@sun-asterisk/sungen 3.2.32 → 3.2.34

package/src/orchestrator/templates/ai-src/skills/sungen-gherkin-syntax/SKILL.md

@@ -629,7 +629,9 @@ on mobile a dialog-scoped ref still resolves against the whole screen (`SG-W021`
 | Tag | Effect |
 |---|---|
 | `@spec:<id>` | **Traceability**: this scenario proves requirement `<id>` from `requirements/spec.md` (`@spec:FR-003`). Repeatable — a scenario may prove several. Read by `sungen audit` (requirement coverage) and by delivery (the Coverage sheet's requirement table). Without it the link is only inferred from wording and cannot survive an edit. |
+| `@ui:<UI-NNN>` | **Design traceability**: this scenario proves design item `UI-NNN` from `requirements/ui-inventory.md` (`@ui:UI-014`). Repeatable. Read by `sungen audit` (`UI-ITEM-UNCOVERED`). An item deliberately not tested is recorded as a comment instead: `# UI-DEFERRED: UI-002..013 — <reason>`. Rules: `sungen-tc-generation` § Design Inventory. |
 | `@manual` | Skip in generation |
+| `@known-defect` | The scenario is EXPECTED to be red — it asserts a live defect. It still runs and still reports as failed (never skipped, never "fixed"), and it stays a pass-through tag so `--grep @known-defect` selects it. **`[mobile]`** Its failure does NOT escalate the next scenario's `@reset:` tier: a red that was designed in is not evidence of leftover state |
 | `@auth:role` | Use auth storage state for role |
 | `@no-auth` | Disable inherited auth |
 | `@steps:name` | Define reusable step block (base scenario) |
@@ -639,7 +641,8 @@ on mobile a dialog-scoped ref still resolves against the whole screen (`SG-W021`
 | `@cleanup:scroll` | Auto-cleanup: scroll to top after each test (cleanupPage) |
 | `@cleanup:storage` | Auto-cleanup: clear sessionStorage after each test (cleanupPage) |
 | `@screenshot:on-failure` | Auto-capture screenshot when test fails (base.ts fixture) |
-| `@parallel` | Opt-out: fresh page per test instead of serial default (for independent scenarios). Compiles on mobile too, but a single device/emulator gets no session-isolation benefit from it |
+| `@serial` | Opt-in: one shared page for an ordered, genuinely stateful sequence → `test.describe.serial()`. A failure skips the rest |
+| `@parallel` | Names the default (fresh page per test) — changes nothing. Compiles on mobile too, but a single device/emulator gets no session-isolation benefit from it |
 | `@beforeAll` | Hook: runs once before all tests → `test.beforeAll()` |
 | `@afterEach` | Hook: runs after each test → `test.afterEach()` (custom cleanup) |
 | `@afterAll` | Hook: runs once after all tests → `test.afterAll()` |
@@ -695,6 +698,14 @@ Rules that matter when generating:
   relaunch does not undo a half-written local DB. One rung, one scenario, and never from `none`.
   A **retry** of the failed scenario is NOT escalated — it re-runs at its declared tier, so a
   state-dependent failure fails both attempts and is reported honestly instead of passing on attempt 2.
+  A scenario tagged **`@known-defect`** escalates nothing either: its red was designed in, so making
+  the next scenario pay a harder reset for it buys nothing, on every run, forever.
+- **Escalation never picks a tier this session cannot perform.** Reinstalling needs the app's build
+  file, so when the app is attached by package the ladder caps at `data` (Android) / `soft` (iOS) and
+  says it capped, instead of throwing. Nothing asked for the escalated tier, and the throw would land
+  in `beforeEach` — a hook failure, which skips every remaining scenario in the FILE. The cap never
+  goes BELOW the declared tier: a tier you asked for that the session cannot do still throws by name,
+  because that one is an unmet contract rather than a recovery the runner chose.
 - **`SUNGEN_RESET_TIER=data` raises the floor for a whole run** without regenerating specs — CI can run
   stricter than local dev. It never lowers a declared tier.
 - **`data`/`full` on iOS reinstall the app**, because `mobile: clearApp` is Android-only (UiAutomator2).
@@ -765,14 +776,14 @@ npx playwright test --grep "@smoke|@high"    # smoke OR high
 
 ### Serial vs Parallel (test execution mode)
 
-**Default: serial** — `test.describe.serial()` with shared page. Background runs once in `beforeAll`. Fail → skip remaining.
+**Default: isolated** — `test.describe()` with a fresh page per test. Background runs as `beforeEach`; a failing scenario does not skip the others. Write every scenario so it stands alone — no scenario may rely on state another one left behind. `@parallel` only names this default.
 
