
     @login_required
     @is_admin_or_owner_required
     @account_initialization_required
-    def delete(self, provider: str):
-        _, current_tenant_id = current_account_with_tenant()
+    @with_current_tenant_id
+    def delete(self, current_tenant_id: str, provider: str):
         args = ParserDeleteCredential.model_validate(console_ns.payload)
 
         model_provider_service = ModelProviderService()
@@ -406,8 +405,8 @@ class ModelProviderModelCredentialSwitchApi(Resource):
     @login_required
     @is_admin_or_owner_required
     @account_initialization_required
-    def post(self, provider: str):
-        _, current_tenant_id = current_account_with_tenant()
+    @with_current_tenant_id
+    def post(self, current_tenant_id: str, provider: str):
         args = ParserSwitch.model_validate(console_ns.payload)
 
         service = ModelProviderService()
@@ -430,9 +429,8 @@ class ModelProviderModelEnableApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def patch(self, provider: str):
-        _, tenant_id = current_account_with_tenant()
-
+    @with_current_tenant_id
+    def patch(self, tenant_id: str, provider: str):
         args = ParserDeleteModels.model_validate(console_ns.payload)
 
         model_provider_service = ModelProviderService()
@@ -452,9 +450,8 @@ class ModelProviderModelDisableApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def patch(self, provider: str):
-        _, tenant_id = current_account_with_tenant()
-
+    @with_current_tenant_id
+    def patch(self, tenant_id: str, provider: str):
         args = ParserDeleteModels.model_validate(console_ns.payload)
 
         model_provider_service = ModelProviderService()
@@ -480,8 +477,8 @@ class ModelProviderModelValidateApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def post(self, provider: str):
-        _, tenant_id = current_account_with_tenant()
+    @with_current_tenant_id
+    def post(self, tenant_id: str, provider: str):
         args = ParserValidate.model_validate(console_ns.payload)
 
         model_provider_service = ModelProviderService()
@@ -515,9 +512,9 @@ class ModelProviderModelParameterRuleApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def get(self, provider: str):
+    @with_current_tenant_id
+    def get(self, tenant_id: str, provider: str):
         args = ParserParameter.model_validate(request.args.to_dict(flat=True))
-        _, tenant_id = current_account_with_tenant()
 
         model_provider_service = ModelProviderService()
         parameter_rules = model_provider_service.get_model_parameter_rules(
@@ -532,8 +529,8 @@ class ModelProviderAvailableModelApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
-    def get(self, model_type: str):
-        _, tenant_id = current_account_with_tenant()
+    @with_current_tenant_id
+    def get(self, tenant_id: str, model_type: str):
         model_provider_service = ModelProviderService()
         models = model_provider_service.get_models_by_model_type(tenant_id=tenant_id, model_type=model_type)
 

api/controllers/console/workspace/plugin.py

@@ -1,16 +1,21 @@
 import io
 from collections.abc import Mapping
-from typing import Any, Literal
+from typing import Any, Literal, TypedDict
 
 from flask import request, send_file
 from flask_restx import Resource
-from pydantic import BaseModel, Field
+from pydantic import BaseModel, ConfigDict, Field
 from werkzeug.datastructures import FileStorage
 from werkzeug.exceptions import Forbidden
 
 from configs import dify_config
 from controllers.common.fields import SuccessResponse
-from controllers.common.schema import register_enum_models, register_response_schema_models, register_schema_models
+from controllers.common.schema import (
+    query_params_from_model,
+    register_enum_models,
+    register_response_schema_models,
+    register_schema_models,
+)
 from controllers.console import console_ns
 from controllers.console.workspace import plugin_permission_required
 from controllers.console.wraps import account_initialization_required, is_admin_or_owner_required, setup_required
@@ -25,6 +30,14 @@ from services.plugin.plugin_parameter_service import PluginParameterService
 from services.plugin.plugin_permission_service import PluginPermissionService
 
 
+class AutoUpgradeSettingsResponse(TypedDict):
+    strategy_setting: TenantPluginAutoUpgradeStrategy.StrategySetting
+    upgrade_time_of_day: int
+    upgrade_mode: TenantPluginAutoUpgradeStrategy.UpgradeMode
+    exclude_plugins: list[str]
+    include_plugins: list[str]
+
+
 class ParserList(BaseModel):
     page: int = Field(default=1, ge=1, description="Page number")
     page_size: int = Field(default=256, ge=1, le=256, description="Page size (1-256)")
@@ -88,8 +101,8 @@ class ParserUninstall(BaseModel):
 
 
 class ParserPermissionChange(BaseModel):
-    install_permission: TenantPluginPermission.InstallPermission
-    debug_permission: TenantPluginPermission.DebugPermission
+    install_permission: TenantPluginPermission.InstallPermission = TenantPluginPermission.InstallPermission.EVERYONE
+    debug_permission: TenantPluginPermission.DebugPermission = TenantPluginPermission.DebugPermission.EVERYONE
 
 
 class ParserDynamicOptions(BaseModel):
@@ -125,13 +138,40 @@ class PluginAutoUpgradeSettingsPayload(BaseModel):
     include_plugins: list[str] = Field(default_factory=list)
 
 
-class ParserPreferencesChange(BaseModel):
-    permission: PluginPermissionSettingsPayload
+class PluginAutoUpgradeChangeResponse(ResponseModel):
+    success: bool
+    message: str | None = None
+
+
+class PluginAutoUpgradeSettingsResponseModel(ResponseModel):
+    strategy_setting: TenantPluginAutoUpgradeStrategy.StrategySetting
+    upgrade_time_of_day: int
+    upgrade_mode: TenantPluginAutoUpgradeStrategy.UpgradeMode
+    exclude_plugins: list[str]
+    include_plugins: list[str]
+
+
+class PluginAutoUpgradeFetchResponse(ResponseModel):
+    category: TenantPluginAutoUpgradeStrategy.PluginCategory
+    auto_upgrade: PluginAutoUpgradeSettingsResponseModel
+
+
+class ParserAutoUpgradeChange(BaseModel):
+    model_config = ConfigDict(extra="forbid")
+
+    category: TenantPluginAutoUpgradeStrategy.PluginCategory
     auto_upgrade: PluginAutoUpgradeSettingsPayload
 
 
+class ParserAutoUpgradeFetch(BaseModel):
+    category: TenantPluginAutoUpgradeStrategy.PluginCategory
+
+
 class ParserExcludePlugin(BaseModel):
+    model_config = ConfigDict(extra="forbid")
+
     plugin_id: str
+    category: TenantPluginAutoUpgradeStrategy.PluginCategory
 
 
 class ParserReadme(BaseModel):
@@ -164,21 +204,53 @@ register_schema_models(
     ParserPermissionChange,
     ParserDynamicOptions,
     ParserDynamicOptionsWithCredentials,
-    ParserPreferencesChange,
+    ParserAutoUpgradeChange,
+    ParserAutoUpgradeFetch,
     ParserExcludePlugin,
     ParserReadme,
 )
-register_response_schema_models(console_ns, PluginDebuggingKeyResponse, SuccessResponse)
+register_response_schema_models(
+    console_ns,
+    PluginAutoUpgradeChangeResponse,
+    PluginAutoUpgradeFetchResponse,
+    PluginAutoUpgradeSettingsResponseModel,
+    PluginDebuggingKeyResponse,
+    SuccessResponse,
+)
 
 register_enum_models(
     console_ns,
     TenantPluginPermission.DebugPermission,
+    TenantPluginAutoUpgradeStrategy.PluginCategory,
     TenantPluginAutoUpgradeStrategy.UpgradeMode,
     TenantPluginAutoUpgradeStrategy.StrategySetting,
     TenantPluginPermission.InstallPermission,
 )
 
 
+def _default_auto_upgrade_settings(
+    tenant_id: str,
+    category: TenantPluginAutoUpgradeStrategy.PluginCategory,
+) -> AutoUpgradeSettingsResponse:
+    return {
+        "strategy_setting": PluginAutoUpgradeService.default_strategy_setting_for_category(category),
+        "upgrade_time_of_day": PluginAutoUpgradeService.default_upgrade_time_of_day(tenant_id),
+        "upgrade_mode": TenantPluginAutoUpgradeStrategy.UpgradeMode.EXCLUDE,
+        "exclude_plugins": [],
+        "include_plugins": [],
+    }
+
+
+def _auto_upgrade_settings_to_dict(strategy: TenantPluginAutoUpgradeStrategy) -> AutoUpgradeSettingsResponse:
+    return {
+        "strategy_setting": strategy.strategy_setting,
+        "upgrade_time_of_day": strategy.upgrade_time_of_day,
+        "upgrade_mode": strategy.upgrade_mode,
+        "exclude_plugins": strategy.exclude_plugins,
+        "include_plugins": strategy.include_plugins,
+    }
+
+
 def _read_upload_content(file: FileStorage, max_size: int) -> bytes:
     """
     Read the uploaded file and validate its actual size before delegating to the plugin service.
@@ -632,11 +704,13 @@ class PluginChangePermissionApi(Resource):
 
         tenant_id = current_tenant_id
 
-        return {
-            "success": PluginPermissionService.change_permission(
-                tenant_id, args.install_permission, args.debug_permission
-            )
-        }
+        set_permission_result = PluginPermissionService.change_permission(
+            tenant_id, args.install_permission, args.debug_permission
+        )
+        if not set_permission_result:
+            return jsonable_encoder({"success": False, "message": "Failed to set permission"})
+
+        return jsonable_encoder({"success": True})
 
 
 @console_ns.route("/workspaces/current/plugin/permission/fetch")
@@ -725,9 +799,10 @@ class PluginFetchDynamicSelectOptionsWithCredentialsApi(Resource):
         return jsonable_encoder({"options": options})
 
 
-@console_ns.route("/workspaces/current/plugin/preferences/change")
-class PluginChangePreferencesApi(Resource):
-    @console_ns.expect(console_ns.models[ParserPreferencesChange.__name__])
+@console_ns.route("/workspaces/current/plugin/auto-upgrade/change")
+class PluginChangeAutoUpgradeApi(Resource):
+    @console_ns.expect(console_ns.models[ParserAutoUpgradeChange.__name__])
+    @console_ns.response(200, "Success", console_ns.models[PluginAutoUpgradeChangeResponse.__name__])
     @setup_required
     @login_required
     @account_initialization_required
@@ -736,38 +811,17 @@ class PluginChangePreferencesApi(Resource):
         if not user.is_admin_or_owner:
             raise Forbidden()
 
-        args = ParserPreferencesChange.model_validate(console_ns.payload)
-
-        permission = args.permission
-
-        install_permission = permission.install_permission
-        debug_permission = permission.debug_permission
+        args = ParserAutoUpgradeChange.model_validate(console_ns.payload)
 
         auto_upgrade = args.auto_upgrade
-
-        strategy_setting = auto_upgrade.strategy_setting
-        upgrade_time_of_day = auto_upgrade.upgrade_time_of_day
-        upgrade_mode = auto_upgrade.upgrade_mode
-        exclude_plugins = auto_upgrade.exclude_plugins
-        include_plugins = auto_upgrade.include_plugins
-
-        # set permission
-        set_permission_result = PluginPermissionService.change_permission(
-            tenant_id,
-            install_permission,
-            debug_permission,
-        )
-        if not set_permission_result:
-            return jsonable_encoder({"success": False, "message": "Failed to set permission"})
-
-        # set auto upgrade strategy
         set_auto_upgrade_strategy_result = PluginAutoUpgradeService.change_strategy(
             tenant_id,
-            strategy_setting,
-            upgrade_time_of_day,
-            upgrade_mode,
-            exclude_plugins,
-            include_plugins,
+            auto_upgrade.strategy_setting,
+            auto_upgrade.upgrade_time_of_day,
+            auto_upgrade.upgrade_mode,
+            auto_upgrade.exclude_plugins,
+            auto_upgrade.include_plugins,
+            category=args.category,
         )
         if not set_auto_upgrade_strategy_result:
             return jsonable_encoder({"success": False, "message": "Failed to set auto upgrade strategy"})
@@ -775,48 +829,36 @@ class PluginChangePreferencesApi(Resource):
         return jsonable_encoder({"success": True})
 
 
-@console_ns.route("/workspaces/current/plugin/preferences/fetch")
-class PluginFetchPreferencesApi(Resource):
+@console_ns.route("/workspaces/current/plugin/auto-upgrade/fetch")
+class PluginFetchAutoUpgradeApi(Resource):
+    @console_ns.doc(params=query_params_from_model(ParserAutoUpgradeFetch))
+    @console_ns.response(200, "Success", console_ns.models[PluginAutoUpgradeFetchResponse.__name__])
     @setup_required
     @login_required
     @account_initialization_required
     def get(self):
         _, tenant_id = current_account_with_tenant()
 
-        permission = PluginPermissionService.get_permission(tenant_id)
-        permission_dict = {
-            "install_permission": TenantPluginPermission.InstallPermission.EVERYONE,
-            "debug_permission": TenantPluginPermission.DebugPermission.EVERYONE,
-        }
+        args = ParserAutoUpgradeFetch.model_validate(request.args.to_dict(flat=True))
+        auto_upgrade = PluginAutoUpgradeService.get_strategy(tenant_id, args.category)
+        auto_upgrade_dict = (
+            _auto_upgrade_settings_to_dict(auto_upgrade)
+            if auto_upgrade
+            else _default_auto_upgrade_settings(tenant_id, args.category)
+        )
 
-        if permission:
-            permission_dict["install_permission"] = permission.install_permission
-            permission_dict["debug_permission"] = permission.debug_permission
-
-        auto_upgrade = PluginAutoUpgradeService.get_strategy(tenant_id)
-        auto_upgrade_dict = {
-            "strategy_setting": TenantPluginAutoUpgradeStrategy.StrategySetting.DISABLED,
-            "upgrade_time_of_day": 0,
-            "upgrade_mode": TenantPluginAutoUpgradeStrategy.UpgradeMode.EXCLUDE,
-            "exclude_plugins": [],
-            "include_plugins": [],
-        }
-
-        if auto_upgrade:
-            auto_upgrade_dict = {
-                "strategy_setting": auto_upgrade.strategy_setting,
-                "upgrade_time_of_day": auto_upgrade.upgrade_time_of_day,
-                "upgrade_mode": auto_upgrade.upgrade_mode,
-                "exclude_plugins": auto_upgrade.exclude_plugins,
-                "include_plugins": auto_upgrade.include_plugins,
+        return jsonable_encoder(
+            {
+                "category": args.category,
+                "auto_upgrade": auto_upgrade_dict,
             }
-
-        return jsonable_encoder({"permission": permission_dict, "auto_upgrade": auto_upgrade_dict})
+        )
 
 
-@console_ns.route("/workspaces/current/plugin/preferences/autoupgrade/exclude")
+@console_ns.route("/workspaces/current/plugin/auto-upgrade/exclude")
 class PluginAutoUpgradeExcludePluginApi(Resource):
     @console_ns.expect(console_ns.models[ParserExcludePlugin.__name__])
+    @console_ns.response(200, "Success", console_ns.models[SuccessResponse.__name__])
     @setup_required
     @login_required
     @account_initialization_required
@@ -826,7 +868,9 @@ class PluginAutoUpgradeExcludePluginApi(Resource):
 
         args = ParserExcludePlugin.model_validate(console_ns.payload)
 
-        return jsonable_encoder({"success": PluginAutoUpgradeService.exclude_plugin(tenant_id, args.plugin_id)})
+        return jsonable_encoder(
+            {"success": PluginAutoUpgradeService.exclude_plugin(tenant_id, args.plugin_id, args.category)}
+        )
 
 
 @console_ns.route("/workspaces/current/plugin/readme")

api/controllers/console/workspace/workspace.py

@@ -166,10 +166,10 @@ class TenantListApi(Resource):
                 if tenant_plan:
                     plan = tenant_plan["plan"] or CloudPlan.SANDBOX
                 else:
-                    features = FeatureService.get_features(tenant.id)
+                    features = FeatureService.get_features(tenant.id, exclude_vector_space=True)
                     plan = features.billing.subscription.plan or CloudPlan.SANDBOX
             elif not is_enterprise_only:
-                features = FeatureService.get_features(tenant.id)
+                features = FeatureService.get_features(tenant.id, exclude_vector_space=True)
                 plan = features.billing.subscription.plan or CloudPlan.SANDBOX
 
