pi-ui-extend 1.0.24 → 1.0.30

package/external/pi-tools-suite/src/async-subagents/private-skills/browser-qa/SKILL.md

@@ -15,6 +15,13 @@ Never read, print, grep, copy, or edit credential values from
 
 ## Workflow
 
+Treat target discovery as a 30-second preflight and invoke the runner within 45
+seconds of starting. Use at most one runner invocation unless the task explicitly
+requests multiple auth profiles. If you cannot identify a reachable target and
+a supported deterministic assertion inside that preflight, return a structured
+`BLOCKED` result immediately. Do not consume the launcher budget on further
+source reading, server polling, capability probing, or retries.
+
 1. Resolve `scripts/browser-qa-runner.mjs` relative to this skill.
 2. Use the launcher-provided `$PI_SUBAGENT_AGENT_DIR/browser-qa/` workspace.
    The launcher creates its private `flows/` directory and the runner rejects
@@ -28,10 +35,13 @@ Never read, print, grep, copy, or edit credential values from
    list without creating `.pi/qa_auth.jsonc`. Otherwise skip profile discovery
    and use public mode. Choose an auth profile only when the task names its id,
    safe profile traits make the choice unambiguous, or a public run proves that
-   the requested page requires login.
+   the requested page requires login. If form authentication is required but
+   there is no usable profile, follow **Form-auth scaffolding** below instead of
+   asking the user to discover selectors.
 5. Inspect the target code and write a declarative JSONC flow under
-   `$PI_SUBAGENT_AGENT_DIR/browser-qa/flows/`. Never put credentials or
-   executable JavaScript in it.
+   `$PI_SUBAGENT_AGENT_DIR/browser-qa/flows/`. Never put credentials or raw
+   executable JavaScript in it. The `evaluate` action exposes only the safe
+   operations documented below; it does not accept expressions or scripts.
 6. Run public QA with
    `node <runner> run --base-url <url> --flow <flow.jsonc>`. The URL's exact
    origin becomes the fail-closed allowlist. Only for authenticated QA, add
@@ -39,7 +49,7 @@ Never read, print, grep, copy, or edit credential values from
    Profile id, URL, and flow path are non-secret; never pass credentials as
    arguments or environment variables.
 7. Report deterministic assertions and every artifact returned by the runner.
-   For each screenshot, video, or trace, emit a separate clickable Markdown
+   For each screenshot, video, trace, or retained download, emit a separate clickable Markdown
    link using its `uri` and also show its absolute `path`. Do this for failed
    runs too whenever `artifacts` is present; never report only `evidenceDir`.
    Visual inspection supplements assertions; it does not replace them.
@@ -56,9 +66,10 @@ Never read, print, grep, copy, or edit credential values from
   plus accessible `name`; `label`; `placeholder`; visible `text`; CSS only as a
   last resort. Use `exact: true` when similar elements could make a match
   ambiguous.
-- Let locator actions auto-wait. Use `waitFor` for an explicit UI state and use
-  `waitForTimeout` only for a short, unavoidable animation/debounce—not as a
-  substitute for an assertion. Set flow `timeoutMs` only as high as the target
+- Let locator actions auto-wait. Assertions retry until the flow timeout, so
+  prefer them over a preceding sleep. Use `waitFor` for an explicit setup state
+  and `waitForTimeout` only for short input settling or an unavoidable
+  animation/debounce. Set flow `timeoutMs` only as high as the target
   legitimately needs.
 - Place `authRejectedIf` immediately after navigation or any transition that
   may reveal expired authentication.
@@ -71,22 +82,97 @@ ambiguous failure, or deciding what evidence proves the result.
 
 ## Flow contract
 
-The flow is `{ "steps": [...] }` with at most 100 steps. Supported actions:
+The flow is `{ "steps": [...] }`, no larger than 16 MiB, with at most 100
+steps. The larger bound exists only for bounded in-memory upload payloads.
