InfiComponent Lab · acceptance test plan

Requirement baseline · scope · data · execution policy · entry and exit criteria

Objective and scope

Verify the supplied LAB01–LAB16 requirements against the unmodified packaged executable using InfiAutomation MCP 0.5.0 at C:/InfiAutomation. Exercise real renderer workflows, the persistent REST backend, native Electron Menu API, optional CDP attachment and explicit Windows UIA OS dialogs.

Supplementary controls are checked separately. The four interactive follow-ups remain visible in the case catalogue; neither mapping nor a passed smoke case claims exhaustive coverage.

Complete supplied requirements · MCP-generated Markdown plan · Structured plan · Executable suite

Environment, entry criteria and test data

Windows win32-x64; MCP 0.5.0 on Node v24.22.2; Electron 45.2.1, Chromium 156.0.8078.75, Playwright 1.64.1. Packaged application version 1.0.0.

Entry criteria: complete win-unpacked folder; executable readable; MCP dist/cli.js and Node 24+ available; native UIA worker and .NET 8 Desktop Runtime available; interactive unlocked Windows desktop; available loopback ports.

Synthetic credentials: tester@example.test / LabPass!2026. Names, emails, notes, tasks and files in this run are synthetic. Seed: 18 customers, 8 orders, 3 tasks. Profile: C:/InfiComponentLab\.runtime\acceptance-execution-20261104-acceptance-rerun-1793786183771. The no-test-mode target uses its own profile and loopback port.

Execution sequence and isolation

  1. Call MCP doctor, launch owned packaged Electron and inspect session_info, ui_windows, ui_menus and all 21 page snapshots.
  2. Import and read the complete requirements, save requirement-linked plan, generate schemaVersion 2 suite with healing off.
  3. Run LAB01–LAB16 sequentially. Assert positive, negative, boundary, state-change and persistence outcomes.
  4. Run supplementary controls and persistence across owned launches. Run reset only after persistent-state cases finish.
  5. Launch a separate profile without --test-mode and verify disabled UI reset plus HTTP 403.
  6. Run optional CDP and explicit Windows UIA checks in a separate newly created profile; record connection capability boundaries.
  7. Poll test_status to completion, retrieve report_get and traceability_get, retain first-run failures and fresh rerun evidence.

Cases begin with a ready Workspace assertion before native IPC navigation. Navigation waits for the destination page. The suite never edits the application or enables healing, and no visual baselines are generated or approved.

Coverage and acceptance gates

The core gate requires all 16 LAB cases passing explicit assertions with zero healed outcomes. Native export additionally requires an existing parseable JSON file with 18 seeded customers. The connection gate requires runtime/menu evidence, a native API capability rejection over CDP, preserved state after detach, and observed UIA OS dialog controls.

Safety gates: only fresh synthetic profiles; owned backend/CDP configuration at loopback; enabled context isolation, renderer sandbox and web security; test-mode reset confirmation; reset denial outside test mode; no automatic baseline approval. Runtime listener socket inspection and exploit testing are outside the recorded configuration audit.

Interactive follow-ups must be completed before claiming exhaustive supplementary coverage. The report can show core acceptance passed while these remain not run.