             # Create a dictionary with tenant attributes

api/controllers/console/wraps.py

@@ -96,21 +96,28 @@ def cloud_edition_billing_resource_check[**P, R](resource: str) -> Callable[[Cal
         @wraps(view)
         def decorated(*args: P.args, **kwargs: P.kwargs):
             _, current_tenant_id = current_account_with_tenant()
-            features = FeatureService.get_features(current_tenant_id)
+            if resource == "vector_space":
+                if not dify_config.BILLING_ENABLED:
+                    return view(*args, **kwargs)
+
+                vector_space = FeatureService.get_vector_space(current_tenant_id)
+                if 0 < vector_space.limit <= vector_space.size:
+                    abort(
+                        403,
+                        "The capacity of the knowledge storage space has reached the limit of your subscription.",
+                    )
+                return view(*args, **kwargs)
+
+            features = FeatureService.get_features(current_tenant_id, exclude_vector_space=True)
             if features.billing.enabled:
                 members = features.members
                 apps = features.apps
-                vector_space = features.vector_space
                 documents_upload_quota = features.documents_upload_quota
                 annotation_quota_limit = features.annotation_quota_limit
                 if resource == "members" and 0 < members.limit <= members.size:
                     abort(403, "The number of members has reached the limit of your subscription.")
                 elif resource == "apps" and 0 < apps.limit <= apps.size:
                     abort(403, "The number of apps has reached the limit of your subscription.")
-                elif resource == "vector_space" and 0 < vector_space.limit <= vector_space.size:
-                    abort(
-                        403, "The capacity of the knowledge storage space has reached the limit of your subscription."
-                    )
                 elif resource == "documents" and 0 < documents_upload_quota.limit <= documents_upload_quota.size:
                     # The api of file upload is used in the multiple places,
                     # so we need to check the source of the request from datasets
@@ -140,7 +147,7 @@ def cloud_edition_billing_knowledge_limit_check[**P, R](
         @wraps(view)
         def decorated(*args: P.args, **kwargs: P.kwargs):
             _, current_tenant_id = current_account_with_tenant()
-            features = FeatureService.get_features(current_tenant_id)
+            features = FeatureService.get_features(current_tenant_id, exclude_vector_space=True)
             if features.billing.enabled:
                 if resource == "add_segment":
                     if features.billing.subscription.plan == CloudPlan.SANDBOX:
@@ -291,7 +298,7 @@ def knowledge_pipeline_publish_enabled[**P, R](view: Callable[P, R]) -> Callable
     @wraps(view)
     def decorated(*args: P.args, **kwargs: P.kwargs):
         _, current_tenant_id = current_account_with_tenant()
-        features = FeatureService.get_features(current_tenant_id)
+        features = FeatureService.get_features(current_tenant_id, exclude_vector_space=True)
         if features.knowledge_pipeline.publish_enabled:
             return view(*args, **kwargs)
         abort(403)

api/controllers/openapi/__init__.py

@@ -37,6 +37,13 @@ from controllers.openapi._models import (
     DeviceMutateRequest,
     DeviceMutateResponse,
     DevicePollRequest,
+    MemberActionResponse,
+    MemberInvitePayload,
+    MemberInviteResponse,
+    MemberListQuery,
+    MemberListResponse,
+    MemberResponse,
+    MemberRoleUpdatePayload,
     MessageMetadata,
     PermittedExternalAppsListQuery,
     PermittedExternalAppsListResponse,
@@ -63,6 +70,9 @@ register_schema_models(
     DevicePollRequest,
     DeviceLookupQuery,
     DeviceMutateRequest,
+    MemberInvitePayload,
+    MemberListQuery,
+    MemberRoleUpdatePayload,
     PermittedExternalAppsListQuery,
 )
 register_response_schema_models(
@@ -86,6 +96,10 @@ register_response_schema_models(
     WorkspaceSummaryResponse,
     WorkspaceListResponse,
     WorkspaceDetailResponse,
+    MemberResponse,
+    MemberListResponse,
+    MemberInviteResponse,
+    MemberActionResponse,
     DeviceCodeResponse,
     DeviceLookupResponse,
     DeviceMutateResponse,

api/controllers/openapi/_models.py

@@ -6,7 +6,7 @@ from typing import Any, Literal
 
 from pydantic import BaseModel, ConfigDict, Field, field_validator
 
-from libs.helper import UUIDStrOrEmpty, uuid_value
+from libs.helper import EmailStr, UUIDStrOrEmpty, uuid_value
 from models.model import AppMode
 
 # Server-side cap on `limit` query param for /openapi/v1/* list endpoints.
@@ -342,3 +342,61 @@ class ApprovalGrantClaimsPayload(BaseModel):
     user_code: str = Field(min_length=1, max_length=32)
     nonce: str = Field(min_length=1, max_length=128)
     csrf_token: str = Field(min_length=1, max_length=128)
+
+
+# Closed enum for invite/update-role payloads. Owner is intentionally not
+# assignable through these endpoints — ownership transfer goes through the
+# console's three-step email-verification flow.
+MemberAssignableRole = Literal["normal", "admin"]
+
+
+class MemberResponse(BaseModel):
+    id: str
+    name: str
+    email: str
+    role: str
+    status: str
+    avatar: str | None = None
+
+
+class MemberListResponse(BaseModel):
+    page: int
+    limit: int
+    total: int
+    has_more: bool
+    data: list[MemberResponse]
+
+
+class MemberListQuery(BaseModel):
+    """Strict (extra='forbid')."""
+
+    model_config = ConfigDict(extra="forbid")
+
+    page: int = Field(1, ge=1)
+    limit: int = Field(20, ge=1, le=MAX_PAGE_LIMIT)
+
+
+class MemberInvitePayload(BaseModel):
+    model_config = ConfigDict(extra="forbid")
+
+    email: EmailStr
+    role: MemberAssignableRole
+
+
+class MemberRoleUpdatePayload(BaseModel):
+    model_config = ConfigDict(extra="forbid")
+
+    role: MemberAssignableRole
+
+
+class MemberInviteResponse(BaseModel):
+    result: Literal["success"] = "success"
+    email: str
+    role: str
+    member_id: str
+    invite_url: str
+    tenant_id: str
+
+
+class MemberActionResponse(BaseModel):
+    result: Literal["success"] = "success"

api/controllers/openapi/account.py

@@ -4,7 +4,7 @@ from datetime import UTC, datetime
 
 from flask import request
 from flask_restx import Resource
-from werkzeug.exceptions import BadRequest, NotFound
+from werkzeug.exceptions import NotFound
 
 from controllers.openapi import openapi_ns
 from controllers.openapi._models import (
@@ -17,18 +17,17 @@ from controllers.openapi._models import (
     SessionRow,
     WorkspacePayload,
 )
+from controllers.openapi.auth.composition import auth_router
+from controllers.openapi.auth.data import AuthData
 from extensions.ext_database import db
 from extensions.ext_redis import redis_client
 from libs.oauth_bearer import (
-    ACCEPT_USER_ANY,
-    AuthContext,
-    SubjectType,
+    Scope,
+    TokenType,
     get_auth_ctx,
-    validate_bearer,
 )
 from libs.rate_limit import (
     LIMIT_ME_PER_ACCOUNT,
-    LIMIT_ME_PER_EMAIL,
     enforce,
 )
 from services.account_service import AccountService, TenantService
@@ -42,32 +41,18 @@ from services.oauth_device_flow import (
 @openapi_ns.route("/account")
 class AccountApi(Resource):
     @openapi_ns.response(200, "Account info", openapi_ns.models[AccountResponse.__name__])
-    @validate_bearer(accept=ACCEPT_USER_ANY)
-    def get(self):
-        ctx = get_auth_ctx()
+    @auth_router.guard(scope=Scope.FULL, allowed_token_types=frozenset({TokenType.OAUTH_ACCOUNT}))
+    def get(self, *, auth_data: AuthData):
+        enforce(LIMIT_ME_PER_ACCOUNT, key=f"account:{auth_data.account_id}")
 
-        if ctx.subject_type == SubjectType.EXTERNAL_SSO:
-            enforce(LIMIT_ME_PER_EMAIL, key=f"subject:{ctx.subject_email}")
-        else:
-            enforce(LIMIT_ME_PER_ACCOUNT, key=f"account:{ctx.account_id}")
-
-        if ctx.subject_type == SubjectType.EXTERNAL_SSO:
-            return AccountResponse(
-                subject_type=ctx.subject_type,
-                subject_email=ctx.subject_email,
-                subject_issuer=ctx.subject_issuer,
-                account=None,
-                workspaces=[],
-                default_workspace_id=None,
-            ).model_dump(mode="json")
-
-        account = AccountService.get_account_by_id(db.session, str(ctx.account_id)) if ctx.account_id else None
-        memberships = TenantService.get_account_memberships(db.session, str(ctx.account_id)) if ctx.account_id else []
+        account_id_str = str(auth_data.account_id) if auth_data.account_id else None
+        account = AccountService.get_account_by_id(db.session, account_id_str) if account_id_str else None
+        memberships = TenantService.get_account_memberships(db.session, account_id_str) if account_id_str else []
         default_ws_id = _pick_default_workspace(memberships)
 
         return AccountResponse(
-            subject_type=ctx.subject_type,
-            subject_email=ctx.subject_email or (account.email if account else None),
+            subject_type="account",
+            subject_email=account.email if account else None,
             account=_account_payload(account) if account else None,
             workspaces=[_workspace_payload(m) for m in memberships],
             default_workspace_id=default_ws_id,
@@ -77,19 +62,17 @@ class AccountApi(Resource):
 @openapi_ns.route("/account/sessions/self")
 class AccountSessionsSelfApi(Resource):
     @openapi_ns.response(200, "Session revoked", openapi_ns.models[RevokeResponse.__name__])
-    @validate_bearer(accept=ACCEPT_USER_ANY)
-    def delete(self):
-        ctx = get_auth_ctx()
-        _require_oauth_subject(ctx)
-        revoke_oauth_token(db.session, redis_client, str(ctx.token_id))
+    @auth_router.guard(scope=Scope.FULL, allowed_token_types=frozenset({TokenType.OAUTH_ACCOUNT}))
+    def delete(self, *, auth_data: AuthData):
+        revoke_oauth_token(db.session, redis_client, str(auth_data.token_id))
         return RevokeResponse(status="revoked").model_dump(mode="json"), 200
 
 
 @openapi_ns.route("/account/sessions")
 class AccountSessionsApi(Resource):
     @openapi_ns.response(200, "Session list", openapi_ns.models[SessionListResponse.__name__])
-    @validate_bearer(accept=ACCEPT_USER_ANY)
-    def get(self):
+    @auth_router.guard(scope=Scope.FULL, allowed_token_types=frozenset({TokenType.OAUTH_ACCOUNT}))
+    def get(self, *, auth_data: AuthData):
         ctx = get_auth_ctx()
         now = datetime.now(UTC)
         page = int(request.args.get("page", "1"))
@@ -122,10 +105,9 @@ class AccountSessionsApi(Resource):
 @openapi_ns.route("/account/sessions/<string:session_id>")
 class AccountSessionByIdApi(Resource):
     @openapi_ns.response(200, "Session revoked", openapi_ns.models[RevokeResponse.__name__])
-    @validate_bearer(accept=ACCEPT_USER_ANY)
-    def delete(self, session_id: str):
+    @auth_router.guard(scope=Scope.FULL, allowed_token_types=frozenset({TokenType.OAUTH_ACCOUNT}))
+    def delete(self, session_id: str, *, auth_data: AuthData):
         ctx = get_auth_ctx()
-        _require_oauth_subject(ctx)
 
         # 404 (not 403) on cross-subject so the endpoint doesn't leak
         # token IDs that belong to other subjects.
@@ -136,13 +118,6 @@ class AccountSessionByIdApi(Resource):
         return RevokeResponse(status="revoked").model_dump(mode="json"), 200
 
 
-def _require_oauth_subject(ctx: AuthContext) -> None:
-    if not ctx.source.startswith("oauth"):
-        raise BadRequest(
-            "this endpoint revokes OAuth bearer tokens; use /openapi/v1/personal-access-tokens/self for PATs"
-        )
-
-
 def _iso(dt: datetime | None) -> str | None:
     if dt is None:
         return None

api/controllers/openapi/app_run.py

@@ -16,7 +16,8 @@ import services
 from controllers.openapi import openapi_ns
 from controllers.openapi._audit import emit_app_run
 from controllers.openapi._models import AppRunRequest
-from controllers.openapi.auth.composition import OAUTH_BEARER_PIPELINE
+from controllers.openapi.auth.composition import auth_router
+from controllers.openapi.auth.data import AuthData
 from controllers.service_api.app.error import (
     AppUnavailableError,
     CompletionRequestError,
@@ -124,8 +125,9 @@ _DISPATCH: dict[AppMode, Callable[[App, Any, AppRunRequest], Any]] = {
 class AppRunApi(Resource):
     @openapi_ns.expect(openapi_ns.models[AppRunRequest.__name__])
     @openapi_ns.response(200, "Run result (SSE stream)")
-    @OAUTH_BEARER_PIPELINE.guard(scope=Scope.APPS_RUN)
-    def post(self, app_id: str, app_model: App, caller, caller_kind: str):
+    @auth_router.guard(scope=Scope.APPS_RUN)
+    def post(self, app_id: str, *, auth_data: AuthData):
+        app_model, caller, caller_kind = auth_data.require_app_context()
         body = request.get_json(silent=True) or {}
         try:
             payload = AppRunRequest.model_validate(body)
@@ -158,8 +160,9 @@ class AppRunApi(Resource):
 @openapi_ns.route("/apps/<string:app_id>/tasks/<string:task_id>/stop")
 class AppRunTaskStopApi(Resource):
     @openapi_ns.response(200, "Task stopped")
-    @OAUTH_BEARER_PIPELINE.guard(scope=Scope.APPS_RUN)
-    def post(self, app_id: str, task_id: str, app_model: App, caller, caller_kind: str):
+    @auth_router.guard(scope=Scope.APPS_RUN)
+    def post(self, app_id: str, task_id: str, *, auth_data: AuthData):
+        app_model, caller, caller_kind = auth_data.require_app_context()
         AppQueueManager.set_stop_flag_no_user_check(task_id)
         GraphEngineManager(redis_client).send_stop_command(task_id)
         return {"result": "success"}

api/controllers/openapi/apps.py

@@ -1,9 +1,4 @@
-"""GET /openapi/v1/apps and per-app reads.
-
-Decorator order: `method_decorators` is innermost-first. `validate_bearer`
-is last → outermost → publishes the auth ContextVar before `require_scope`
-reads it.
-"""
+"""GET /openapi/v1/apps and per-app reads."""
 
 from __future__ import annotations
 
@@ -28,31 +23,17 @@ from controllers.openapi._models import (
     AppListRow,
     TagItem,
 )
-from controllers.openapi.auth.surface_gate import accept_subjects
+from controllers.openapi.auth.composition import auth_router
+from controllers.openapi.auth.data import AuthData
 from controllers.service_api.app.error import AppUnavailableError
 from core.app.app_config.common.parameters_mapping import get_parameters_from_feature_dict
 from extensions.ext_database import db
-from libs.oauth_bearer import (
-    ACCEPT_USER_ANY,
-    AuthContext,
-    Scope,
-    SubjectType,
-    get_auth_ctx,
-    require_scope,
-    require_workspace_member,
-    validate_bearer,
-)
+from libs.oauth_bearer import Scope, TokenType
 from models import App
 from services.account_service import TenantService
 from services.app_service import AppListParams, AppService
 from services.tag_service import TagService
 
-_APPS_READ_DECORATORS = [
-    require_scope(Scope.APPS_READ),
-    accept_subjects(SubjectType.ACCOUNT),
-    validate_bearer(accept=ACCEPT_USER_ANY),
-]
-
 _ALLOWED_DESCRIBE_FIELDS: frozenset[str] = frozenset({"info", "parameters", "input_schema"})
 
 
@@ -66,13 +47,9 @@ _EMPTY_PARAMETERS: dict[str, Any] = {
 
 
 class AppReadResource(Resource):
-    """Base for per-app read endpoints; subclasses call `_load()` for SSO/membership/exists checks."""
-
-    method_decorators = _APPS_READ_DECORATORS
-
-    def _load(self, app_id: str, workspace_id: str | None = None) -> tuple[App, AuthContext]:
-        ctx: AuthContext = get_auth_ctx()
+    """Base for per-app read endpoints; subclasses call `_load()` for membership/exists checks."""
 