+Supported actions:
 
 - navigation: `goto`, `reload`, `waitFor`, `waitForTimeout`
 - interaction: `click`, `doubleClick`, `hover`, `fill`, `press`, `check`,
-  `uncheck`, `selectOption`
+  `uncheck`, `selectOption`, `wheel`, `evaluate`, `dragTo`, `uploadFiles`,
+  `openPopup`, `download`
 - assertions: `assertVisible`, `assertHidden`, `assertEnabled`,
   `assertDisabled`, `assertChecked`, `assertUnchecked`, `assertText`,
-  `assertValue`, `assertCount`, `assertURL`
+  `assertValue`, `assertAttribute`, `assertCount`, `assertURL`,
+  `assertDOMMetric`
 - evidence/auth: `screenshot`, `authRejectedIf`
 
 Locators accept one of `testId`, `role` (plus optional `name`), `label`,
 `placeholder`, `text`, or `css`; add `exact: true` where useful. String
 assertions require exactly one of `equals` or `includes`.
+`assertAttribute` additionally requires a bounded `attribute` name and is
+useful for `aria-*`, `data-*`, `href`, and similar observable state. All
+assertions retry until `timeoutMs` and report generic failures without exposing
+the actual text, value, or attribute content.
+
+Set an optional top-level `viewport` with integer `width` and `height` from
+`320×240` through `3840×2160`; the default is `1280×720`. The same dimensions
+are used for the browser viewport and recorded video.
+
+The top-level `environment` may set `locale`, `timezoneId`, `colorScheme`, and
+`reducedMotion`. Defaults are deterministic: `en-US`, `UTC`, `light`, and
+`reduce`. Color scheme accepts `light`, `dark`, or `no-preference`; reduced
+motion accepts `reduce` or `no-preference`. The resolved environment is returned
+in the result alongside the viewport.
+
+Triggering interactions (`click`, `doubleClick`, `press`, `check`, `uncheck`,
+and `selectOption`) may declare race-free expectations that are armed before
+the interaction:
+
+- `expectResponse`: exact origin-relative `path` (without query/fragment),
+  uppercase `method`, and integer `status` from 100 through 599;
+- `expectDialog`: `type` (`alert`, `beforeunload`, `confirm`, or `prompt`), a
+  nested `message` matcher with exactly one of `equals`/`includes`, and boolean
+  `accept`.
+
+The runner never records response bodies, headers, URLs, actual dialog text, or
+prompt defaults in observations or failure reasons. A mismatching dialog is
+dismissed so it cannot deadlock the browser.
+
+`dragTo` requires a source `locator` and `dropTarget`, with optional bounded
+`sourcePosition` and `dropPosition` `{ x, y }`. `uploadFiles` accepts only
+in-memory entries `{ name, mimeType, base64 }`; up to 10 files, 5 MiB each and
+10 MiB total. An empty array clears the input. Filesystem paths and directories
+are not supported.
+
+`download` atomically clicks its locator and requires a nested `filename`
+matcher. `maxBytes` defaults to 5 MiB and is capped at 25 MiB. Downloads are
+deleted after validation unless `retain: true` and a safe `name` are supplied;
+retained bytes appear in `artifacts.downloads` under a runner-generated `.bin`
+name. The runner cancels while its private copy grows past `maxBytes`, but this
+is an evidence-retention bound rather than a network-bandwidth guarantee because
+the browser may already hold temporary bytes. Retained bytes are scanned for
+configured authentication before publication. The actual server filename is
+never placed in diagnostics.
+
+`openPopup` atomically clicks a locator, captures a same-origin popup, and stores
+it under a safe `name` (maximum three). Target later actions with
+`{ "target": { "type": "popup", "name": "..." } }`. Target same-origin
+iframes with `{ "target": { "type": "frame", "locator": { ... } } }`; add
+`page` with a popup name for a frame inside that popup. Frame origin is checked
+from its live document before every scoped step. Cross-origin popups/frames are
+rejected. Popup recordings are returned as separate video artifacts.
+
+`wheel` accepts finite `deltaX`/`deltaY` values and requires at least one
+non-zero delta. With a locator, the runner hovers that element before sending
+the wheel input. Safe `evaluate` operations are:
+
+- `scrollTo`: optional locator plus numeric `x`/`y` or the string `"max"`;
+- `scrollBy`: optional locator plus numeric `deltaX`/`deltaY`;
+- `metrics`: optional locator plus an optional safe `name`; values are returned
+  in the runner's `observations` array.