Requirement-linked scenarios

CaseRequirementScopeExpected result
LAB01LAB01LAB01 — Packaged Workspace has seven groups with three working subpages eachExpect page-title = A workspace built to be tested.; Unique Workspace title; Expect 7 elements: .nav-group; Expect 3 elements: [data-testid^="nav-workspace-"]; Expect page-title = A workspace built to be tested.; Expect page-title = A workspace built to be tested.; Expect page-title = A workspace built to be tested.; Expect page-title = Activity log; Expect page-title = A workspace built to be tested.; Expect page-title = Sign-in sandbox; Expect 3 elements: [data-testid^="nav-forms-"]; Expect page-title = Registration form; Expect page-title = Registration form; Expect page-title = Registration form; Expect page-title = Input laboratory; Expect page-title = Registration form; Expect page-title = Validation & editor; Expect 3 elements: [data-testid^="nav-records-"]; Expect page-title = Customers; Expect page-title = Customers; Expect page-title = Customers; Expect page-title = Orders; Expect page-title = Customers; Expect page-title = Advanced data grid; Expect 3 elements: [data-testid^="nav-components-"]; Expect page-title = Buttons & menus; Expect page-title = Buttons & menus; Expect page-title = Buttons & menus; Expect page-title = Dialogs & overlays; Expect page-title = Buttons & menus; Expect page-title = Tabs, tree & accordion; Expect 3 elements: [data-testid^="nav-files-"]; Expect page-title = Upload center; Expect page-title = Upload center; Expect page-title = Upload center; Expect page-title = Downloads & exports; Expect page-title = Upload center; Expect page-title = Frames & media; Expect 3 elements: [data-testid^="nav-workflows-"]; Expect page-title = Task board; Expect page-title = Task board; Expect page-title = Task board; Expect page-title = Order wizard; Expect page-title = Task board; Expect page-title = Async & recovery; Expect 3 elements: [data-testid^="nav-settings-"]; Expect page-title = Preferences; Expect page-title = Preferences; Expect page-title = Preferences; Expect page-title = Windows & connections; Expect page-title = Preferences; Expect page-title = Runtime & test data; Health identifies synthetic test-mode app
LAB02LAB02LAB02 — Registration rejects missing name/email/terms and records a backend submissionExpect page-title = A workspace built to be tested.; Expect page-title = Registration form; Expect page-title = Registration form; Expect page-result = required fields; Expect page-result = required fields; Expect page-result = required fields; Expect page-result = required fields; Backend independently rejects missing terms; Expect page-result = Registration saved for Acceptance Tester; Submission produced a persisted form event
LAB03LAB03LAB03 — Dates, multiple selection, keyboard slider and framework expose submitted valuesExpect page-title = A workspace built to be tested.; Expect page-title = Registration form; Expect page-title = Input laboratory; Expect Confidence value; Expect Confidence value; Expect 2 elements: select[name="skills"] option:checked; Expect Test framework value; Expect input-values = "quantity": "9"; Expect input-values = "date": "2026-11-15"; Expect input-values = "time": "14:45"; Expect input-values = "datetime": "2026-11-15T14:45"; Expect input-values = "month": "2026-12"; Expect input-values = "week": "2026-W47"; Expect input-values = "confidence": "100"; Expect input-values = "Playwright"; Expect input-values = "Electron"; Expect input-values = "framework": "InfiAutomation"
LAB04LAB04LAB04 — Project format, team boundaries, password length and confirmation are enforcedExpect page-title = A workspace built to be tested.; Expect page-title = Registration form; Expect page-title = Validation & editor; Expect page-result = Check project code; Expect page-result = Check project code; Expect page-result = Check project code; Expect page-result = Check project code; Expect page-result = Check project code; Expect page-result = Check project code; Expect page-result = Passphrases do not match.; Expect page-result = Project ABC-123 is valid.; Expect page-result = Project ABC-123 is valid.
LAB05LAB05LAB05 — Customers create, edit, search, persist and delete across UI and RESTExpect page-title = A workspace built to be tested.; Expect page-title = Customers; Expect page-title = Customers; Expect 1 elements: [data-testid="customer-row"]; Created customer exists in REST; Expect page-title = A workspace built to be tested.; Expect page-title = A workspace built to be tested.; Expect page-title = Customers; Expect page-title = Customers; Expect 1 elements: [data-testid="customer-row"]; Edited customer persisted; Expect 1 elements: [data-testid="customer-row"]; Expect empty-customers = No customers match your filters.; Deleted customer is absent
LAB06LAB06LAB06 — Grid selects five of eighteen records and paginates to the final threeExpect page-title = A workspace built to be tested.; Expect page-title = Customers; Expect page-title = Advanced data grid; Expect .pagination text; Expect 5 elements: #grid-table tbody tr; Expect #selection-count text; Expect .pagination text; Expect #selection-count text; Expect .pagination text; Expect 3 elements: #grid-table tbody tr; Expected enabled=false: Next page; Expect .pagination text