+    def _load(self, app_id: str, workspace_id: str | None = None) -> App:
         try:
             parsed_uuid = _uuid.UUID(app_id)
             is_uuid = True
@@ -99,8 +76,7 @@ class AppReadResource(Resource):
                 raise Conflict("".join(lines))
             app = matches[0]
 
-        require_workspace_member(ctx, str(app.tenant_id))
-        return app, ctx
+        return app
 
 
 def parameters_payload(app: App) -> dict:
@@ -114,13 +90,14 @@ def parameters_payload(app: App) -> dict:
 class AppDescribeApi(AppReadResource):
     @openapi_ns.doc(params=query_params_from_model(AppDescribeQuery))
     @openapi_ns.response(200, "App description", openapi_ns.models[AppDescribeResponse.__name__])
-    def get(self, app_id: str):
+    @auth_router.guard(scope=Scope.APPS_READ, allowed_token_types=frozenset({TokenType.OAUTH_ACCOUNT}))
+    def get(self, app_id: str, *, auth_data: AuthData):
         try:
             query = AppDescribeQuery.model_validate(request.args.to_dict(flat=True))
         except ValidationError as exc:
             raise UnprocessableEntity(exc.json())
 
-        app, _ = self._load(app_id, workspace_id=query.workspace_id)
+        app = self._load(app_id, workspace_id=query.workspace_id)
 
         requested = query.fields
         want_info = requested is None or "info" in requested
@@ -168,20 +145,16 @@ class AppDescribeApi(AppReadResource):
 
 @openapi_ns.route("/apps")
 class AppListApi(Resource):
-    method_decorators = _APPS_READ_DECORATORS
-
     @openapi_ns.doc(params=query_params_from_model(AppListQuery))
     @openapi_ns.response(200, "App list", openapi_ns.models[AppListResponse.__name__])
-    def get(self):
-        ctx: AuthContext = get_auth_ctx()
-
+    @auth_router.guard(scope=Scope.APPS_READ, allowed_token_types=frozenset({TokenType.OAUTH_ACCOUNT}))
+    def get(self, *, auth_data: AuthData):
         try:
             query: AppListQuery = AppListQuery.model_validate(request.args.to_dict(flat=True))
         except ValidationError as exc:
             raise UnprocessableEntity(exc.json())
 
         workspace_id = query.workspace_id
-        require_workspace_member(ctx, workspace_id)
 
         empty = (
             AppListResponse(page=query.page, limit=query.limit, total=0, has_more=False, data=[]).model_dump(
@@ -237,7 +210,7 @@ class AppListApi(Resource):
             openapi_visible=True,
         )
 
-        pagination = AppService().get_paginate_apps(str(ctx.account_id), workspace_id, params)
+        pagination = AppService().get_paginate_apps(str(auth_data.account_id), workspace_id, params)
         if pagination is None:
             return empty
 

api/controllers/openapi/apps_permitted_external.py

@@ -18,37 +18,27 @@ from controllers.openapi._models import (
     PermittedExternalAppsListQuery,
     PermittedExternalAppsListResponse,
 )
-from controllers.openapi.auth.surface_gate import accept_subjects
+from controllers.openapi.auth.composition import auth_router
+from controllers.openapi.auth.data import AuthData, Edition
 from extensions.ext_database import db
-from libs.device_flow_security import enterprise_only
-from libs.oauth_bearer import (
-    ACCEPT_USER_ANY,
-    Scope,
-    SubjectType,
-    require_scope,
-    validate_bearer,
-)
+from libs.oauth_bearer import Scope, TokenType
 from models import App
 from services.account_service import TenantService
 from services.app_service import AppService
 from services.enterprise.app_permitted_service import list_permitted_apps
-from services.openapi.license_gate import license_required
 
 
 @openapi_ns.route("/permitted-external-apps")
 class PermittedExternalAppsListApi(Resource):
-    method_decorators = [
-        require_scope(Scope.APPS_READ_PERMITTED_EXTERNAL),
-        license_required,
-        accept_subjects(SubjectType.EXTERNAL_SSO),
-        validate_bearer(accept=ACCEPT_USER_ANY),
-        enterprise_only,
-    ]
-
     @openapi_ns.response(
         200, "Permitted external apps list", openapi_ns.models[PermittedExternalAppsListResponse.__name__]
     )
-    def get(self):
+    @auth_router.guard(
+        scope=Scope.APPS_READ_PERMITTED_EXTERNAL,
+        allowed_token_types=frozenset({TokenType.OAUTH_EXTERNAL_SSO}),
+        edition=frozenset({Edition.EE}),
+    )
+    def get(self, *, auth_data: AuthData):
         try:
             query = PermittedExternalAppsListQuery.model_validate(request.args.to_dict(flat=True))
         except ValidationError as exc:

api/controllers/openapi/auth/__init__.py

@@ -1,3 +1,3 @@
-from controllers.openapi.auth.composition import OAUTH_BEARER_PIPELINE
+from controllers.openapi.auth.composition import auth_router
 
-__all__ = ["OAUTH_BEARER_PIPELINE"]
+__all__ = ["auth_router"]

api/controllers/openapi/auth/composition.py

@@ -1,46 +1,64 @@
-"""`OAUTH_BEARER_PIPELINE` — the auth scheme for openapi `/run` endpoints.
-
-Endpoints attach via `@OAUTH_BEARER_PIPELINE.guard(scope=…)`. No alternative
-paths. Read endpoints (`/apps`, `/info`, `/parameters`, `/describe`) skip
-the pipeline and use `validate_bearer + require_scope + require_workspace_member`
-inline — they don't need `AppAuthzCheck`/`CallerMount`.
-"""
-
 from __future__ import annotations
 
-from controllers.openapi.auth.pipeline import Pipeline
-from controllers.openapi.auth.steps import (
-    AppAuthzCheck,
-    AppResolver,
-    BearerCheck,
-    CallerMount,
-    ScopeCheck,
-    SurfaceCheck,
-    WorkspaceMembershipCheck,
+from controllers.openapi.auth.conditions import (
+    EDITION_CE,
+    EDITION_EE,
+    LOADED_APP_IS_PRIVATE,
+    PATH_HAS_APP_ID,
+    WEBAPP_AUTH_ENABLED,
 )
-from controllers.openapi.auth.strategies import (
-    AccountMounter,
-    AclStrategy,
-    AppAuthzStrategy,
-    EndUserMounter,
-    MembershipStrategy,
+from controllers.openapi.auth.data import Edition
+from controllers.openapi.auth.flow import When
+from controllers.openapi.auth.pipeline import AuthPipeline, PipelineRoute, PipelineRouter
+from controllers.openapi.auth.prepare import (
+    load_account,
+    load_app,
+    load_app_access_mode,
+    load_tenant,
+    resolve_external_user,
 )
-from libs.oauth_bearer import SubjectType
-from services.feature_service import FeatureService
-
-
-def _resolve_app_authz_strategy() -> AppAuthzStrategy:
-    if FeatureService.get_system_features().webapp_auth.enabled:
-        return AclStrategy()
-    return MembershipStrategy()
-
-
-OAUTH_BEARER_PIPELINE = Pipeline(
-    BearerCheck(),
-    SurfaceCheck(accepted=frozenset({SubjectType.ACCOUNT})),
-    ScopeCheck(),
-    AppResolver(),
-    WorkspaceMembershipCheck(),
-    AppAuthzCheck(_resolve_app_authz_strategy),
-    CallerMount(AccountMounter(), EndUserMounter()),
+from controllers.openapi.auth.verify import (
+    check_acl,
+    check_app_access,
+    check_membership,
+    check_private_app_permission,
+    check_scope,
+)
+from libs.oauth_bearer import TokenType
+
+account_pipeline = AuthPipeline(
+    prepare=[
+        When(PATH_HAS_APP_ID, then=load_app),
+        When(PATH_HAS_APP_ID, then=load_tenant),
+        load_account,  # all tokens here are account tokens
+        When(PATH_HAS_APP_ID & EDITION_EE, then=load_app_access_mode),
+    ],
+    auth=[
+        check_scope,
+        When(EDITION_CE & PATH_HAS_APP_ID, then=check_membership),
+        When(EDITION_EE & PATH_HAS_APP_ID & ~WEBAPP_AUTH_ENABLED, then=check_app_access),
+        When(PATH_HAS_APP_ID & EDITION_EE & WEBAPP_AUTH_ENABLED, then=check_acl),
+        When(EDITION_EE & LOADED_APP_IS_PRIVATE, then=check_private_app_permission),
+    ],
+)
+
+external_sso_pipeline = AuthPipeline(
+    prepare=[
+        When(PATH_HAS_APP_ID, then=load_app),
+        When(PATH_HAS_APP_ID, then=load_tenant),
+        When(PATH_HAS_APP_ID, then=resolve_external_user),
+        When(PATH_HAS_APP_ID, then=load_app_access_mode),
+    ],
+    auth=[
+        check_scope,
+        When(PATH_HAS_APP_ID & WEBAPP_AUTH_ENABLED, then=check_acl),
+        When(LOADED_APP_IS_PRIVATE, then=check_private_app_permission),
+    ],
+)
+
+auth_router = PipelineRouter(
+    {
+        TokenType.OAUTH_ACCOUNT: PipelineRoute(account_pipeline),
+        TokenType.OAUTH_EXTERNAL_SSO: PipelineRoute(external_sso_pipeline, required_edition=frozenset({Edition.EE})),
+    }
 )

api/controllers/openapi/auth/conditions.py

@@ -0,0 +1,53 @@
+from __future__ import annotations
+
+from collections.abc import Callable
+
+from controllers.openapi.auth.data import AuthData, Edition, RequestContext, current_edition
+from libs.oauth_bearer import TokenType
+from services.enterprise.enterprise_service import WebAppAccessMode
+from services.feature_service import FeatureService
+
+CondFn = Callable[[RequestContext, AuthData | None], bool]
+
+
+class Cond:
+    def __init__(self, fn: CondFn) -> None:
+        self._fn = fn
+
+    def __call__(self, ctx: RequestContext, data: AuthData | None = None) -> bool:
+        return self._fn(ctx, data)
+
+    def __and__(self, other: Cond) -> Cond:
+        return Cond(lambda ctx, data: self(ctx, data) and other(ctx, data))
+
+    def __or__(self, other: Cond) -> Cond:
+        return Cond(lambda ctx, data: self(ctx, data) or other(ctx, data))
+
+    def __invert__(self) -> Cond:
+        return Cond(lambda ctx, data: not self(ctx, data))
+
+
+def request_cond(fn: Callable[[RequestContext], bool]) -> Cond:
+    return Cond(lambda ctx, _: fn(ctx))
+
+
+def data_cond(fn: Callable[[AuthData], bool]) -> Cond:
+    return Cond(lambda _, data: data is not None and fn(data))
+
+
+def config_cond(fn: Callable[[], bool]) -> Cond:
+    return Cond(lambda _, __: fn())
+
+
+TOKEN_IS_OAUTH_ACCOUNT = request_cond(lambda ctx: ctx.token_type == TokenType.OAUTH_ACCOUNT)
+TOKEN_IS_OAUTH_EXTERNAL_SSO = request_cond(lambda ctx: ctx.token_type == TokenType.OAUTH_EXTERNAL_SSO)
+
+PATH_HAS_APP_ID = request_cond(lambda ctx: "app_id" in ctx.path_params)
+
+EDITION_CE = config_cond(lambda: current_edition() == Edition.CE)
+EDITION_EE = config_cond(lambda: current_edition() == Edition.EE)
+EDITION_SAAS = config_cond(lambda: current_edition() == Edition.SAAS)
+
+WEBAPP_AUTH_ENABLED = config_cond(lambda: FeatureService.get_system_features().webapp_auth.enabled)
+
+LOADED_APP_IS_PRIVATE = data_cond(lambda data: data.app_access_mode == WebAppAccessMode.PRIVATE)

api/controllers/openapi/auth/context.py

@@ -1,68 +0,0 @@
-"""Mutable per-request context for the openapi auth pipeline.
-
-Every field starts None / empty and is filled in by a step. The pipeline
-is the only thing that should construct or mutate Context — handlers
-read populated values via the decorator's kwargs unpacking.
-
-Context is intentionally decoupled from Flask's ``Request``: the pipeline
-guard extracts whatever transport-level inputs the steps need (bearer
-token, path params) at the boundary and writes them into Context fields,
-so steps stay testable without a request object and won't leak coupling
-to a specific framework.
-"""
-
-from __future__ import annotations
-
-import uuid
-from collections.abc import Mapping
-from contextvars import Token
-from dataclasses import dataclass, field
-from datetime import datetime
-from typing import TYPE_CHECKING, Literal, Protocol
-
-from werkzeug.exceptions import Unauthorized
-
-from libs.oauth_bearer import AuthContext, Scope, SubjectType
-
-if TYPE_CHECKING:
-    from models import App, Tenant
-
-
-@dataclass
-class Context:
-    required_scope: Scope
-    bearer_token: str | None = None
-    path_params: Mapping[str, str] = field(default_factory=dict)
-    subject_type: SubjectType | None = None
-    subject_email: str | None = None
-    subject_issuer: str | None = None
-    account_id: uuid.UUID | None = None
-    scopes: frozenset[Scope] = field(default_factory=frozenset)
-    token_id: uuid.UUID | None = None
-    token_hash: str | None = None
-    cached_verified_tenants: dict[str, bool] | None = None
-    source: str | None = None
-    expires_at: datetime | None = None
-    app: App | None = None
-    tenant: Tenant | None = None
-    caller: object | None = None
-    caller_kind: Literal["account", "end_user"] | None = None
-    auth_ctx_reset_token: Token[AuthContext] | None = None
-
-    @property
-    def must_tenant(self) -> Tenant:
-        if not self.tenant:
-            raise Unauthorized("tenant is not associated")
-        return self.tenant
-
-    @property
-    def must_subject_type(self) -> SubjectType:
-        if not self.subject_type:
-            raise Unauthorized("subject_type unset — BearerCheck did not run")
-        return self.subject_type
-
-
-class Step(Protocol):
-    """One responsibility. Mutate ctx or raise to short-circuit."""
-
-    def __call__(self, ctx: Context) -> None: ...

api/controllers/openapi/auth/data.py

@@ -0,0 +1,69 @@
+from __future__ import annotations
+
+import uuid
+from enum import StrEnum
+from typing import Literal
+
+from pydantic import BaseModel, ConfigDict, Field
+from werkzeug.exceptions import InternalServerError
+
+from configs import dify_config
+from libs.oauth_bearer import Scope, TokenType
+from models.account import Account, Tenant
+from models.model import App, EndUser
+from services.enterprise.enterprise_service import WebAppAccessMode
+
+
+class Edition(StrEnum):
+    CE = "ce"
+    EE = "ee"
+    SAAS = "saas"
+
+
+def current_edition() -> Edition:
+    if dify_config.EDITION == "CLOUD":
+        return Edition.SAAS
+    if dify_config.ENTERPRISE_ENABLED:
+        return Edition.EE
+    return Edition.CE
+
+
+class ExternalIdentity(BaseModel):
+    model_config = ConfigDict(frozen=True)
+
+    email: str
+    issuer: str | None = None
+
+
+class RequestContext(BaseModel):
+    model_config = ConfigDict(frozen=True)
+
+    token_type: TokenType
+    scope: Scope | None = None
+    path_params: dict[str, str]
+
+
+class AuthData(BaseModel):
+    model_config = ConfigDict(arbitrary_types_allowed=True)
+
+    required_scope: Scope | None = None
+    token_type: TokenType
+    account_id: uuid.UUID | None = None
+    token_hash: str
+    token_id: uuid.UUID | None = None
+    scopes: frozenset[Scope]
+    tenants: dict[str, bool] = Field(default_factory=dict)
+    external_identity: ExternalIdentity | None = None
+    path_params: dict[str, str] = Field(default_factory=dict)
+
+    app: App | None = None
+    tenant: Tenant | None = None
+    app_access_mode: WebAppAccessMode | None = None
+
+    caller: Account | EndUser | None = None
+    caller_kind: Literal["account", "end_user"] | None = None
+
+    def require_app_context(self) -> tuple[App, Account | EndUser, Literal["account", "end_user"]]:
+        if self.app is None or self.caller is None or self.caller_kind is None:
+            raise InternalServerError("pipeline_invariant_violated: app context missing")
+        return self.app, self.caller, self.caller_kind

api/controllers/openapi/auth/flow.py

@@ -0,0 +1,19 @@
+from __future__ import annotations
+
+from collections.abc import Callable
+from typing import Any
+
+from controllers.openapi.auth.conditions import Cond
+from controllers.openapi.auth.data import AuthData, RequestContext
+
+
+class When:
+    def __init__(self, condition: Cond, *, then: Callable[[Any], None]) -> None:
+        self.condition = condition
+        self._step = then
+
+    def applies(self, ctx: RequestContext, data: AuthData | None = None) -> bool:
+        return self.condition(ctx, data)
+
+    def __call__(self, arg: Any) -> None:
+        self._step(arg)

api/controllers/openapi/auth/pipeline.py

@@ -1,51 +1,209 @@
-"""Pipeline IS the auth scheme.
+"""Auth pipeline — entry point for all openapi auth.
 