+
+Element metrics are `scrollLeft`, `scrollTop`, `scrollWidth`, `scrollHeight`,
+`clientWidth`, `clientHeight`, `x`, `y`, `width`, and `height`. Page metrics are
+`scrollX`, `scrollY`, `scrollWidth`, `scrollHeight`, `clientWidth`,
+`clientHeight`, `viewportWidth`, and `viewportHeight`. Use `assertDOMMetric`
+with a `metric` and exactly one of `equals`, `greaterThan`,
+`greaterThanOrEqual`, `lessThan`, or `lessThanOrEqual` for a deterministic
+oracle. Raw JavaScript remains intentionally unsupported.
 
 ```jsonc
 {
+  "viewport": { "width": 844, "height": 847 },
+  "environment": { "locale": "en-GB", "timezoneId": "Europe/London", "colorScheme": "dark" },
   "steps": [
     { "action": "goto", "path": "/settings" },
     { "action": "authRejectedIf", "urlIncludes": "/login" },
@@ -94,7 +180,18 @@ assertions require exactly one of `equals` or `includes`.
       "action": "assertVisible",
       "locator": { "role": "heading", "name": "Settings", "exact": true }
     },
-    { "action": "click", "locator": { "testId": "save-settings" } },
+    { "action": "wheel", "locator": { "css": ".settings-panel" }, "deltaY": 500 },
+    {
+      "action": "assertDOMMetric",
+      "locator": { "css": ".settings-panel" },
+      "metric": "scrollTop",
+      "greaterThan": 0
+    },
+    {
+      "action": "click",
+      "locator": { "testId": "save-settings" },
+      "expectResponse": { "path": "/api/settings", "method": "PUT", "status": 200 }
+    },
     { "action": "assertText", "locator": { "testId": "toast" }, "includes": "Saved" },
     { "action": "screenshot", "name": "settings-saved" }
   ]
@@ -104,7 +201,7 @@ assertions require exactly one of `equals` or `includes`.
 For multiple profiles, invoke the runner separately. Every invocation gets an
 isolated browser context and exclusive evidence directory; the runner closes
 all owned browser resources on success and failure. Flows, screenshots, video,
-sanitized traces, and runner result manifests remain under
+sanitized traces, retained downloads, and runner result manifests remain under
 `$PI_SUBAGENT_AGENT_DIR/browser-qa/` so normal sub-agent shutdown or cleanup
 deletes them with the run directory. For form auth, recording starts on the login
 page and includes field filling and submission; password inputs remain masked,
@@ -113,11 +210,43 @@ after login succeeds so credentials are not captured in the trace.
 
 ## Credentials and blocked runs
 
+### Form-auth scaffolding
+
+When a public HTTPS login page (or loopback HTTP page for local development) is
+known and there is no usable profile, invoke the trusted runner once:
+
+```sh
+node <runner> auth scaffold \
+  --profile <safe-id> \
+  --login-url <public-login-url>
+```
+
+The runner discovers the login form and privately writes a target-specific
+`.pi/qa_auth.jsonc` containing `__PI_QA_SECRET_n__` values. It may create the
+file or replace only the runner's own empty generated template; it refuses to
+overwrite invalid or non-empty auth configuration. Never inspect the result.
+Relay `QA_AUTH_SCAFFOLD_CREATED`, the reported file, profile, placeholder count,
+and action, then ask the user to replace only those placeholder values. Browser
+QA must resume in a new run after the user confirms that edit.
+
+By default, login succeeds when the discovered form becomes hidden. A form-less
+page therefore requires an explicit deterministic signal using either
+`--success-url <glob>` or `--success-selector <selector>` with optional
+`--success-state attached|detached|visible|hidden`. Use `--base-url <url>` only
+to override the profile's same-origin application root. These are public
+metadata; never put credential values in command arguments or environment
+variables.
+
+Do not scaffold cookie, bearer, storage, storage-state, MFA, CAPTCHA, or
+federated login. For those modes, use the generic empty-template flow below and
+ask the user to configure the profile themselves.