LAB07LAB07LAB07 — Native JSON export reports a real path and Compact changes renderer stateExpect page-title = A workspace built to be tested.; Expect command-status = C:/InfiComponentLab\.runtime\acceptance-execution-20261104-acceptance-rerun-1793786183771\lab-data\infi-lab-export.json; Expect command-status = Compact mode on; Expect 1 elements: body.compact; Expect command-status = Compact mode off; Expect 0 elements: body.compact
LAB08LAB08LAB08 — DOM File action, double click and disabled button preserve explicit behaviorExpect page-title = A workspace built to be tested.; Expect page-title = Buttons & menus; Expect page-title = Buttons & menus; Expect page-result = DOM note saved; Expect page-result = Double-click detected; Disabled remains disabled
LAB09LAB09LAB09 — Note dialog cancels without saving and saves only the submitted noteExpect page-title = A workspace built to be tested.; Expect page-title = Buttons & menus; Expect page-title = Dialogs & overlays; Expect 0 elements: #app-dialog[open]; Expect 0 elements: [data-testid="page-result"]:not([hidden]); Expect 0 elements: #app-dialog[open]; Expect page-result = Note saved: Acceptance saved note; Expect 0 elements: #app-dialog[open]; Expect page-result = Note saved: Acceptance saved note
LAB10LAB10LAB10 — Tabs change visible panels and expanded tree selects ReportsExpect page-title = A workspace built to be tested.; Expect page-title = Buttons & menus; Expect page-title = Tabs, tree & accordion; Expect 1 elements: #tab-panel-1:not([hidden]); Expect 1 elements: #tab-panel-0[hidden]; Expect 1 elements: #tab-panel-2:not([hidden]); Expect page-result = Tree selected: Reports; Expect 1 elements: [role="treeitem"][aria-selected="true"][data-tree-leaf="Reports"]
LAB11LAB11LAB11 — Same-origin iframe and open shadow forms retain independent feedbackExpect page-title = A workspace built to be tested.; Expect page-title = Upload center; Expect page-title = Frames & media; Frame saved; Expect shadow-output = Shadow saved: Acceptance shadow; Frame remains independent
LAB12LAB12LAB12 — Invalid credentials reject; login, session read, logout and post-logout denial workExpect page-title = A workspace built to be tested.; Expect page-title = A workspace built to be tested.; Expect page-title = Sign-in sandbox; Expect page-result = incorrect; Expect session-user = Signed in as Alex Morgan; Expect page-result = Session is active; Expect session-user = Signed out; Expect page-result = Sign in to continue
LAB13LAB13LAB13 — Wizard validates all three stages and persists the resulting orderExpect page-title = A workspace built to be tested.; Expect page-title = Task board; Expect page-title = Order wizard; Expect 1 elements: #wizard-form select[name="customer"]; Expect 1 elements: #wizard-form input[name="amount"]; Expect order-review = Customer: Aarav Shah; Expect 1 elements: #wizard-form input[name="confirmed"]:not(:checked); No order without confirmation; Expect order-success = created for Aarav Shah; Order exists in persisted backend; Expect page-title = Orders; Expect #order-table tbody tr:first-child text; Expect #order-table tbody tr:first-child text
LAB14LAB14LAB14 — HTTP 503 error is reported, Success recovers and delayed content becomes readyExpect page-title = A workspace built to be tested.; Expect page-title = Task board; Expect page-title = Async & recovery; Intentional error is actual HTTP 503; Expect request-response = Intentional lab error; Expect request-response = Request completed successfully; Expect delayed-result = Delayed content is ready
LAB15LAB15LAB15 — Uniquely titled Inspector stores its own note and returns to WorkspaceExpect page-title = A workspace built to be tested.; Inspector unique title; Expect inspector-result = Inspector saved: Acceptance independent note; Workspace title; Expect 0 elements: [data-testid="inspector-result"]; Expect inspector-result = Inspector saved: Acceptance independent note
LAB16LAB16LAB16 — API-created synthetic customer is exactly one matching renderer rowExpect page-title = A workspace built to be tested.; Start owned app before REST; Create REST customer; Expect page-title = Customers; Expect page-title = Customers; Expect 1 elements: [data-testid="customer-row"]; Expect [data-testid="customer-row"] text; Expect [data-testid="customer-row"] text
SUP01SUP01SUP01 — Upload/download API returns identical bytes and renderer lists the real fileExpect page-title = A workspace built to be tested.; Start app; Upload synthetic bytes; Download identical bytes; Expect page-title = Upload center; Expect page-title = Upload center; Expect #uploaded-files text; Expect 1 elements: #uploaded-files a[download="acceptance.txt"]
SUP02SUP02SUP02 — Task keyboard alternative creates and moves a persistent taskExpect page-title = A workspace built to be tested.; Expect page-title = Task board; Expect page-title = Task board; Expect page-result = Task moved to Done; Expect Move Acceptance keyboard task value; Task move persisted
SUP03SUP03SUP03 — Tooltip, popover, progress, meter and accessible canvas have observable resultsExpect page-title = A workspace built to be tested.; Expect page-title = Buttons & menus; Expect page-title = Buttons & menus; Tooltip visible; Expect page-title = Buttons & menus; Expect page-title = Dialogs & overlays; Expect 1 elements: #details-popover:popover-open; Expect 0 elements: #details-popover:popover-open; Expect page-title = Buttons & menus; Expect page-title = Tabs, tree & accordion; Expect 1 elements: #demo-progress[value="50"]; Expect 1 elements: meter[value="62"]; Expect page-title = Upload center; Expect page-title = Frames & media; Expect #canvas-result text; Local audio asset exists