-`Pipeline.guard(scope=…)` is the only attachment point for endpoints —
-that is the design lock-in: forgetting an auth layer is structurally
-impossible because there is no "sometimes wrap, sometimes don't" choice.
+`PipelineRouter.guard()` is the only attachment point for endpoints.
+`AuthPipeline` is a pure step-runner with no routing concerns.
+`PipelineRoute` binds a pipeline to optional edition requirements.
 """
 
 from __future__ import annotations
 
+from collections.abc import Callable
+from dataclasses import dataclass
 from functools import wraps
+from typing import Any
 
-from flask import request
+from flask import current_app, request
+from flask_login import user_logged_in
+from werkzeug.exceptions import Forbidden, NotFound, Unauthorized
 
-from controllers.openapi.auth.context import Context, Step
-from libs.oauth_bearer import Scope, extract_bearer, reset_auth_ctx
+from controllers.openapi._audit import emit_wrong_surface
+from controllers.openapi.auth.data import (
+    AuthData,
+    Edition,
+    ExternalIdentity,
+    RequestContext,
+    current_edition,
+)
+from controllers.openapi.auth.flow import When
+from libs.oauth_bearer import (
+    AuthContext,
+    Scope,
+    TokenType,
+    extract_bearer,
+    get_authenticator,
+    reset_auth_ctx,
+    set_auth_ctx,
+)
+from services.feature_service import FeatureService, LicenseStatus
 
 
-class Pipeline:
-    def __init__(self, *steps: Step) -> None:
-        self._steps = steps
+class AuthPipeline:
+    """Pure step-runner — no routing, no guard.
 
-    def run(self, ctx: Context) -> None:
-        for step in self._steps:
-            step(ctx)
+    Both `prepare` and `auth` steps receive the same `AuthData` instance.
+    `prepare` steps populate it; `auth` steps validate it.
+    """
 
-    def guard(self, *, scope: Scope):
-        def decorator(view):
+    def __init__(self, prepare: list, auth: list) -> None:
+        self._prepare = prepare
+        self._auth = auth
+
+    def _run(
+        self,
+        identity: AuthContext,
+        args: tuple,
+        kwargs: dict,
+        view: Callable,
+        *,
+        scope: Scope | None,
+    ) -> Any:
+        req_ctx = RequestContext(
+            token_type=identity.token_type,
+            scope=scope,
+            path_params=dict(request.view_args or {}),
+        )
+
+        data = AuthData(
+            token_type=identity.token_type,
+            account_id=identity.account_id,
+            token_hash=identity.token_hash,
+            token_id=identity.token_id,
+            scopes=frozenset(identity.scopes),
+            tenants=dict(identity.verified_tenants),
+            required_scope=scope,
+            path_params=dict(req_ctx.path_params),
+            external_identity=(
+                ExternalIdentity(email=identity.subject_email, issuer=identity.subject_issuer)
+                if identity.subject_email
+                else None
+            ),
+        )
+
+        for step in self._prepare:
+            if _should_run(step, req_ctx, data=None):
+                step(data)
+
+        for step in self._auth:
+            if _should_run(step, req_ctx, data=data):
+                step(data)
+
+        reset_token = set_auth_ctx(identity)
+        if data.caller:
+            _mount_flask_login(data.caller)
+
+        try:
+            kwargs["auth_data"] = data
+            return view(*args, **kwargs)
+        finally:
+            reset_auth_ctx(reset_token)
+
+
+@dataclass(frozen=True)
+class PipelineRoute:
+    pipeline: AuthPipeline
+    required_edition: frozenset[Edition] | None = None
+
+
+class PipelineRouter:
+    """Entry point for openapi auth.
+
+    `guard()` is the decorator that endpoints attach to. It applies
+    global gates (edition, token type) then dispatches to the matching
+    `PipelineRoute` for the token type.
+    """
+
+    def __init__(self, routes: dict[TokenType, PipelineRoute]) -> None:
+        self._routes = routes
+
+    def guard(
+        self,
+        *,
+        scope: Scope | None = None,
+        allowed_token_types: frozenset[TokenType] | None = None,
+        edition: frozenset[Edition] | None = None,
+    ) -> Callable:
+        def decorator(view: Callable) -> Callable:
             @wraps(view)
-            def decorated(*args, **kwargs):
-                # Extract transport-level inputs at the boundary so steps
-                # stay decoupled from Flask's request object.
-                ctx = Context(
-                    required_scope=scope,
-                    bearer_token=extract_bearer(request),
-                    path_params=dict(request.view_args or {}),
+            def decorated(*args: Any, **kwargs: Any) -> Any:
+                return self._execute(
+                    args,
+                    kwargs,
+                    view,
+                    scope=scope,
+                    allowed_token_types=allowed_token_types,
+                    edition=edition,
                 )
-                try:
-                    self.run(ctx)
-                    kwargs.update(
-                        app_model=ctx.app,
-                        caller=ctx.caller,
-                        caller_kind=ctx.caller_kind,
-                    )
-                    return view(*args, **kwargs)
-                finally:
-                    if ctx.auth_ctx_reset_token is not None:
-                        reset_auth_ctx(ctx.auth_ctx_reset_token)
 
             return decorated
 
         return decorator
+
+    def _execute(
+        self,
+        args: tuple,
+        kwargs: dict,
+        view: Callable,
+        *,
+        scope: Scope | None,
+        allowed_token_types: frozenset[TokenType] | None,
+        edition: frozenset[Edition] | None,
+    ) -> Any:
+        # 404 not 403 — this edition doesn't expose the feature at all
+        if edition is not None and current_edition() not in edition:
+            raise NotFound()
+
+        license_checked = False
+        if edition is not None and Edition.EE in edition:
+            _check_license()
+            license_checked = True
+
+        token = extract_bearer(request)
+        if not token:
+            raise Unauthorized("bearer required")
+
+        identity = get_authenticator().authenticate(token)
+
+        if allowed_token_types is not None and identity.token_type not in allowed_token_types:
+            emit_wrong_surface(
+                subject_type=_subject_type_str(identity),
+                attempted_path=request.path,
+                client_id=getattr(identity, "client_id", None),
+                token_id=str(identity.token_id) if identity.token_id else None,
+            )
+            raise Forbidden("unsupported_token_type")
+
+        route = self._routes.get(identity.token_type)
+        if route is None:
+            raise Forbidden("unsupported_token_type")
+
+        if route.required_edition is not None:
+            if current_edition() not in route.required_edition:
+                raise Forbidden("external_sso_requires_ee")
+            if not license_checked and Edition.EE in route.required_edition:
+                _check_license()
+
+        return route.pipeline._run(identity, args, kwargs, view, scope=scope)
+
+
+def _should_run(step: Any, req_ctx: RequestContext, data: AuthData | None) -> bool:
+    if isinstance(step, When):
+        return step.applies(req_ctx, data)
+    return True
+
+
+def _subject_type_str(identity: Any) -> str | None:
+    subject = getattr(identity, "subject_type", None)
+    if subject is None:
+        return None
+    return subject.value if hasattr(subject, "value") else str(subject)
+
+
+def _check_license() -> None:
+    settings = FeatureService.get_system_features()
+    if settings.license.status in {LicenseStatus.INACTIVE, LicenseStatus.EXPIRED, LicenseStatus.LOST}:
+        raise Forbidden("license_invalid")
+
+
+def _mount_flask_login(user: Any) -> None:
+    current_app.login_manager._update_request_context_with_user(user)  # type: ignore[attr-defined]
+    user_logged_in.send(current_app._get_current_object(), user=user)  # type: ignore[attr-defined]

api/controllers/openapi/auth/prepare.py

@@ -0,0 +1,67 @@
+from __future__ import annotations
+
+from werkzeug.exceptions import Forbidden, InternalServerError, NotFound, Unauthorized
+
+from controllers.openapi.auth.data import AuthData
+from core.app.entities.app_invoke_entities import InvokeFrom
+from extensions.ext_database import db
+from models.account import TenantStatus
+from services.account_service import AccountService, TenantService
+from services.app_service import AppService
+from services.end_user_service import EndUserService
+from services.enterprise.enterprise_service import EnterpriseService, WebAppAccessMode
+
+
+def load_app(data: AuthData) -> None:
+    app_id = data.path_params["app_id"]
+    app = AppService.get_app_by_id(db.session, app_id)
+    if not app or app.status != "normal":
+        raise NotFound("app not found")
+    if not app.enable_api:
+        raise Forbidden("service_api_disabled")
+    data.app = app
+
+
+def load_tenant(data: AuthData) -> None:
+    if data.app is None:
+        raise InternalServerError("pipeline_invariant_violated: app not loaded before load_tenant")
+    tenant = TenantService.get_tenant_by_id(db.session, str(data.app.tenant_id))
+    if tenant is None or tenant.status == TenantStatus.ARCHIVE:
+        raise Forbidden("workspace unavailable")
+    data.tenant = tenant
+
+
+def load_account(data: AuthData) -> None:
+    account = AccountService.get_account_by_id(db.session, str(data.account_id))
+    if account is None:
+        raise Unauthorized("account not found")
+    if data.tenant:
+        account.current_tenant = data.tenant
+    data.caller = account
+    data.caller_kind = "account"
+
+
+def resolve_external_user(data: AuthData) -> None:
+    if data.tenant is None or data.app is None or data.external_identity is None:
+        raise Unauthorized("missing context for external user resolution")
+    end_user = EndUserService.get_or_create_end_user_by_type(
+        InvokeFrom.OPENAPI,
+        tenant_id=str(data.tenant.id),
+        app_id=str(data.app.id),
+        user_id=data.external_identity.email,
+    )
+    data.caller = end_user
+    data.caller_kind = "end_user"
+
+
+def load_app_access_mode(data: AuthData) -> None:
+    if data.app is None:
+        return
+    try:
+        settings = EnterpriseService.WebAppAuth.get_app_access_mode_by_id(app_id=str(data.app.id))
+        if settings is None:
+            data.app_access_mode = None
+            return
+        data.app_access_mode = WebAppAccessMode(settings.access_mode)
+    except ValueError:
+        data.app_access_mode = None

api/controllers/openapi/auth/role_gate.py

@@ -0,0 +1,77 @@
+"""Workspace role gate.
+
+Layered on top of `validate_bearer` + `accept_subjects(SubjectType.ACCOUNT)`
+for routes whose access depends on the caller's `TenantAccountJoin.role`
+in the workspace named by the `workspace_id` path parameter.
+
+Usage::
+
+    @openapi_ns.route("/workspaces/<string:workspace_id>/members")
+    class Members(Resource):
+        @validate_bearer(accept=ACCEPT_USER_ANY)
+        @accept_subjects(SubjectType.ACCOUNT)
+        @require_workspace_role()  # any member
+        def get(self, workspace_id: str): ...
+
+        @validate_bearer(accept=ACCEPT_USER_ANY)
+        @accept_subjects(SubjectType.ACCOUNT)
+        @require_workspace_role(TenantAccountRole.OWNER, TenantAccountRole.ADMIN)
+        def post(self, workspace_id: str): ...
+
+Non-member callers get 404 (matching `GET /openapi/v1/workspaces/<id>`)
+so workspace IDs do not leak across tenants. A member without one of the
+allowed roles gets 403.
+"""
+
+from __future__ import annotations
+
+from collections.abc import Callable
+from functools import wraps
+from typing import TypeVar
+
+from werkzeug.exceptions import Forbidden, NotFound
+
+from extensions.ext_database import db
+from libs.oauth_bearer import try_get_auth_ctx
+from models.account import TenantAccountRole
+from services.account_service import TenantService
+
+F = TypeVar("F", bound=Callable[..., object])
+
+
+def require_workspace_role(*allowed_roles: TenantAccountRole) -> Callable[[F], F]:
+    """Gate a route on the caller's role in ``workspace_id``.
+
+    Pass no roles to require only membership. Pass one or more roles to
+    require the caller's role be in that set.
+    """
+
+    allowed = frozenset(allowed_roles)
+
+    def deco(fn: F) -> F:
+        @wraps(fn)
+        def wrapper(*args: object, **kwargs: object) -> object:
+            ctx = try_get_auth_ctx()
+            if ctx is None or ctx.account_id is None:
+                raise RuntimeError(
+                    "require_workspace_role called without account-bearer context; "
+                    "stack validate_bearer + accept_subjects(SubjectType.ACCOUNT) above it"
+                )
+
+            workspace_id = kwargs.get("workspace_id")
+            if not workspace_id:
+                raise RuntimeError("require_workspace_role expects a 'workspace_id' route parameter")
+
+            role = TenantService.get_account_role_in_tenant(db.session, str(ctx.account_id), str(workspace_id))
+
+            if role is None:
+                raise NotFound("workspace not found")
+
+            if allowed and role not in allowed:
+                raise Forbidden("insufficient workspace role")
+
+            return fn(*args, **kwargs)
+
+        return wrapper  # type: ignore[return-value]
+
+    return deco

api/controllers/openapi/auth/steps.py

@@ -1,170 +0,0 @@
-"""Pipeline steps. Each is one responsibility.
-
-`BearerCheck` is the only step that touches the token registry; downstream
-steps see only the populated `Context`. `BearerCheck` also publishes the
-resolved identity to the openapi auth ``ContextVar`` (the same one the
-decorator-level :func:`libs.oauth_bearer.validate_bearer` writes to) so the
-surface gate and any handler reading the request-scoped context has a single
-source of truth across both auth-attach paths. The reset token is stashed
-on `ctx.auth_ctx_reset_token`; `Pipeline.guard` resets the ContextVar in
-its `finally` so worker-thread reuse can't leak identity across requests.
-"""
-
-from __future__ import annotations
-
-from collections.abc import Callable
-
-from werkzeug.exceptions import BadRequest, Forbidden, NotFound, Unauthorized
-
-from configs import dify_config
-from controllers.openapi.auth.context import Context
-from controllers.openapi.auth.strategies import AppAuthzStrategy, CallerMounter
-from controllers.openapi.auth.surface_gate import check_surface
-from extensions.ext_database import db
-from libs.oauth_bearer import (
-    AuthContext,
-    InvalidBearerError,
-    Scope,
-    SubjectType,
-    check_workspace_membership,
-    get_authenticator,
-    set_auth_ctx,
-)
-from models import TenantStatus
-from services.account_service import TenantService
-from services.app_service import AppService
-
-
-class BearerCheck:
-    """Resolve bearer → populate identity fields. Rate-limit is enforced
-    inside `BearerAuthenticator.authenticate`, so no separate step here.
-    Also publishes the resolved `AuthContext` via
-    :func:`libs.oauth_bearer.set_auth_ctx` — same shape the decorator-level
-    ``validate_bearer`` writes — so the surface gate + downstream readers
-    don't see two different identity sources. The reset token is parked on
-    ``ctx.auth_ctx_reset_token`` for `Pipeline.guard` to consume."""
-
-    def __call__(self, ctx: Context) -> None:
-        if not ctx.bearer_token:
-            raise Unauthorized("bearer required")
-
-        try:
-            authn = get_authenticator().authenticate(ctx.bearer_token)
-        except InvalidBearerError as e:
-            raise Unauthorized(str(e))
-
-        ctx.subject_type = authn.subject_type
-        ctx.subject_email = authn.subject_email
-        ctx.subject_issuer = authn.subject_issuer
-        ctx.account_id = authn.account_id
-        ctx.scopes = frozenset(authn.scopes)
-        ctx.source = authn.source
-        ctx.token_id = authn.token_id
-        ctx.expires_at = authn.expires_at
-        ctx.token_hash = authn.token_hash
-        ctx.cached_verified_tenants = dict(authn.verified_tenants)
-        ctx.auth_ctx_reset_token = set_auth_ctx(authn)
-
-
-class ScopeCheck:
-    """Verify ctx.scopes (already populated by BearerCheck) covers required."""
-
-    def __call__(self, ctx: Context) -> None:
-        if Scope.FULL in ctx.scopes or ctx.required_scope in ctx.scopes:
-            return
-        raise Forbidden("insufficient_scope")
-
-
-class SurfaceCheck:
-    """Reject the request if the resolved subject is not in `accepted`."""
-
-    def __init__(self, *, accepted: frozenset[SubjectType]) -> None:
-        self._accepted = accepted
-
-    def __call__(self, ctx: Context) -> None:
-        check_surface(self._accepted)
-
-
-class AppResolver:
-    """Read ``app_id`` from ``ctx.path_params``; populate ctx.app + ctx.tenant.
-
-    Every endpoint using the OAuth bearer pipeline must declare
-    ``<string:app_id>`` in its route — that is the design lock-in (no body /
-    header coupling). ``Pipeline.guard`` lifts ``request.view_args`` into
-    ``ctx.path_params`` at the boundary so this step doesn't need to know
-    about the request object.
-    """
-
-    def __call__(self, ctx: Context) -> None:
-        app_id = ctx.path_params.get("app_id")
-        if not app_id:
-            raise BadRequest("app_id is required in path")
-        app = AppService.get_app_by_id(db.session, app_id)
-        if not app or app.status != "normal":
-            raise NotFound("app not found")
-        if not app.enable_api:
-            raise Forbidden("service_api_disabled")
-        tenant = TenantService.get_tenant_by_id(db.session, str(app.tenant_id))
-        if tenant is None or tenant.status == TenantStatus.ARCHIVE:
-            raise Forbidden("workspace unavailable")
-        ctx.app, ctx.tenant = app, tenant
-
-
-class WorkspaceMembershipCheck:
-    """Layer 0 — workspace membership gate.
-
-    CE-only (skipped when ENTERPRISE_ENABLED). Account-subject bearers
-    (dfoa_) only — SSO subjects skip.
-    """
-
-    def __call__(self, ctx: Context) -> None:
-        if dify_config.ENTERPRISE_ENABLED:
-            return
-        if ctx.subject_type != SubjectType.ACCOUNT:
-            return
-        if ctx.account_id is None or ctx.tenant is None:
-            raise Unauthorized("account_id or tenant unset — BearerCheck or AppResolver did not run")
-        if ctx.token_hash is None:
-            raise Unauthorized("token_hash unset — BearerCheck did not run")
-
-        check_workspace_membership(
-            account_id=ctx.account_id,
-            tenant_id=ctx.must_tenant.id,
-            token_hash=ctx.token_hash,
-            cached_verdicts=ctx.cached_verified_tenants or {},
-        )
-
-
-class AppAuthzCheck:
-    def __init__(self, resolve_strategy: Callable[[], AppAuthzStrategy]) -> None:
-        self._resolve = resolve_strategy
-
-    def __call__(self, ctx: Context) -> None:
-        if not self._resolve().authorize(ctx):
-            raise Forbidden("subject_no_app_access")
-
-
-class CallerMount:
-    def __init__(self, *mounters: CallerMounter) -> None:
-        self._mounters = mounters
-
-    def __call__(self, ctx: Context) -> None:
-        if ctx.subject_type is None:
-            raise Unauthorized("subject_type unset — BearerCheck did not run")
-        for m in self._mounters:
-            if m.applies_to(ctx.must_subject_type):
-                m.mount(ctx)
-                return
-        raise Unauthorized("no caller mounter for subject type")
-
-
-__all__ = [
-    "AppAuthzCheck",
-    "AppResolver",
-    "AuthContext",
-    "BearerCheck",
-    "CallerMount",
-    "ScopeCheck",
-    "SurfaceCheck",
-    "WorkspaceMembershipCheck",
-]