+
 Do not request credentials merely because `.pi/qa_auth.jsonc` is absent. If the
 task explicitly requires authenticated behavior, or a public run reaches the
 flow's `authRejectedIf` check, and `profiles` returned no usable profile, run
-`node <runner> profiles --require-auth`. Only this explicit authenticated path
-may create the private empty template.
+form-auth scaffolding when supported. Otherwise run
+`node <runner> profiles --require-auth`. Only these explicit authenticated paths
+may create the private auth file.
 
 If that command or an authenticated run returns `QA_AUTH_UPDATE_REQUIRED`, stop
 and explicitly report that authenticated browser QA requires credentials or an
@@ -125,15 +254,17 @@ auth-config update. Ask the user to fill the reported file and rerun QA. If
 `templateCreated` is true, say that a private empty template was created at that
 path. Relay only the runner's profile, file, reason, action, and template-created
 state; never read the generated file or attempt to recover by exposing or
-replaying credentials.
+replaying credentials. If the action is `fill_credentials`, also relay the
+placeholder count and tell the user to replace only the generated placeholder
+values.
 
 For any other blocked run, report the runner status and redacted reason. Do not
 claim that browser QA passed based on source inspection, unit tests, or a build.
 
 After any runner invocation that actually performed browser testing, include
 all non-empty `artifacts.screenshots`, `artifacts.videos`, and
-`artifacts.traces` groups in the final response. These links are mandatory so
-the user can open the evidence directly.
+`artifacts.traces`, and `artifacts.downloads` groups in the final response.
+These links are mandatory so the user can open the evidence directly.
 
-See `references/qa-auth.example.jsonc`, `references/qa-flow.example.jsonc`, and
-`references/qa-design.md`.
+See `references/qa-auth.example.jsonc`, `references/qa-flow.example.jsonc`,
+`references/qa-design.md`, and `references/auth-scaffold-spec.md`.

package/external/pi-tools-suite/src/async-subagents/private-skills/browser-qa/references/auth-scaffold-spec.md

@@ -0,0 +1,78 @@
+# Spec: Browser QA form-auth scaffold
+
+## Type
+
+Change
+
+## Goal
+
+Let the browser-QA sub-agent prepare a target-specific form-auth configuration
+without reading, writing, receiving, or printing credential values. The user
+should only need to replace clearly marked secret placeholders.
+
+## Scope
+
+- Add a trusted runner command that inspects a public login form.
+- Generate `.pi/qa_auth.jsonc` with discovered field and submit selectors.
+- Keep generated values as non-secret placeholders until the user edits them.
+- Preserve the rule that agents never read or edit `.pi/qa_auth.jsonc`.
+
+## Non-goals
+
+- Collecting credentials interactively through the agent.
+- Guessing cookie, token, storage-state, MFA, or federated-login credentials.
+- Modifying an existing non-empty auth configuration.
+
+## Behavior
+
+- `auth scaffold` requires a project-local browser-QA sub-agent workspace, an
+  HTTPS login URL (or loopback HTTP for local development), and a safe profile
+  id.
+- The trusted runner opens the unauthenticated login page with fail-closed
+  origin routing and discovers the most likely login form, fillable fields,
+  submit control, and form container.
+- It writes a private `0600` `.pi/qa_auth.jsonc` containing only generated
+  selectors and `__PI_QA_SECRET_n__` placeholders. It may replace the runner's
+  empty generated template, but never a non-empty profiles object.
+- By default, successful login means that the discovered form becomes hidden.
+  A form-less page requires an explicit success URL pattern or selector/state.
+- A run cannot use a profile while generated placeholders remain. The runner
+  returns `QA_AUTH_UPDATE_REQUIRED` without exposing selectors or values.
+
+## Invariants
+
+- No credential value is accepted through command arguments or environment
+  variables.
+- Scaffold status output contains only path, profile id, counts, and action.
+- The command never prints inspected DOM text, input values, URLs, or selectors.
+- Agents continue to treat `.pi/qa_auth.jsonc` as opaque and user-owned.
+
+## Edge cases
+
+- Reject pages without a discoverable fillable field or submit control, and
+  form-less pages without an explicit success condition.
+- Reject remote plain-HTTP login pages.
+- Reject cross-origin base URLs and malformed success options.