SUP04SUP04SUP04 — Rich text value appears in valid project resultExpect page-title = A workspace built to be tested.; Expect page-title = Registration form; Expect page-title = Validation & editor; Expect page-result = Release notes: Acceptance rich text note
SUP05SUP05SUP05 — Preferences save real theme, compact mode, locale and notification valuesExpect page-title = A workspace built to be tested.; Expect page-title = Preferences; Expect page-title = Preferences; Expect page-result = Preferences saved; Expect 1 elements: body.dark.compact; Preferences persisted in backend
SUP06SUP05, LAB05, LAB13, LAB16SUP06 — A new owned launch reads persisted preferences, customer and orderExpect page-title = A workspace built to be tested.; Expect page-title = Preferences; Expect page-title = Preferences; Expect Theme value; Expect Locale value; Compact preference persists; Expect 1 elements: body.dark.compact; Order persisted across launch; Expect page-title = Customers; Expect page-title = Customers; Expect 1 elements: [data-testid="customer-row"]
SUP07SUP06SUP07 — Command search and activity filtering expose specific resultsExpect page-title = A workspace built to be tested.; Expect page-title = Registration form; Expect page-title = A workspace built to be tested.; Expect page-title = Activity log; Expect 1 elements: #event-list .activity-item; Expect #event-list text; Expect 0 elements: #event-list .activity-item
SUP08SUP07SUP08 — Slow request can cancel and then recover successfullyExpect page-title = A workspace built to be tested.; Expect page-title = Task board; Expect page-title = Async & recovery; Expect request-response = Request cancelled; Expected enabled=true: Send request; Expected enabled=false: Cancel request; Expect request-response = Request completed successfully
SAFE01SAFE01SAFE01 — Test-mode reset requires confirmation, cancellation preserves data, confirmed reset restores seedExpect page-title = A workspace built to be tested.; Expect page-title = Preferences; Expect page-title = Runtime & test data; Cancelled reset preserved synthetic customer; Reset modal closed; Confirmed reset restores all seeded state
SAFE02SAFE02SAFE02 — Without --test-mode, UI/native reset are unavailable and backend reset returns 403Expect page-title = A workspace built to be tested.; Expect page-title = Preferences; Expect page-title = Runtime & test data; Expected enabled=false: Reset test data; Backend rejects reset outside test mode
CONN01CONN01Owned launch reports real runtime and native menu discoveryRead the actual owned session_info and ui_menus results. Assert electron-launch, ownership owned, nativeElectronMenus true, seven native menu roots and actual Electron/Chromium versions.
CONN02CONN02CDP reports attached renderer-only capabilities and rejects native Menu APISpawn a new test instance with a fresh data directory and --remote-debugging-port=0. Attach to the observed loopback WebSocket endpoint through MCP. Read session_info and ui_windows. Assert attached ownership and false mainProcess/nativeElectronMenus. Request an explicit native menu action and assert MCP returns a capability error.
CONN03CONN03CDP detach preserves app and current entered value on reconnectEnter a synthetic registration name through CDP. Close only the attached MCP session. Assert the application process remains alive. Reconnect to the same endpoint and assert the entered name remains unchanged.
NATIVE01CONN04Explicit Windows UIA cancels an observed real OS message dialogOpen the actual native message dialog from the renderer. Explicitly attach a kind: desktop MCP session to the owned process PID. Read the real Windows UIA snapshot and capture a screenshot. Click the uniquely observed Cancel button and detach UIA. Assert the renderer reports Native action cancelled.
NATIVE02CONN04Explicit Windows UIA cancels an observed real OS file pickerOpen the actual OS file picker. Explicitly attach Windows UIA to the owned PID. Read observed native controls and capture a screenshot. Click the observed Cancel button. Assert the renderer reports Native file selection cancelled.
SAFE03SAFE03Packaged configuration retains isolation, sandbox, web security and loopback bindingRead the packaged main-process and backend source, rather than development-source assumptions. Assert contextIsolation true, nodeIntegration false, sandbox true and webSecurity true. Assert backend listener and optional CDP configuration bind to 127.0.0.1. Record observed loopback renderer/CDP URLs. This is a configuration audit; it is not exploit testing.
NATIVE03SUP01, CONN04HTML file chooser uses observed UIA controls; upload persists and downloads identical bytesCreate uia-upload-fixture.txt containing only synthetic text inside the report folder. Open the HTML upload chooser and explicitly attach Windows UIA. Observe the File name editor and Open split button. Fill the fixture path and invoke the observed Open control. Assert the renderer shows the fixture filename before submitting the upload. Click Upload selected file and assert the saved filename. Use MCP api_request to read persisted upload metadata and download the file, comparing exact bytes.