api/controllers/openapi/auth/strategies.py

@@ -1,168 +0,0 @@
-"""Strategy classes for the openapi auth pipeline.
-
-App authorization (Acl/Membership) and caller mounting (Account/EndUser)
-vary along independent axes; each strategy is one class so the pipeline
-composition stays a flat list.
-"""
-
-from __future__ import annotations
-
-from typing import Protocol
-
-from flask import current_app
-from flask_login import user_logged_in
-
-from controllers.openapi.auth.context import Context
-from core.app.entities.app_invoke_entities import InvokeFrom
-from extensions.ext_database import db
-from libs.oauth_bearer import SubjectType
-from services.account_service import AccountService, TenantService
-from services.end_user_service import EndUserService
-from services.enterprise.enterprise_service import (
-    EnterpriseService,
-    WebAppAccessMode,
-)
-
-
-class AppAuthzStrategy(Protocol):
-    def authorize(self, ctx: Context) -> bool: ...
-
-
-class AclStrategy:
-    """Per-app ACL, evaluated in two stages.
-
-    The EE gateway has already enforced tenancy and workspace membership
-    by the time this strategy runs, so AclStrategy only owns per-app ACL:
-
-    1. Subject vs access-mode compatibility (pure rule table). External-SSO
-       bearers belong to public-facing apps only; account bearers cover the
-       full set. A mismatch is an immediate deny — no IO.
-    2. For modes that pair with the subject, decide whether the inner
-       permission API must run. Only `PRIVATE` (per-app selected-user list)
-       requires it; the remaining modes are pass-through.
-    """
-
-    _ALLOWED_MODES_BY_SUBJECT: dict[SubjectType, frozenset[WebAppAccessMode]] = {
-        SubjectType.ACCOUNT: frozenset(
-            {
-                WebAppAccessMode.PUBLIC,
-                WebAppAccessMode.SSO_VERIFIED,
-                WebAppAccessMode.PRIVATE_ALL,
-                WebAppAccessMode.PRIVATE,
-            }
-        ),
-        SubjectType.EXTERNAL_SSO: frozenset(
-            {
-                WebAppAccessMode.PUBLIC,
-                WebAppAccessMode.SSO_VERIFIED,
-            }
-        ),
-    }
-
-    _MODES_REQUIRING_INNER_CHECK: frozenset[WebAppAccessMode] = frozenset({WebAppAccessMode.PRIVATE})
-
-    def authorize(self, ctx: Context) -> bool:
-        if ctx.app is None:
-            return False
-        access_mode = self._fetch_access_mode(ctx.app.id)
-        if access_mode is None:
-            return False
-        if not self._subject_allowed_for_mode(ctx.must_subject_type, access_mode):
-            return False
-        if access_mode not in self._MODES_REQUIRING_INNER_CHECK:
-            return True
-        return self._inner_permission_check(ctx)
-
-    @staticmethod
-    def _fetch_access_mode(app_id: str) -> WebAppAccessMode | None:
-        settings = EnterpriseService.WebAppAuth.get_app_access_mode_by_id(app_id=app_id)
-        if settings is None:
-            return None
-        try:
-            return WebAppAccessMode(settings.access_mode)
-        except ValueError:
-            return None
-
-    @classmethod
-    def _subject_allowed_for_mode(cls, subject_type: SubjectType, access_mode: WebAppAccessMode) -> bool:
-        return access_mode in cls._ALLOWED_MODES_BY_SUBJECT.get(subject_type, frozenset())
-
-    def _inner_permission_check(self, ctx: Context) -> bool:
-        if ctx.app is None:
-            return False
-        user_id = self._resolve_user_id(ctx)
-        if user_id is None:
-            return False
-        return EnterpriseService.WebAppAuth.is_user_allowed_to_access_webapp(
-            user_id=user_id,
-            app_id=ctx.app.id,
-        )
-
-    @staticmethod
-    def _resolve_user_id(ctx: Context) -> str | None:
-        if ctx.subject_type == SubjectType.ACCOUNT:
-            return str(ctx.account_id) if ctx.account_id is not None else None
-        if ctx.subject_email is None:
-            return None
-        account = AccountService.get_account_by_email(db.session, ctx.subject_email)
-        return str(account.id) if account is not None else None
-
-
-class MembershipStrategy:
-    """Tenant-membership fallback.
-
-    Used when webapp-auth is disabled (CE deployment). Account-bearing
-    subjects pass if they have a TenantAccountJoin row; EXTERNAL_SSO is
-    denied (it requires the webapp-auth surface).
-    """
-
-    def authorize(self, ctx: Context) -> bool:
-        if ctx.subject_type == SubjectType.EXTERNAL_SSO:
-            return False
-        if ctx.tenant is None:
-            return False
-        return TenantService.account_belongs_to_tenant(db.session, ctx.account_id, ctx.tenant.id)
-
-
-def _login_as(user) -> None:
-    """Set Flask-Login request user so downstream services see the caller."""
-    current_app.login_manager._update_request_context_with_user(user)  # type:ignore
-    user_logged_in.send(current_app._get_current_object(), user=user)  # type:ignore
-
-
-class CallerMounter(Protocol):
-    def applies_to(self, subject_type: SubjectType) -> bool: ...
-
-    def mount(self, ctx: Context) -> None: ...
-
-
-class AccountMounter:
-    def applies_to(self, subject_type: SubjectType) -> bool:
-        return subject_type == SubjectType.ACCOUNT
-
-    def mount(self, ctx: Context) -> None:
-        if ctx.account_id is None:
-            raise RuntimeError("AccountMounter: account_id unset — BearerCheck did not run")
-        account = AccountService.get_account_by_id(db.session, str(ctx.account_id))
-        if account is None:
-            raise RuntimeError("AccountMounter: account row missing for resolved bearer")
-        account.current_tenant = ctx.must_tenant
-        _login_as(account)
-        ctx.caller, ctx.caller_kind = account, "account"
-
-
-class EndUserMounter:
-    def applies_to(self, subject_type: SubjectType) -> bool:
-        return subject_type == SubjectType.EXTERNAL_SSO
-
-    def mount(self, ctx: Context) -> None:
-        if ctx.tenant is None or ctx.app is None or ctx.subject_email is None:
-            raise RuntimeError("EndUserMounter: tenant/app/subject_email unset — earlier steps did not run")
-        end_user = EndUserService.get_or_create_end_user_by_type(
-            InvokeFrom.OPENAPI,
-            tenant_id=ctx.tenant.id,
-            app_id=ctx.app.id,
-            user_id=ctx.subject_email,
-        )
-        _login_as(end_user)
-        ctx.caller, ctx.caller_kind = end_user, "end_user"

api/controllers/openapi/auth/verify.py

@@ -0,0 +1,82 @@
+from __future__ import annotations
+
+from werkzeug.exceptions import Forbidden, Unauthorized
+
+from controllers.openapi.auth.data import AuthData
+from extensions.ext_database import db
+from libs.oauth_bearer import Scope, TokenType, check_workspace_membership
+from services.account_service import AccountService, TenantService
+from services.enterprise.enterprise_service import EnterpriseService, WebAppAccessMode
+
+
+def check_scope(data: AuthData) -> None:
+    if data.required_scope is None:
+        return
+    if Scope.FULL in data.scopes or data.required_scope in data.scopes:
+        return
+    raise Forbidden("insufficient_scope")
+
+
+def check_membership(data: AuthData) -> None:
+    if data.tenant is None:
+        raise Unauthorized("tenant unset")
+    if data.account_id is None:
+        raise Unauthorized("account_id unset")
+    check_workspace_membership(
+        account_id=data.account_id,
+        tenant_id=data.tenant.id,
+        token_hash=data.token_hash,
+        membership_cache=data.tenants,
+    )
+
+
+def check_app_access(data: AuthData) -> None:
+    if data.tenant is None:
+        return
+    if not TenantService.account_belongs_to_tenant(db.session, data.account_id, data.tenant.id):
+        raise Forbidden("subject_no_app_access")
+
+
+_ALLOWED_MODES_BY_TOKEN_TYPE: dict[TokenType, frozenset[WebAppAccessMode]] = {
+    TokenType.OAUTH_ACCOUNT: frozenset(
+        {
+            WebAppAccessMode.PUBLIC,
+            WebAppAccessMode.SSO_VERIFIED,
+            WebAppAccessMode.PRIVATE_ALL,
+            WebAppAccessMode.PRIVATE,
+        }
+    ),
+    TokenType.OAUTH_EXTERNAL_SSO: frozenset(
+        {
+            WebAppAccessMode.PUBLIC,
+            WebAppAccessMode.SSO_VERIFIED,
+        }
+    ),
+}
+
+
+def check_acl(data: AuthData) -> None:
+    if data.app is None or data.app_access_mode is None:
+        raise Forbidden("app or access mode not loaded")
+    allowed_modes = _ALLOWED_MODES_BY_TOKEN_TYPE.get(data.token_type, frozenset())
+    if data.app_access_mode not in allowed_modes:
+        raise Forbidden("subject_not_allowed_for_access_mode")
+
+
+def check_private_app_permission(data: AuthData) -> None:
+    if data.app is None:
+        raise Forbidden("app not loaded")
+    user_id = _resolve_user_id(data)
+    if user_id is None:
+        raise Forbidden("cannot resolve user for private app check")
+    if not EnterpriseService.WebAppAuth.is_user_allowed_to_access_webapp(user_id=user_id, app_id=data.app.id):
+        raise Forbidden("user_not_allowed_for_private_app")
+
+
+def _resolve_user_id(data: AuthData) -> str | None:
+    if data.token_type == TokenType.OAUTH_ACCOUNT:
+        return str(data.account_id) if data.account_id is not None else None
+    if data.external_identity is None:
+        return None
+    account = AccountService.get_account_by_email(db.session, data.external_identity.email)
+    return str(account.id) if account is not None else None

api/controllers/openapi/files.py

@@ -17,11 +17,11 @@ from controllers.common.errors import (
     UnsupportedFileTypeError,
 )
 from controllers.openapi import openapi_ns
-from controllers.openapi.auth.composition import OAUTH_BEARER_PIPELINE
+from controllers.openapi.auth.composition import auth_router
+from controllers.openapi.auth.data import AuthData
 from extensions.ext_database import db
 from fields.file_fields import FileResponse
 from libs.oauth_bearer import Scope
-from models import Account, App
 from services.file_service import FileService
 
 
@@ -39,8 +39,9 @@ class AppFileUploadApi(Resource):
         }
     )
     @openapi_ns.response(HTTPStatus.CREATED, "File uploaded", openapi_ns.models[FileResponse.__name__])
-    @OAUTH_BEARER_PIPELINE.guard(scope=Scope.APPS_RUN)
-    def post(self, app_id: str, app_model: App, caller: Account, caller_kind: str):
+    @auth_router.guard(scope=Scope.APPS_RUN)
+    def post(self, app_id: str, *, auth_data: AuthData):
+        app_model, caller, _ = auth_data.require_app_context()
         if "file" not in request.files:
             raise NoFileUploadedError()
         if len(request.files) > 1:

api/controllers/openapi/human_input_form.py

@@ -17,7 +17,8 @@ from werkzeug.exceptions import BadRequest, NotFound
 from controllers.common.human_input import HumanInputFormSubmitPayload, stringify_form_default_values
 from controllers.common.schema import register_schema_models
 from controllers.openapi import openapi_ns
-from controllers.openapi.auth.composition import OAUTH_BEARER_PIPELINE
+from controllers.openapi.auth.composition import auth_router
+from controllers.openapi.auth.data import AuthData
 from core.workflow.human_input_policy import HumanInputSurface, is_recipient_type_allowed_for_surface
 from extensions.ext_database import db
 from libs.helper import to_timestamp
@@ -55,8 +56,9 @@ def _ensure_form_is_allowed_for_openapi(form) -> None:
 @openapi_ns.route("/apps/<string:app_id>/form/human_input/<string:form_token>")
 class OpenApiWorkflowHumanInputFormApi(Resource):
     @openapi_ns.response(200, "Form definition")
-    @OAUTH_BEARER_PIPELINE.guard(scope=Scope.APPS_RUN)
-    def get(self, app_id: str, form_token: str, app_model: App, caller, caller_kind: str):
+    @auth_router.guard(scope=Scope.APPS_RUN)
+    def get(self, app_id: str, form_token: str, *, auth_data: AuthData):
+        app_model, caller, caller_kind = auth_data.require_app_context()
         service = HumanInputService(db.engine)
         form = service.get_form_by_token(form_token)
         if form is None:
@@ -69,8 +71,9 @@ class OpenApiWorkflowHumanInputFormApi(Resource):
 
     @openapi_ns.expect(openapi_ns.models[HumanInputFormSubmitPayload.__name__])
     @openapi_ns.response(200, "Form submitted")
-    @OAUTH_BEARER_PIPELINE.guard(scope=Scope.APPS_RUN)
-    def post(self, app_id: str, form_token: str, app_model: App, caller, caller_kind: str):
+    @auth_router.guard(scope=Scope.APPS_RUN)
+    def post(self, app_id: str, form_token: str, *, auth_data: AuthData):
+        app_model, caller, caller_kind = auth_data.require_app_context()
         payload = HumanInputFormSubmitPayload.model_validate(request.get_json(silent=True) or {})
 
         service = HumanInputService(db.engine)

api/controllers/openapi/workflow_events.py

@@ -17,7 +17,8 @@ from sqlalchemy.orm import sessionmaker
 from werkzeug.exceptions import NotFound, UnprocessableEntity
 
 from controllers.openapi import openapi_ns
-from controllers.openapi.auth.composition import OAUTH_BEARER_PIPELINE
+from controllers.openapi.auth.composition import auth_router
+from controllers.openapi.auth.data import AuthData
 from core.app.apps.advanced_chat.app_generator import AdvancedChatAppGenerator
 from core.app.apps.base_app_generator import BaseAppGenerator
 from core.app.apps.common.workflow_response_converter import WorkflowResponseConverter
@@ -28,7 +29,7 @@ from core.workflow.human_input_policy import HumanInputSurface
 from extensions.ext_database import db
 from libs.oauth_bearer import Scope
 from models.enums import CreatorUserRole
-from models.model import App, AppMode
+from models.model import AppMode
 from repositories.factory import DifyAPIRepositoryFactory
 from services.workflow_event_snapshot_service import build_workflow_event_stream
 
@@ -36,8 +37,9 @@ from services.workflow_event_snapshot_service import build_workflow_event_stream
 @openapi_ns.route("/apps/<string:app_id>/tasks/<string:task_id>/events")
 class OpenApiWorkflowEventsApi(Resource):
     @openapi_ns.response(200, "SSE event stream")
-    @OAUTH_BEARER_PIPELINE.guard(scope=Scope.APPS_RUN)
-    def get(self, app_id: str, task_id: str, app_model: App, caller, caller_kind: str):
+    @auth_router.guard(scope=Scope.APPS_RUN)
+    def get(self, app_id: str, task_id: str, *, auth_data: AuthData):
+        app_model, caller, caller_kind = auth_data.require_app_context()
         app_mode = AppMode.value_of(app_model.mode)
         if app_mode not in {AppMode.WORKFLOW, AppMode.ADVANCED_CHAT}:
             raise UnprocessableEntity("mode_not_supported_for_event_reconnect")

api/controllers/openapi/workspaces.py

@@ -1,41 +1,129 @@
-"""User-scoped workspace reads under /openapi/v1/workspaces. Bearer-authed
-counterparts to the cookie-authed /console/api/workspaces endpoints.
+"""User-scoped workspace reads and member management under /openapi/v1/workspaces.
 