+- Reject symlinked or permissive auth paths.
+- Refuse to overwrite invalid or non-empty auth configuration.
+
+## Related files
+
+- `scripts/browser-qa-runner.mjs`
+- `../SKILL.md`
+- `../../../../../test/async-subagents/browser-qa-runner.test.ts`
+
+## Verification
+
+- Unit tests for discovery output, placeholder blocking, private permissions,
+  empty-template replacement, and non-empty-config refusal.
+- Existing browser-QA runner tests and repository checks continue to pass.
+
+## Evidence
+
+- Confirmed by code: current runner owns Playwright, origin isolation, private
+  auth-file creation, validation, and redacted status output.
+- Confirmed by tests: current runner rejects permissive/symlinked auth files and
+  creates its empty template with private permissions.
+- Confirmed by docs: the private skill forbids agents from reading or editing
+  `.pi/qa_auth.jsonc`.

package/external/pi-tools-suite/src/async-subagents/private-skills/browser-qa/references/qa-design.md

@@ -47,6 +47,86 @@ observable intermediate state. Sleeping longer hides races instead of proving
 behavior. If a normal operation legitimately needs more time, adjust the flow's
 `timeoutMs` rather than inserting repeated sleeps.
 
+All assertion actions retry until that timeout. This makes an action followed
+directly by `assertText`, `assertVisible`, `assertURL`, or another assertion
+safe for asynchronously rendered outcomes. Use `assertAttribute` for
+observable state such as `aria-expanded`, `aria-invalid`, or `data-state`
+instead of reading DOM state through executable JavaScript.
+
+## Responsive and scrolling scenarios
+
+Set the flow's top-level `viewport` whenever the bug depends on a breakpoint or
+available height. Assert `viewportWidth` or `viewportHeight` with
+`assertDOMMetric` when the dimensions themselves are part of the proof; the
+runner also includes the applied viewport in its result.
+
+Use `wheel` to reproduce real pointer-wheel input. Add a locator when the wheel
+must target a nested scrolling container—the runner hovers it before sending
+the input. Because browser scrolling may be scheduled after the wheel event,
+wait only for a short known settling interval when a direct metric assertion is
+otherwise racy.
+
+Use safe `evaluate` `scrollTo`/`scrollBy` operations for deterministic setup or
+to distinguish input handling from layout behavior. Use the `metrics` operation
+to retain a named page/element snapshot in result `observations`, and use
+`assertDOMMetric` for pass/fail. Raw JavaScript expressions are intentionally
+excluded: flows remain declarative and cannot inspect authentication storage or
+execute arbitrary same-origin requests.
+
+## Deterministic browser environment
+
+Set top-level `environment` when locale, timezone, color scheme, or motion
+preferences can change the behavior. The runner otherwise uses stable defaults
+(`en-US`, `UTC`, `light`, `reduce`) instead of inheriting host settings. Prefer
+asserting product-visible copy or state derived from those settings; do not use
+screenshot pixels as the only oracle.
+
+## Causal network and dialog expectations
+
+Put `expectResponse` or `expectDialog` on the interaction that causes the event.
+The runner arms both listeners before the interaction, avoiding the race in a
+separate “click, then wait” sequence. Response expectations deliberately match
+only an exact allowlisted-origin pathname, HTTP method, and status. This proves
+that a matching request started and received a response within the action
+window without retaining its URL query, headers, or body.
+
+Dialog expectations match a fixed dialog type and exact/included message, then
+accept or dismiss it declaratively. A mismatch is dismissed before the step
+fails so the page cannot freeze. Actual event metadata is never included in
+failure diagnostics. Do not place secrets in expected messages or response
+paths even though the runner keeps diagnostics generic.
+
+## Drag, upload, and download scenarios
+
+Use `dragTo` for native DOM drag/drop and assert the resulting product state.
+Optional source/drop positions are relative bounded coordinates. Canvas-only,
+OS-native, or custom synthetic-event drag protocols remain unsupported; do not
+work around that with executable JavaScript.
+
+Uploads are memory-only base64 payloads declared in the flow. This intentionally
+prevents a flow from selecting arbitrary project files, credential config, or
+directories. Keep fixtures minimal and non-secret. An empty file list clears a
+file input.