Interactive follow-ups and limits

MAN01 — CSV download and selected-row CSV content not-run

  1. Select five current-page rows in the advanced grid. Export selected and save inside a dedicated test output directory.
  2. Parse selected-customers.csv and assert exactly five data records, the six expected columns, matching selected customer IDs and valid escaping.
  3. Download all-customer CSV and verify its records against the current REST customer state.

Expected: Downloaded bytes parse correctly and correspond to the selected/all current records.

The MCP 0.5.0 typed actions do not expose a browser download-event capture/save-path action. Native JSON export bytes were independently verified.

MAN02 — Physical task drag-and-drop gesture not-run

  1. Create a synthetic task in To do.
  2. Physically drag its card to Done using observed UI coordinates.
  3. Read REST state and assert the same task ID now has lane Done.
  4. Reload the board and confirm the card remains in Done.

Expected: The gesture updates persistent state. Keyboard alternative is already covered by SUP02.

The MCP 0.5.0 typed UI action schema has no drag action. The accessible keyboard alternative passed.

MAN03 — Audible playback and canvas pixel changes not-run

  1. Play, pause and seek the local audio sample and observe actual audible output/playback time.
  2. Capture canvas pixels before and after Change canvas signal and inspect the signal rendering.

Expected: The local audio plays and controls work; the canvas visual changes consistently with its accessible signal level.

SUP03 verified the audio asset, canvas equivalent action, progress and meter. It did not assert audible output or compare pixels; no visual baseline was created.

MAN04 — Rich text bold and italic formatting not-run

  1. Enter synthetic release notes. Select a text range and apply Bold selected text and Italic selected text.
  2. Verify the selected characters visibly use the expected formatting and that other characters remain unchanged.
  3. Submit a valid project and verify the resulting release-note text.

Expected: Formatting applies to the selected range and the textual result remains correct.

SUP04 verified contenteditable input and submitted text. Character-range formatting was not exercised by the typed MCP suite.

Evidence and reproduction

Every typed-suite UI step has a screenshot; API steps retain explicit HTTP status/body evidence. The native export is copied into the report. Connection checks preserve runtime capabilities, snapshots and native screenshots. Traces, raw results, requirement hashes and run manifests remain available.

From the project directory, run the following commands in order. The runner makes new timestamped profiles on each execution; use a new report ID to preserve prior outputs.

& 'C:/InfiAutomation\.tools\node-v24.22.2-win-x64\node.exe' scripts/acceptance-mcp.cjs inspect <new-run-id>
& 'C:/InfiAutomation\.tools\node-v24.22.2-win-x64\node.exe' scripts/acceptance-mcp.cjs execute <new-run-id>
& 'C:/InfiAutomation\.tools\node-v24.22.2-win-x64\node.exe' scripts/acceptance-mcp.cjs connections <new-run-id>

The packaged EXE and app source are unchanged by this test task. Generated tests, plans, reports and synthetic profiles are the deliverables.