-Account bearers (dfoa_) see every tenant they're a member of. External
-SSO bearers (dfoe_) have no account_id and so see an empty list — that
-matches /openapi/v1/account.
+Bearer-authed counterparts to the cookie-authed /console/api/workspaces
+endpoints. Account bearers (dfoa_) see every tenant they're a member of.
+External SSO bearers (dfoe_) have no account_id and so see an empty list —
+that matches /openapi/v1/account.
+
+Member-management endpoints are gated by both `accept_subjects` (SSO out)
+and `require_workspace_role` (membership / role lookup against the path's
+``workspace_id``).
 """
 
 from __future__ import annotations
 
 from itertools import starmap
+from urllib import parse
 
+from flask import jsonify, make_response, request
 from flask_restx import Resource
-from werkzeug.exceptions import NotFound
+from pydantic import BaseModel, ValidationError
+from werkzeug.exceptions import BadRequest, Forbidden, NotFound
 
+from configs import dify_config
+from controllers.common.schema import query_params_from_model
 from controllers.openapi import openapi_ns
-from controllers.openapi._models import WorkspaceDetailResponse, WorkspaceListResponse, WorkspaceSummaryResponse
-from controllers.openapi.auth.surface_gate import accept_subjects
-from extensions.ext_database import db
-from libs.oauth_bearer import (
-    ACCEPT_USER_ANY,
-    SubjectType,
-    get_auth_ctx,
-    validate_bearer,
+from controllers.openapi._models import (
+    MemberActionResponse,
+    MemberInvitePayload,
+    MemberInviteResponse,
+    MemberListQuery,
+    MemberListResponse,
+    MemberResponse,
+    MemberRoleUpdatePayload,
+    WorkspaceDetailResponse,
+    WorkspaceListResponse,
+    WorkspaceSummaryResponse,
 )
-from models import Tenant, TenantAccountJoin
-from services.account_service import TenantService
+from controllers.openapi.auth.composition import auth_router
+from controllers.openapi.auth.data import AuthData
+from controllers.openapi.auth.role_gate import require_workspace_role
+from extensions.ext_database import db
+from libs.oauth_bearer import Scope, TokenType
+from models import Account, Tenant, TenantAccountJoin
+from models.account import TenantAccountRole, TenantStatus
+from services.account_service import AccountService, RegisterService, TenantService
+from services.errors.account import (
+    AccountAlreadyInTenantError,
+    AccountNotLinkTenantError,
+    AccountRegisterError,
+    CannotOperateSelfError,
+    MemberNotInTenantError,
+    NoPermissionError,
+    RoleAlreadyAssignedError,
+)
+from services.feature_service import FeatureService
+
+
+def _validate_body[M: BaseModel](model: type[M]) -> M:
+    body = request.get_json(silent=True) or {}
+    try:
+        return model.model_validate(body)
+    except ValidationError as exc:
+        raise BadRequest(str(exc))
+
+
+def _member_response(account: Account) -> MemberResponse:
+    return MemberResponse(
+        id=str(account.id),
+        name=account.name,
+        email=account.email,
+        role=account.role.value if account.role else "",
+        status=account.status.value if account.status else "",
+        avatar=account.avatar,
+    )
+
+
+def _load_tenant(workspace_id: str) -> Tenant:
+    tenant = TenantService.get_tenant_by_id(db.session, workspace_id)
+    if tenant is None or tenant.status != TenantStatus.NORMAL:
+        raise NotFound("workspace not found")
+    return tenant
+
+
+def _load_account(account_id: object) -> Account:
+    account = AccountService.get_account_by_id(db.session, str(account_id)) if account_id else None
+    if account is None:
+        raise RuntimeError("authenticated account_id has no Account row")
+    return account
+
+
+def _quota_error(*, code: str, message: str, hint: str) -> Forbidden:
+    err = Forbidden(message)
+    err.response = make_response(
+        jsonify({"code": code, "message": message, "hint": hint}),
+        403,
+    )
+    return err
+
+
+def _check_member_invite_quota(tenant_id: str) -> None:
+    features = FeatureService.get_features(tenant_id)
+
+    if features.billing.enabled:
+        members = features.members
+        if 0 < members.limit <= members.size:
+            raise _quota_error(
+                code="members.limit_exceeded",
+                message="Subscription member limit reached.",
+                hint="Upgrade your plan to invite more members or remove an existing member first.",
+            )
+
+    if features.workspace_members.enabled:
+        if not features.workspace_members.is_available(1):
+            raise _quota_error(
+                code="workspace_members.license_exceeded",
+                message="Workspace member license capacity reached.",
+                hint="Contact your workspace administrator to expand the license seat count.",
+            )
 
 
 @openapi_ns.route("/workspaces")
 class WorkspacesApi(Resource):
     @openapi_ns.response(200, "Workspace list", openapi_ns.models[WorkspaceListResponse.__name__])
-    @validate_bearer(accept=ACCEPT_USER_ANY)
-    @accept_subjects(SubjectType.ACCOUNT)
-    def get(self):
-        ctx = get_auth_ctx()
-
-        rows = TenantService.get_workspaces_for_account(db.session, str(ctx.account_id))
+    @auth_router.guard(scope=Scope.WORKSPACE_READ, allowed_token_types=frozenset({TokenType.OAUTH_ACCOUNT}))
+    def get(self, *, auth_data: AuthData):
+        rows = TenantService.get_workspaces_for_account(db.session, str(auth_data.account_id))
 
         return WorkspaceListResponse(workspaces=list(starmap(_workspace_summary, rows))).model_dump(mode="json"), 200
 
@@ -43,12 +131,9 @@ class WorkspacesApi(Resource):
 @openapi_ns.route("/workspaces/<string:workspace_id>")
 class WorkspaceByIdApi(Resource):
     @openapi_ns.response(200, "Workspace detail", openapi_ns.models[WorkspaceDetailResponse.__name__])
-    @validate_bearer(accept=ACCEPT_USER_ANY)
-    @accept_subjects(SubjectType.ACCOUNT)
-    def get(self, workspace_id: str):
-        ctx = get_auth_ctx()
-
-        row = TenantService.find_workspace_for_account(db.session, str(ctx.account_id), workspace_id)
+    @auth_router.guard(scope=Scope.WORKSPACE_READ, allowed_token_types=frozenset({TokenType.OAUTH_ACCOUNT}))
+    def get(self, workspace_id: str, *, auth_data: AuthData):
+        row = TenantService.find_workspace_for_account(db.session, str(auth_data.account_id), workspace_id)
         # 404 (not 403) on non-member so workspace IDs don't leak across tenants.
         if row is None:
             raise NotFound("workspace not found")
@@ -57,6 +142,172 @@ class WorkspaceByIdApi(Resource):
         return _workspace_detail(tenant, membership).model_dump(mode="json"), 200
 
 
+@openapi_ns.route("/workspaces/<string:workspace_id>/switch")
+class WorkspaceSwitchApi(Resource):
+    """Server-side switch — equivalent to the console's POST /workspaces/switch.
+
+    CLI `difyctl use workspace <id>` calls this; it does NOT mutate
+    ``hosts.yml`` on its own. Failure here must abort the local write so
+    that ``hosts.yml`` never diverges from the server's ``current`` state.
+    """
+
+    @openapi_ns.response(200, "Workspace detail", openapi_ns.models[WorkspaceDetailResponse.__name__])
+    @auth_router.guard(scope=Scope.WORKSPACE_READ, allowed_token_types=frozenset({TokenType.OAUTH_ACCOUNT}))
+    @require_workspace_role()
+    def post(self, workspace_id: str, *, auth_data: AuthData):
+        account = _load_account(auth_data.account_id)
+
+        try:
+            TenantService.switch_tenant(account, workspace_id)
+        except AccountNotLinkTenantError:
+            raise NotFound("workspace not found")
+
+        row = TenantService.find_workspace_for_account(db.session, str(auth_data.account_id), workspace_id)
+        if row is None:
+            raise NotFound("workspace not found")
+        tenant, membership = row
+        return _workspace_detail(tenant, membership).model_dump(mode="json"), 200
+
+
+@openapi_ns.route("/workspaces/<string:workspace_id>/members")
+class WorkspaceMembersApi(Resource):
+    """List + invite members.
+
+    GET is any-member. POST requires admin/owner — owner can never be
+    assigned through invite (ownership transfer is console-only).
+    """
+
+    @openapi_ns.doc(params=query_params_from_model(MemberListQuery))
+    @openapi_ns.response(200, "Member list", openapi_ns.models[MemberListResponse.__name__])
+    @auth_router.guard(scope=Scope.WORKSPACE_READ, allowed_token_types=frozenset({TokenType.OAUTH_ACCOUNT}))
+    @require_workspace_role()
+    def get(self, workspace_id: str, *, auth_data: AuthData):
+        try:
+            query = MemberListQuery.model_validate(request.args.to_dict(flat=True))
+        except ValidationError as exc:
+            raise BadRequest(str(exc))
+
+        tenant = _load_tenant(workspace_id)
+        members = TenantService.get_tenant_members(tenant)
+        total = len(members)
+        start = (query.page - 1) * query.limit
+        page_items = members[start : start + query.limit]
+        return MemberListResponse(
+            page=query.page,
+            limit=query.limit,
+            total=total,
+            has_more=query.page * query.limit < total,
+            data=[_member_response(m) for m in page_items],
+        ).model_dump(mode="json"), 200
+
+    @openapi_ns.expect(openapi_ns.models[MemberInvitePayload.__name__])
+    @openapi_ns.response(201, "Member invited", openapi_ns.models[MemberInviteResponse.__name__])
+    @auth_router.guard(scope=Scope.WORKSPACE_WRITE, allowed_token_types=frozenset({TokenType.OAUTH_ACCOUNT}))
+    @require_workspace_role(TenantAccountRole.OWNER, TenantAccountRole.ADMIN)
+    def post(self, workspace_id: str, *, auth_data: AuthData):
+        payload = _validate_body(MemberInvitePayload)
+        inviter = _load_account(auth_data.account_id)
+        tenant = _load_tenant(workspace_id)
+
+        _check_member_invite_quota(str(tenant.id))
+
+        try:
+            token = RegisterService.invite_new_member(
+                tenant=tenant,
+                email=payload.email,
+                language=None,
+                role=payload.role,
+                inviter=inviter,
+            )
+        except AccountAlreadyInTenantError as exc:
+            raise BadRequest(str(exc))
+        except NoPermissionError as exc:
+            raise BadRequest(str(exc))
+        except AccountRegisterError as exc:
+            raise BadRequest(str(exc))
+
+        normalized_email = payload.email.lower()
+        member = AccountService.get_account_by_email_with_case_fallback(normalized_email)
+        if member is None:
+            # invite_new_member just created or fetched this account.
+            raise RuntimeError("invited member missing from DB after invite")
+
+        encoded_email = parse.quote(normalized_email)
+        invite_url = f"{dify_config.CONSOLE_WEB_URL}/activate?email={encoded_email}&token={token}"
+        return MemberInviteResponse(
+            email=normalized_email,
+            role=payload.role,
+            member_id=str(member.id),
+            invite_url=invite_url,
+            tenant_id=str(tenant.id),
+        ).model_dump(mode="json"), 201
+
+
+@openapi_ns.route("/workspaces/<string:workspace_id>/members/<string:member_id>")
+class WorkspaceMemberApi(Resource):
+    """Remove a member.
+
+    Self-removal and owner-removal are explicitly rejected by the service
+    layer (CannotOperateSelfError, NoPermissionError) — both surface as
+    400 per the spec, with the service's message preserved.
+    """
+
+    @openapi_ns.response(200, "Member removed", openapi_ns.models[MemberActionResponse.__name__])
+    @auth_router.guard(scope=Scope.WORKSPACE_WRITE, allowed_token_types=frozenset({TokenType.OAUTH_ACCOUNT}))
+    @require_workspace_role(TenantAccountRole.OWNER, TenantAccountRole.ADMIN)
+    def delete(self, workspace_id: str, member_id: str, *, auth_data: AuthData):
+        operator = _load_account(auth_data.account_id)
+        tenant = _load_tenant(workspace_id)
+        member = AccountService.get_account_by_id(db.session, member_id)
+        if member is None:
+            raise NotFound("member not found")
+
+        try:
+            TenantService.remove_member_from_tenant(tenant, member, operator)
+        except CannotOperateSelfError as exc:
+            raise BadRequest(str(exc))
+        except NoPermissionError as exc:
+            raise BadRequest(str(exc))
+        except MemberNotInTenantError as exc:
+            raise NotFound(str(exc))
+
+        return MemberActionResponse().model_dump(mode="json"), 200
+
+
+@openapi_ns.route("/workspaces/<string:workspace_id>/members/<string:member_id>/role")
+class WorkspaceMemberRoleApi(Resource):
+    """Change a member's role.
+
+    Owner cannot be assigned here (closed enum). Admin cannot demote the
+    standing owner (service NoPermissionError → 400, per spec).
+    """
+
+    @openapi_ns.expect(openapi_ns.models[MemberRoleUpdatePayload.__name__])
+    @openapi_ns.response(200, "Role updated", openapi_ns.models[MemberActionResponse.__name__])
+    @auth_router.guard(scope=Scope.WORKSPACE_WRITE, allowed_token_types=frozenset({TokenType.OAUTH_ACCOUNT}))
+    @require_workspace_role(TenantAccountRole.OWNER, TenantAccountRole.ADMIN)
+    def put(self, workspace_id: str, member_id: str, *, auth_data: AuthData):
+        payload = _validate_body(MemberRoleUpdatePayload)
+        operator = _load_account(auth_data.account_id)
+        tenant = _load_tenant(workspace_id)
+        member = AccountService.get_account_by_id(db.session, member_id)
+        if member is None:
+            raise NotFound("member not found")
+
+        try:
+            TenantService.update_member_role(tenant, member, payload.role, operator)
+        except CannotOperateSelfError as exc:
+            raise BadRequest(str(exc))
+        except NoPermissionError as exc:
+            raise BadRequest(str(exc))
+        except MemberNotInTenantError as exc:
+            raise NotFound(str(exc))
+        except RoleAlreadyAssignedError as exc:
+            raise BadRequest(str(exc))
+
+        return MemberActionResponse().model_dump(mode="json"), 200
+
+
 def _workspace_summary(tenant: Tenant, membership: TenantAccountJoin) -> WorkspaceSummaryResponse:
     return WorkspaceSummaryResponse(
         id=str(tenant.id),

api/controllers/service_api/app/annotation.py

@@ -6,7 +6,7 @@ from flask_restx import Resource
 from flask_restx.api import HTTPStatus
 from pydantic import BaseModel, Field, TypeAdapter
 
-from controllers.common.schema import register_schema_models
+from controllers.common.schema import query_params_from_model, register_schema_models
 from controllers.console.wraps import edit_permission_required
 from controllers.service_api import service_api_ns
 from controllers.service_api.wraps import validate_app_token
@@ -32,8 +32,19 @@ class AnnotationReplyActionPayload(BaseModel):
     embedding_model_name: str = Field(description="Embedding model name")
 
 
+class AnnotationListQuery(BaseModel):
+    page: int = Field(default=1, ge=1, description="Page number")
+    limit: int = Field(default=20, ge=1, description="Number of annotations per page")
+    keyword: str = Field(default="", description="Keyword to search annotations")
+
+
 register_schema_models(
-    service_api_ns, AnnotationCreatePayload, AnnotationReplyActionPayload, Annotation, AnnotationList
+    service_api_ns,
+    AnnotationCreatePayload,
+    AnnotationReplyActionPayload,
+    AnnotationListQuery,
+    Annotation,
+    AnnotationList,
 )
 