+
+Use the atomic `download` action rather than clicking a download link directly.
+Always match the suggested filename and choose a tight `maxBytes`. Retain a
+download only when its contents are needed as evidence; otherwise the runner
+deletes it after validation. Retained downloads use generated private names,
+not server-provided paths, and are scanned for configured authentication before
+publication. `maxBytes` bounds the runner's private evidence copy and triggers
+cancellation while it grows, but it is not a network-bandwidth guarantee: the
+browser can receive temporary bytes before cancellation.
+
+## Same-origin frames and popups
+
+Use a scoped `target` for iframe or named popup interactions. The runner checks
+the live frame origin before every scoped step and checks a popup after it loads;
+both must remain in `allowedOrigins`. This supports embedded application UI and
+same-origin auxiliary windows without opening a route around the network guard.
+Cross-origin login, payment, and third-party widgets remain intentionally out of
+scope. Each popup adds a separate private video artifact, so open only the
+windows needed for the proof.
+
 ## Authentication transitions
 
 Add `authRejectedIf` directly after initial navigation and after transitions

package/external/pi-tools-suite/src/async-subagents/private-skills/browser-qa/references/qa-flow.example.jsonc

@@ -1,5 +1,12 @@
 {
   "timeoutMs": 15000,
+  "viewport": { "width": 844, "height": 847 },
+  "environment": {
+    "locale": "en-GB",
+    "timezoneId": "Europe/London",
+    "colorScheme": "dark",
+    "reducedMotion": "reduce"
+  },
   "steps": [
     { "action": "goto", "path": "/settings" },
     { "action": "authRejectedIf", "urlIncludes": "/login" },
@@ -12,8 +19,54 @@
       "locator": { "label": "Display name", "exact": true },
       "value": "QA Example"
     },
-    { "action": "click", "locator": { "testId": "save-settings" } },
+    { "action": "evaluate", "operation": "scrollTo", "locator": { "css": ".settings-panel" }, "y": 0 },
+    { "action": "wheel", "locator": { "css": ".settings-panel" }, "deltaY": 400 },
+    { "action": "assertDOMMetric", "locator": { "css": ".settings-panel" }, "metric": "scrollTop", "greaterThan": 0 },
+    { "action": "evaluate", "operation": "metrics", "locator": { "css": ".settings-panel" }, "name": "settings-scroll" },
+    {
+      "action": "dragTo",
+      "locator": { "testId": "available-widget" },
+      "dropTarget": { "testId": "enabled-widgets" }
+    },
+    {
+      "action": "uploadFiles",
+      "locator": { "label": "Import settings", "exact": true },
+      "files": [
+        { "name": "settings.json", "mimeType": "application/json", "base64": "eyJ0aGVtZSI6ImRhcmsifQ==" }
+      ]
+    },
+    {
+      "action": "click",
+      "locator": { "testId": "save-settings" },
+      "expectResponse": { "path": "/api/settings", "method": "PUT", "status": 200 },
+      "expectDialog": {
+        "type": "confirm",
+        "message": { "equals": "Save these settings?" },
+        "accept": true
+      }
+    },
+    { "action": "assertAttribute", "locator": { "testId": "save-settings" }, "attribute": "aria-busy", "equals": "false" },
     { "action": "assertText", "locator": { "testId": "toast" }, "includes": "Saved" },
+    { "action": "openPopup", "locator": { "testId": "preview-settings" }, "name": "preview" },
+    {
+      "action": "assertText",
+      "target": { "type": "popup", "name": "preview" },
+      "locator": { "role": "heading", "name": "Settings preview", "exact": true },
+      "equals": "Settings preview"
+    },
+    {
+      "action": "assertVisible",
+      "target": { "type": "frame", "locator": { "testId": "settings-help" } },
+      "locator": { "role": "heading", "name": "Settings help", "exact": true }
+    },
+    {
+      "action": "download",
+      "locator": { "testId": "export-settings" },
+      "filename": { "equals": "settings.json" },
+      "maxBytes": 1048576,
+      "retain": true,
+      "name": "settings-export"
+    },
     { "action": "screenshot", "name": "settings-saved" }
   ]
 }