 
@@ -100,6 +111,7 @@ class AnnotationReplyActionStatusApi(Resource):
 class AnnotationListApi(Resource):
     @service_api_ns.doc("list_annotations")
     @service_api_ns.doc(description="List annotations for the application")
+    @service_api_ns.doc(params=query_params_from_model(AnnotationListQuery))
     @service_api_ns.doc(
         responses={
             200: "Annotations retrieved successfully",
@@ -114,18 +126,18 @@ class AnnotationListApi(Resource):
     @validate_app_token
     def get(self, app_model: App):
         """List annotations for the application."""
-        page = request.args.get("page", default=1, type=int)
-        limit = request.args.get("limit", default=20, type=int)
-        keyword = request.args.get("keyword", default="", type=str)
+        query = AnnotationListQuery.model_validate(request.args.to_dict(flat=True))
 
-        annotation_list, total = AppAnnotationService.get_annotation_list_by_app_id(app_model.id, page, limit, keyword)
+        annotation_list, total = AppAnnotationService.get_annotation_list_by_app_id(
+            app_model.id, query.page, query.limit, query.keyword
+        )
         annotation_models = TypeAdapter(list[Annotation]).validate_python(annotation_list, from_attributes=True)
         response = AnnotationList(
             data=annotation_models,
-            has_more=len(annotation_list) == limit,
-            limit=limit,
+            has_more=len(annotation_list) == query.limit,
+            limit=query.limit,
             total=total,
-            page=page,
+            page=query.page,
         )
         return response.model_dump(mode="json")
 

api/controllers/service_api/dataset/hit_testing.py

@@ -1,11 +1,14 @@
 from uuid import UUID
 
-from controllers.common.schema import register_schema_model
+from controllers.common.schema import register_response_schema_models, register_schema_models
 from controllers.console.datasets.hit_testing_base import DatasetsHitTestingBase, HitTestingPayload
 from controllers.service_api import service_api_ns
 from controllers.service_api.wraps import DatasetApiResource, cloud_edition_billing_rate_limit_check
+from fields.hit_testing_fields import HitTestingResponse
+from libs.helper import dump_response
 
-register_schema_model(service_api_ns, HitTestingPayload)
+register_schema_models(service_api_ns, HitTestingPayload)
+register_response_schema_models(service_api_ns, HitTestingResponse)
 
 
 @service_api_ns.route("/datasets/<uuid:dataset_id>/hit-testing", "/datasets/<uuid:dataset_id>/retrieve")
@@ -13,16 +16,16 @@ class HitTestingApi(DatasetApiResource, DatasetsHitTestingBase):
     @service_api_ns.doc("dataset_hit_testing")
     @service_api_ns.doc(description="Perform hit testing on a dataset")
     @service_api_ns.doc(params={"dataset_id": "Dataset ID"})
-    @service_api_ns.doc(
-        responses={
-            200: "Hit testing results",
-            401: "Unauthorized - invalid API token",
-            404: "Dataset not found",
-        }
+    @service_api_ns.response(
+        200,
+        "Hit testing results",
+        model=service_api_ns.models[HitTestingResponse.__name__],
     )
+    @service_api_ns.response(401, "Unauthorized - invalid API token")
+    @service_api_ns.response(404, "Dataset not found")
     @service_api_ns.expect(service_api_ns.models[HitTestingPayload.__name__])
     @cloud_edition_billing_rate_limit_check("knowledge", "dataset")
-    def post(self, tenant_id, dataset_id: UUID):
+    def post(self, tenant_id: str, dataset_id: UUID) -> dict[str, object]:
         """Perform hit testing on a dataset.
 
         Tests retrieval performance for the specified dataset.
@@ -33,4 +36,4 @@ class HitTestingApi(DatasetApiResource, DatasetsHitTestingBase):
         args = self.parse_args(service_api_ns.payload)
         self.hit_testing_args_check(args)
 
-        return self.perform_hit_testing(dataset, args)
+        return dump_response(HitTestingResponse, self.perform_hit_testing(dataset, args))

api/controllers/service_api/wraps.py

@@ -13,6 +13,7 @@ from pydantic import BaseModel
 from sqlalchemy import select
 from werkzeug.exceptions import Forbidden, NotFound, Unauthorized
 
+from configs import dify_config
 from enums.cloud_plan import CloudPlan
 from extensions.ext_database import db
 from extensions.ext_redis import redis_client
@@ -140,20 +141,26 @@ def cloud_edition_billing_resource_check[**P, R](
     def interceptor(view: Callable[P, R]):
         def decorated(*args: P.args, **kwargs: P.kwargs):
             api_token = validate_and_get_api_token(api_token_type)
-            features = FeatureService.get_features(api_token.tenant_id)
+            if resource == "vector_space":
+                if not dify_config.BILLING_ENABLED:
+                    return view(*args, **kwargs)
+
+                vector_space = FeatureService.get_vector_space(api_token.tenant_id)
+                if 0 < vector_space.limit <= vector_space.size:
+                    raise Forbidden("The capacity of the vector space has reached the limit of your subscription.")
+                return view(*args, **kwargs)
+
+            features = FeatureService.get_features(api_token.tenant_id, exclude_vector_space=True)
 
             if features.billing.enabled:
                 members = features.members
                 apps = features.apps
-                vector_space = features.vector_space
                 documents_upload_quota = features.documents_upload_quota
 
                 if resource == "members" and 0 < members.limit <= members.size:
                     raise Forbidden("The number of members has reached the limit of your subscription.")
                 elif resource == "apps" and 0 < apps.limit <= apps.size:
                     raise Forbidden("The number of apps has reached the limit of your subscription.")
-                elif resource == "vector_space" and 0 < vector_space.limit <= vector_space.size:
-                    raise Forbidden("The capacity of the vector space has reached the limit of your subscription.")
                 elif resource == "documents" and 0 < documents_upload_quota.limit <= documents_upload_quota.size:
                     raise Forbidden("The number of documents has reached the limit of your subscription.")
                 else:
@@ -174,7 +181,7 @@ def cloud_edition_billing_knowledge_limit_check[**P, R](
         @wraps(view)
         def decorated(*args: P.args, **kwargs: P.kwargs):
             api_token = validate_and_get_api_token(api_token_type)
-            features = FeatureService.get_features(api_token.tenant_id)
+            features = FeatureService.get_features(api_token.tenant_id, exclude_vector_space=True)
             if features.billing.enabled:
                 if resource == "add_segment":
                     if features.billing.subscription.plan == CloudPlan.SANDBOX:

api/controllers/web/app.py

@@ -12,7 +12,7 @@ from controllers.common.schema import register_response_schema_models, register_
 from core.app.app_config.common.parameters_mapping import get_parameters_from_feature_dict
 from libs.passport import PassportService
 from libs.token import extract_webapp_passport
-from models.model import App, AppMode
+from models.model import App, AppMode, EndUser
 from services.app_service import AppService
 from services.enterprise.enterprise_service import EnterpriseService
 from services.feature_service import FeatureService
@@ -56,7 +56,7 @@ class AppParameterApi(WebApiResource):
             500: "Internal Server Error",
         }
     )
-    def get(self, app_model: App, end_user):
+    def get(self, app_model: App, end_user: EndUser):
         """Retrieve app parameters."""
         if app_model.mode in {AppMode.ADVANCED_CHAT, AppMode.WORKFLOW}:
             workflow = app_model.workflow
@@ -92,7 +92,7 @@ class AppMeta(WebApiResource):
             500: "Internal Server Error",
         }
     )
-    def get(self, app_model: App, end_user):
+    def get(self, app_model: App, end_user: EndUser):
         """Get app meta"""
         return AppService().get_app_meta(app_model)
 

api/controllers/web/completion.py

@@ -29,7 +29,7 @@ from core.errors.error import (
 from graphon.model_runtime.errors.invoke import InvokeError
 from libs import helper
 from libs.helper import uuid_value
-from models.model import AppMode
+from models.model import App, AppMode, EndUser
 from services.app_generate_service import AppGenerateService
 from services.app_task_service import AppTaskService
 from services.errors.llm import InvokeRateLimitError
@@ -86,7 +86,7 @@ class CompletionApi(WebApiResource):
             500: "Internal Server Error",
         }
     )
-    def post(self, app_model, end_user):
+    def post(self, app_model: App, end_user: EndUser):
         if app_model.mode != AppMode.COMPLETION:
             raise NotCompletionAppError()
 
@@ -140,7 +140,7 @@ class CompletionStopApi(WebApiResource):
         }
     )
     @web_ns.response(200, "Success", web_ns.models[SimpleResultResponse.__name__])
-    def post(self, app_model, end_user, task_id: str):
+    def post(self, app_model: App, end_user: EndUser, task_id: str):
         if app_model.mode != AppMode.COMPLETION:
             raise NotCompletionAppError()
 
@@ -169,7 +169,7 @@ class ChatApi(WebApiResource):
             500: "Internal Server Error",
         }
     )
-    def post(self, app_model, end_user):
+    def post(self, app_model: App, end_user: EndUser):
         app_mode = AppMode.value_of(app_model.mode)
         if app_mode not in {AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT}:
             raise NotChatAppError()
@@ -226,7 +226,7 @@ class ChatStopApi(WebApiResource):
         }
     )
     @web_ns.response(200, "Success", web_ns.models[SimpleResultResponse.__name__])
-    def post(self, app_model, end_user, task_id: str):
+    def post(self, app_model: App, end_user: EndUser, task_id: str):
         app_mode = AppMode.value_of(app_model.mode)
         if app_mode not in {AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT}:
             raise NotChatAppError()

api/controllers/web/conversation.py

@@ -19,7 +19,7 @@ from fields.conversation_fields import (
     SimpleConversation,
 )
 from libs.helper import uuid_value
-from models.model import AppMode
+from models.model import App, AppMode, EndUser
 from services.conversation_service import ConversationService
 from services.errors.conversation import ConversationNotExistsError, LastConversationNotExistsError
 from services.web_conversation_service import WebConversationService
@@ -81,7 +81,7 @@ class ConversationListApi(WebApiResource):
             500: "Internal Server Error",
         }
     )
-    def get(self, app_model, end_user):
+    def get(self, app_model: App, end_user: EndUser):
         app_mode = AppMode.value_of(app_model.mode)
         if app_mode not in {AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT}:
             raise NotChatAppError()
@@ -127,7 +127,7 @@ class ConversationApi(WebApiResource):
             500: "Internal Server Error",
         }
     )
-    def delete(self, app_model, end_user, c_id: UUID):
+    def delete(self, app_model: App, end_user: EndUser, c_id: UUID):
         app_mode = AppMode.value_of(app_model.mode)
         if app_mode not in {AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT}:
             raise NotChatAppError()
@@ -166,7 +166,7 @@ class ConversationRenameApi(WebApiResource):
             500: "Internal Server Error",
         }
     )
-    def post(self, app_model, end_user, c_id: UUID):
+    def post(self, app_model: App, end_user: EndUser, c_id: UUID):
         app_mode = AppMode.value_of(app_model.mode)
         if app_mode not in {AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT}:
             raise NotChatAppError()
@@ -204,7 +204,7 @@ class ConversationPinApi(WebApiResource):
         }
     )
     @web_ns.response(200, "Conversation pinned successfully", web_ns.models[ResultResponse.__name__])
-    def patch(self, app_model, end_user, c_id: UUID):
+    def patch(self, app_model: App, end_user: EndUser, c_id: UUID):
         app_mode = AppMode.value_of(app_model.mode)
         if app_mode not in {AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT}:
             raise NotChatAppError()
@@ -235,7 +235,7 @@ class ConversationUnPinApi(WebApiResource):
         }
     )
     @web_ns.response(200, "Conversation unpinned successfully", web_ns.models[ResultResponse.__name__])
-    def patch(self, app_model, end_user, c_id: UUID):
+    def patch(self, app_model: App, end_user: EndUser, c_id: UUID):
         app_mode = AppMode.value_of(app_model.mode)
         if app_mode not in {AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT}:
             raise NotChatAppError()

api/controllers/web/files.py

@@ -13,6 +13,7 @@ from controllers.web import web_ns
 from controllers.web.wraps import WebApiResource
 from extensions.ext_database import db
 from fields.file_fields import FileResponse
+from models.model import App, EndUser
 from services.file_service import FileService
 
 register_schema_models(web_ns, FileResponse)
@@ -31,7 +32,7 @@ class FileApi(WebApiResource):
         }
     )
     @web_ns.response(201, "File uploaded successfully", web_ns.models[FileResponse.__name__])
-    def post(self, app_model, end_user):
+    def post(self, app_model: App, end_user: EndUser):
         """Upload a file for use in web applications.
 
         Accepts file uploads for use within web applications, supporting

api/controllers/web/message.py

@@ -27,7 +27,7 @@ from fields.message_fields import SuggestedQuestionsResponse, WebMessageInfinite
 from graphon.model_runtime.errors.invoke import InvokeError
 from libs import helper
 from models.enums import FeedbackRating
-from models.model import AppMode
+from models.model import App, AppMode, EndUser
 from services.app_generate_service import AppGenerateService
 from services.errors.app import MoreLikeThisDisabledError
 from services.errors.conversation import ConversationNotExistsError
@@ -81,7 +81,7 @@ class MessageListApi(WebApiResource):
             500: "Internal Server Error",
         }
     )
-    def get(self, app_model, end_user):
+    def get(self, app_model: App, end_user: EndUser):
         app_mode = AppMode.value_of(app_model.mode)
         if app_mode not in {AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT}:
             raise NotChatAppError()
@@ -133,7 +133,7 @@ class MessageFeedbackApi(WebApiResource):
         }
     )
     @web_ns.response(200, "Feedback submitted successfully", web_ns.models[ResultResponse.__name__])
-    def post(self, app_model, end_user, message_id: UUID):
+    def post(self, app_model: App, end_user: EndUser, message_id: UUID):
         message_id_str = str(message_id)
 
         payload = MessageFeedbackPayload.model_validate(web_ns.payload or {})
@@ -167,7 +167,7 @@ class MessageMoreLikeThisApi(WebApiResource):
             500: "Internal Server Error",
         }
     )
-    def get(self, app_model, end_user, message_id: UUID):
+    def get(self, app_model: App, end_user: EndUser, message_id: UUID):
         if app_model.mode != "completion":
             raise NotCompletionAppError()
 
@@ -223,7 +223,7 @@ class MessageSuggestedQuestionApi(WebApiResource):
             500: "Internal Server Error",
         }
     )
-    def get(self, app_model, end_user, message_id: UUID):
+    def get(self, app_model: App, end_user: EndUser, message_id: UUID):
         app_mode = AppMode.value_of(app_model.mode)
         if app_mode not in {AppMode.CHAT, AppMode.AGENT_CHAT, AppMode.ADVANCED_CHAT}:
             raise NotChatAppError()

api/controllers/web/remote_files.py

@@ -13,6 +13,7 @@ from core.helper import ssrf_proxy
 from extensions.ext_database import db
 from fields.file_fields import FileWithSignedUrl, RemoteFileInfo
 from graphon.file import helpers as file_helpers
+from models.model import App, EndUser
 from services.file_service import FileService
 
 from ..common.schema import register_response_schema_models, register_schema_models
@@ -41,7 +42,7 @@ class RemoteFileInfoApi(WebApiResource):
         }
     )
     @web_ns.response(200, "Remote file info", web_ns.models[RemoteFileInfo.__name__])
-    def get(self, app_model, end_user, url):
+    def get(self, app_model: App, end_user: EndUser, url: str):
         """Get information about a remote file.
 
         Retrieves basic information about a file located at a remote URL,
@@ -85,7 +86,7 @@ class RemoteFileUploadApi(WebApiResource):
         }
     )
     @web_ns.response(201, "Remote file uploaded", web_ns.models[FileWithSignedUrl.__name__])
-    def post(self, app_model, end_user):
+    def post(self, app_model: App, end_user: EndUser):
         """Upload a file from a remote URL.
 
         Downloads a file from the provided remote URL and uploads it

api/controllers/web/saved_message.py

@@ -11,6 +11,7 @@ from controllers.web.error import NotCompletionAppError
 from controllers.web.wraps import WebApiResource
 from fields.conversation_fields import ResultResponse
 from fields.message_fields import SavedMessageInfiniteScrollPagination, SavedMessageItem
+from models.model import App, EndUser
 from services.errors.message import MessageNotExistsError
 from services.saved_message_service import SavedMessageService
 
@@ -43,7 +44,7 @@ class SavedMessageListApi(WebApiResource):
             500: "Internal Server Error",
         }
     )
-    def get(self, app_model, end_user):
+    def get(self, app_model: App, end_user: EndUser):
         if app_model.mode != "completion":
             raise NotCompletionAppError()
 
@@ -77,7 +78,7 @@ class SavedMessageListApi(WebApiResource):
         }
     )
     @web_ns.response(200, "Message saved successfully", web_ns.models[ResultResponse.__name__])
-    def post(self, app_model, end_user):
+    def post(self, app_model: App, end_user: EndUser):
         if app_model.mode != "completion":
             raise NotCompletionAppError()
 
@@ -106,7 +107,7 @@ class SavedMessageApi(WebApiResource):
             500: "Internal Server Error",
         }
     )
-    def delete(self, app_model, end_user, message_id: UUID):
+    def delete(self, app_model: App, end_user: EndUser, message_id: UUID):
         message_id_str = str(message_id)
 
         if app_model.mode != "completion":

api/controllers/web/site.py

@@ -10,7 +10,7 @@ from controllers.web.wraps import WebApiResource
 from extensions.ext_database import db
 from libs.helper import AppIconUrlField
 from models.account import TenantStatus
-from models.model import App, Site
+from models.model import App, EndUser, Site
 from services.feature_service import FeatureService
 
 
@@ -70,7 +70,7 @@ class AppSiteApi(WebApiResource):
         }
     )
     @marshal_with(app_fields)
-    def get(self, app_model, end_user):
+    def get(self, app_model: App, end_user: EndUser):
         """Retrieve app site info."""
         # get site
         site = db.session.scalar(select(Site).where(Site.app_id == app_model.id).limit(1))
@@ -78,10 +78,10 @@ class AppSiteApi(WebApiResource):
         if not site:
             raise Forbidden()
 
-        if app_model.tenant.status == TenantStatus.ARCHIVE:
+        if app_model.tenant and app_model.tenant.status == TenantStatus.ARCHIVE:
             raise Forbidden()
 
-        can_replace_logo = FeatureService.get_features(app_model.tenant_id).can_replace_logo
+        can_replace_logo = FeatureService.get_features(app_model.tenant_id, exclude_vector_space=True).can_replace_logo
 
         return AppSiteInfo(app_model.tenant, app_model, site, end_user.id, can_replace_logo)
 
@@ -119,6 +119,6 @@ def serialize_site(site: Site) -> dict[str, Any]:
 
 
 def serialize_app_site_payload(app_model: App, site: Site, end_user_id: str | None) -> dict[str, Any]:
-    can_replace_logo = FeatureService.get_features(app_model.tenant_id).can_replace_logo
+    can_replace_logo = FeatureService.get_features(app_model.tenant_id, exclude_vector_space=True).can_replace_logo
     app_site_info = AppSiteInfo(app_model.tenant, app_model, site, end_user_id, can_replace_logo)
     return cast(dict[str, Any], marshal(app_site_info, AppSiteApi.app_fields))

api/core/app/apps/advanced_chat/app_runner.py

@@ -27,6 +27,7 @@ from core.moderation.base import ModerationError
 from core.moderation.input_moderation import InputModeration
 from core.repositories.factory import WorkflowExecutionRepository, WorkflowNodeExecutionRepository
 from core.workflow.node_factory import get_default_root_node_id
+from core.workflow.nodes.agent_v2.session_cleanup_layer import build_workflow_agent_session_cleanup_layer
 from core.workflow.system_variables import (
     build_bootstrap_variables,
     build_system_variables,
@@ -239,6 +240,7 @@ class AdvancedChatAppRunner(WorkflowBasedAppRunner):
         )
 
         workflow_entry.graph_engine.layer(persistence_layer)
+        workflow_entry.graph_engine.layer(build_workflow_agent_session_cleanup_layer())
         conversation_variable_layer = ConversationVariablePersistenceLayer(
             ConversationVariableUpdater(session_factory.get_session_maker())
         )

api/core/app/apps/base_app_runner.py

@@ -7,6 +7,7 @@ from typing import TYPE_CHECKING, Any, Union
 
 from core.app.app_config.entities import ExternalDataVariableEntity, PromptTemplateEntity
 from core.app.apps.base_app_queue_manager import AppQueueManager, PublishFrom
+from core.app.apps.exc import GenerateTaskStoppedError
 from core.app.entities.app_invoke_entities import (
     AppGenerateEntity,
     EasyUIBasedAppGenerateEntity,
@@ -292,46 +293,51 @@ class AppRunner:
         prompt_messages: list[PromptMessage] = []
         text = ""
         usage = None
-        for result in invoke_result:
-            if not agent:
-                queue_manager.publish(QueueLLMChunkEvent(chunk=result), PublishFrom.APPLICATION_MANAGER)
-            else:
-                queue_manager.publish(QueueAgentMessageEvent(chunk=result), PublishFrom.APPLICATION_MANAGER)
+        try:
+            for result in invoke_result:
+                if not agent:
+                    queue_manager.publish(QueueLLMChunkEvent(chunk=result), PublishFrom.APPLICATION_MANAGER)
+                else:
+                    queue_manager.publish(QueueAgentMessageEvent(chunk=result), PublishFrom.APPLICATION_MANAGER)
 
-            message = result.delta.message
-            if isinstance(message.content, str):
-                text += message.content
-            elif isinstance(message.content, list):
-                for content in message.content:
-                    if isinstance(content, str):
-                        text += content
-                    elif isinstance(content, TextPromptMessageContent):
-                        text += content.data
-                    elif isinstance(content, ImagePromptMessageContent):
-                        if message_id and user_id and tenant_id:
-                            try:
-                                self._handle_multimodal_image_content(
-                                    content=content,
-                                    message_id=message_id,
-                                    user_id=user_id,
-                                    tenant_id=tenant_id,
-                                    queue_manager=queue_manager,
-                                )
-                            except Exception:
-                                _logger.exception("Failed to handle multimodal image output")
+                message = result.delta.message
+                if isinstance(message.content, str):
+                    text += message.content
+                elif isinstance(message.content, list):
+                    for content in message.content:
+                        if isinstance(content, str):
+                            text += content
+                        elif isinstance(content, TextPromptMessageContent):
+                            text += content.data
+                        elif isinstance(content, ImagePromptMessageContent):
+                            if message_id and user_id and tenant_id:
+                                try:
+                                    self._handle_multimodal_image_content(
+                                        content=content,
+                                        message_id=message_id,
+                                        user_id=user_id,
+                                        tenant_id=tenant_id,
+                                        queue_manager=queue_manager,
+                                    )
+                                except Exception:
+                                    _logger.exception("Failed to handle multimodal image output")
+                            else:
+                                _logger.warning("Received multimodal output but missing required parameters")
                         else:
-                            _logger.warning("Received multimodal output but missing required parameters")
-                    else:
-                        text += content.data if hasattr(content, "data") else str(content)
+                            text += content.data if hasattr(content, "data") else str(content)
 
-            if not model:
-                model = result.model
+                if not model:
+                    model = result.model
 
-            if not prompt_messages:
-                prompt_messages = list(result.prompt_messages)
+                if not prompt_messages:
+                    prompt_messages = list(result.prompt_messages)
 
-            if result.delta.usage:
-                usage = result.delta.usage
+                if result.delta.usage:
+                    usage = result.delta.usage
+        except GenerateTaskStoppedError:
+            # Explicitly close provider stream to stop in-flight token generation ASAP.
+            invoke_result.close()
+            raise
 
         if usage is None:
             usage = LLMUsage.empty_usage()

api/core/app/apps/common/workflow_response_converter.py

@@ -562,15 +562,16 @@ class WorkflowResponseConverter:
         outputs, outputs_truncated = self._truncate_mapping(encoded_outputs)
         metadata = self._merge_metadata(event.execution_metadata, snapshot)
 
-        if isinstance(event, QueueNodeSucceededEvent):
-            status = WorkflowNodeExecutionStatus.SUCCEEDED
-            error_message = event.error
-        elif isinstance(event, QueueNodeFailedEvent):
-            status = WorkflowNodeExecutionStatus.FAILED
-            error_message = event.error
-        else:
-            status = WorkflowNodeExecutionStatus.EXCEPTION
-            error_message = event.error
+        match event:
+            case QueueNodeSucceededEvent():
+                status = WorkflowNodeExecutionStatus.SUCCEEDED
+                error_message = event.error
+            case QueueNodeFailedEvent():
+                status = WorkflowNodeExecutionStatus.FAILED
+                error_message = event.error
+            case _:
+                status = WorkflowNodeExecutionStatus.EXCEPTION
+                error_message = event.error
 
         return NodeFinishStreamResponse(
             task_id=task_id,

api/core/app/apps/workflow/app_runner.py

@@ -10,6 +10,7 @@ from core.app.entities.app_invoke_entities import InvokeFrom, WorkflowAppGenerat
 from core.app.workflow.layers.persistence import PersistenceWorkflowInfo, WorkflowPersistenceLayer
 from core.repositories.factory import WorkflowExecutionRepository, WorkflowNodeExecutionRepository
 from core.workflow.node_factory import get_default_root_node_id
+from core.workflow.nodes.agent_v2.session_cleanup_layer import build_workflow_agent_session_cleanup_layer
 from core.workflow.system_variables import build_bootstrap_variables, build_system_variables
 from core.workflow.variable_pool_initializer import add_node_inputs_to_pool, add_variables_to_pool
 from core.workflow.workflow_entry import WorkflowEntry
@@ -166,6 +167,7 @@ class WorkflowAppRunner(WorkflowBasedAppRunner):
         )
 
         workflow_entry.graph_engine.layer(persistence_layer)
+        workflow_entry.graph_engine.layer(build_workflow_agent_session_cleanup_layer())
         for layer in self._graph_engine_layers:
             workflow_entry.graph_engine.layer(layer)
 

api/core/app/workflow/layers/persistence.py

@@ -47,6 +47,12 @@ from graphon.graph_events import (
 )
 from graphon.node_events import NodeRunResult
 from libs.datetime_utils import naive_utc_now
+from services.workflow.inspector_events import (
+    publish_node_changed as _inspector_publish_node_changed,
+)
+from services.workflow.inspector_events import (
+    publish_workflow_completed as _inspector_publish_workflow_completed,
+)
 
 
 @dataclass(slots=True)
@@ -163,6 +169,7 @@ class WorkflowPersistenceLayer(GraphEngineLayer):
 
         self._workflow_execution_repository.save(execution)
         self._enqueue_trace_task(execution)
+        _inspector_publish_workflow_completed(workflow_run_id=execution.id_, status=str(execution.status.value))
 
     def _handle_graph_run_partial_succeeded(self, event: GraphRunPartialSucceededEvent) -> None:
         execution = self._get_workflow_execution()
@@ -173,6 +180,7 @@ class WorkflowPersistenceLayer(GraphEngineLayer):
 
         self._workflow_execution_repository.save(execution)
         self._enqueue_trace_task(execution)
+        _inspector_publish_workflow_completed(workflow_run_id=execution.id_, status=str(execution.status.value))
 
     def _handle_graph_run_failed(self, event: GraphRunFailedEvent) -> None:
         execution = self._get_workflow_execution()
@@ -184,6 +192,7 @@ class WorkflowPersistenceLayer(GraphEngineLayer):
         self._fail_running_node_executions(error_message=event.error)
         self._workflow_execution_repository.save(execution)
         self._enqueue_trace_task(execution)
+        _inspector_publish_workflow_completed(workflow_run_id=execution.id_, status=str(execution.status.value))
 
     def _handle_graph_run_aborted(self, event: GraphRunAbortedEvent) -> None:
         execution = self._get_workflow_execution()
@@ -194,6 +203,7 @@ class WorkflowPersistenceLayer(GraphEngineLayer):
         self._fail_running_node_executions(error_message=execution.error_message or "")
         self._workflow_execution_repository.save(execution)
         self._enqueue_trace_task(execution)
+        _inspector_publish_workflow_completed(workflow_run_id=execution.id_, status=str(execution.status.value))
 
     def _handle_graph_run_paused(self, event: GraphRunPausedEvent) -> None:
         execution = self._get_workflow_execution()
@@ -241,6 +251,7 @@ class WorkflowPersistenceLayer(GraphEngineLayer):
             created_at=event.start_at,
         )
         self._node_snapshots[event.id] = snapshot
+        _inspector_publish_node_changed(workflow_run_id=execution.id_, node_id=event.node_id, status="running")
 
     def _handle_node_retry(self, event: NodeRunRetryEvent) -> None:
         domain_execution = self._get_node_execution(event.id)
@@ -248,6 +259,11 @@ class WorkflowPersistenceLayer(GraphEngineLayer):
         domain_execution.error = event.error
         self._workflow_node_execution_repository.save(domain_execution)
         self._workflow_node_execution_repository.save_execution_data(domain_execution)
+        _inspector_publish_node_changed(
+            workflow_run_id=self._get_workflow_execution().id_,
+            node_id=domain_execution.node_id,
+            status="retry",
+        )
 
     def _handle_node_succeeded(self, event: NodeRunSucceededEvent) -> None:
         domain_execution = self._get_node_execution(event.id)
@@ -257,6 +273,11 @@ class WorkflowPersistenceLayer(GraphEngineLayer):
             WorkflowNodeExecutionStatus.SUCCEEDED,
             finished_at=event.finished_at,
         )
+        _inspector_publish_node_changed(
+            workflow_run_id=self._get_workflow_execution().id_,
+            node_id=domain_execution.node_id,
+            status="succeeded",
+        )
 
     def _handle_node_failed(self, event: NodeRunFailedEvent) -> None:
         domain_execution = self._get_node_execution(event.id)
@@ -267,6 +288,11 @@ class WorkflowPersistenceLayer(GraphEngineLayer):
             error=event.error,
             finished_at=event.finished_at,
         )
+        _inspector_publish_node_changed(
+            workflow_run_id=self._get_workflow_execution().id_,
+            node_id=domain_execution.node_id,
+            status="failed",
+        )
 
     def _handle_node_exception(self, event: NodeRunExceptionEvent) -> None:
         domain_execution = self._get_node_execution(event.id)
@@ -277,6 +303,11 @@ class WorkflowPersistenceLayer(GraphEngineLayer):
             error=event.error,
             finished_at=event.finished_at,
         )
+        _inspector_publish_node_changed(
+            workflow_run_id=self._get_workflow_execution().id_,
+            node_id=domain_execution.node_id,
+            status="exception",
+        )
 
     def _handle_node_pause_requested(self, event: NodeRunPauseRequestedEvent) -> None:
         domain_execution = self._get_node_execution(event.id)

api/core/base/tts/app_generator_tts_publisher.py

@@ -91,26 +91,28 @@ class AppGeneratorTTSPublisher:
                         )
                         future_queue.put(futures_result)
                     break
-                elif isinstance(message.event, QueueAgentMessageEvent | QueueLLMChunkEvent):
-                    message_content = message.event.chunk.delta.message.content
-                    if not message_content:
-                        continue
-                    match message_content:
-                        case str():
-                            self.msg_text += message_content
-                        case list():
-                            for content in message_content:
-                                if not isinstance(content, TextPromptMessageContent):
-                                    continue
-                                self.msg_text += content.data
-                elif isinstance(message.event, QueueTextChunkEvent):
-                    self.msg_text += message.event.text
-                elif isinstance(message.event, QueueNodeSucceededEvent):
-                    if message.event.outputs is None:
-                        continue
-                    output = message.event.outputs.get("output", "")
-                    if isinstance(output, str):
-                        self.msg_text += output
+                else:
+                    match message.event:
+                        case QueueAgentMessageEvent() | QueueLLMChunkEvent():
+                            message_content = message.event.chunk.delta.message.content
+                            if not message_content:
+                                continue
+                            match message_content:
+                                case str():
+                                    self.msg_text += message_content
+                                case list():
+                                    for content in message_content:
+                                        if not isinstance(content, TextPromptMessageContent):
+                                            continue
+                                        self.msg_text += content.data
+                        case QueueTextChunkEvent():
+                            self.msg_text += message.event.text
+                        case QueueNodeSucceededEvent():
+                            if message.event.outputs is None:
+                                continue
+                            output = message.event.outputs.get("output", "")
+                            if isinstance(output, str):
+                                self.msg_text += output
                 self.last_message = message
                 sentence_arr, text_tmp = self._extract_sentence(self.msg_text)
                 if len(sentence_arr) >= min(self.max_sentence, 7):

api/core/provider_manager.py

@@ -534,7 +534,9 @@ class ProviderManager:
         cache_key = f"tenant:{tenant_id}:model_load_balancing_enabled"
         cache_result = redis_client.get(cache_key)
         if cache_result is None:
-            model_load_balancing_enabled = FeatureService.get_features(tenant_id).model_load_balancing_enabled
+            model_load_balancing_enabled = FeatureService.get_features(
+                tenant_id, exclude_vector_space=True
+            ).model_load_balancing_enabled
             redis_client.setex(cache_key, 120, str(model_load_balancing_enabled))
         else:
             cache_result = cache_result.decode("utf-8")