-**`@parallel` opt-out** — `test.describe()` with fresh page per test. Background runs as `beforeEach`. Use when scenarios are truly independent (validation rules, permission tests).
+**`@serial` opt-in** — `test.describe.serial()` with one shared page. Background runs once in `beforeAll`; a failure skips the rest. Use it only for a genuinely stateful, ordered sequence.
 
 | Mode | Generated | Page | Background | On fail |
 |---|---|---|---|---|
-| Serial (default) | `test.describe.serial()` | Shared | `beforeAll` (1 goto) | Skip remaining |
-| `@parallel` | `test.describe()` | Fresh per test | `beforeEach` (N goto) | Continue |
+| Isolated (default, `@parallel`) | `test.describe()` | Fresh per test | `beforeEach` (N goto) | Continue |
+| `@serial` | `test.describe.serial()` | Shared | `beforeAll` (1 goto) | Skip remaining |
 
 **Multiple auth groups need the default isolation.** A feature mixing `@auth:user` with `@no-auth` compiles as-is — features are parallel by default, and `@parallel` only names that default. What breaks it is `@serial`: one shared browser context cannot carry two different sessions, so the compiler errors and asks you to drop `@serial` or split the feature.
 
@@ -933,7 +944,13 @@ Feature: User Management
 
 ### Layer 2: `@afterEach` scenario (custom cleanup)
 
-Only when `@cleanup:*` tags aren't enough — feature-specific logic.
+Only when `@cleanup:*` tags aren't enough — feature-specific logic. `@afterEach` gets the per-test
+`page`, so UI cleanup steps work in it.
+
+It runs after every test it applies to — and it does NOT apply to the `@no-auth` group when the
+feature mixes `@no-auth` with other scenarios: an anonymous test has an empty session, so a cleanup
+that needs the signed-in page would fail it although its own steps passed. The hook is emitted inside
+each other group instead. A feature whose scenarios are ALL `@no-auth` keeps the hook for them.
 
 ### Layer 3: `@beforeAll` / `@afterAll` (optional)
 
@@ -958,8 +975,9 @@ Scenario: drop the seeded data after the last scenario
   recorded and the next step still runs; a failing action step (`Given`/`When`) stops the hook there.
   The hook fails ONCE with the full list (`Hook preconditions not met`), not with the first failure.
 - `@cases` and `@extend` are not supported on a hook — they are ignored with a compile warning.
-- A hook has no `page` fixture on a per-scenario (non-`@serial`) feature — keep its steps to `@api:`/
-  `@query:` preconditions and `expect` assertions there.
+- `@beforeAll` / `@afterAll` have no `page` fixture on an isolated (non-`@serial`) feature — keep
+  their steps to `@api:`/`@query:` preconditions and `expect` assertions there. `@afterEach` does get
+  the per-test `page`.
 
 **Rendering order in `.spec.ts`:**
 `test.describe` → `test.use(storageState)` → `test.use(autoCleanup)` → `test.beforeAll` → `test.beforeEach` → `test.afterEach` → `test.afterAll` → `test()` blocks.

package/src/orchestrator/templates/ai-src/skills/sungen-selector-fix/SKILL.md

@@ -225,12 +225,33 @@ Skip if `specs/.auth/<role>.json` already exists and a probe navigation reaches
 
 ---
 
+## Where each wave writes its result
+
+`sungen delivery` reads the per-target result file (`specs/generated/<name>/<name>-test-result.json`),
+which the JSON reporter in `playwright.config.ts` writes. Keep it that way in every wave:
+
+- **Never pass `--reporter`** — it replaces the config's reporters, the JSON reporter is switched off, and the run writes no result file.
+- **Subset waves (Phase 1–2) write to `test-results/sungen-wave.json`** — never to the per-target file (a subset would overwrite the unit's full result) and never to the default `test-results/results.json` (delivery falls back to that file when a per-target one is missing, so a 5-scenario smoke would read as the unit's result). `test-results/` is Playwright's output dir, wiped on every run — nothing to clean up.
+- **Full waves (Phase 3–4) write the per-target file** — the Phase 4 file is the result delivery shows.
+- **Scope every run to the unit** — `specs/generated/<name>/` (flow: `specs/generated/flows/<name>/`). Without a path, `--grep` runs every unit in `specs/generated/`.
+
+### Fix retries re-run only what the wave left red (Phase 2–3)
+
+- **The first run of a phase is its full wave.** Each **fix retry** is the same command plus `--last-failed`, written to `test-results/sungen-wave.json` — never to the per-target file.
+- **Retry immediately after the run whose failures you fixed.** `--last-failed` reads the failures of the *previous run only*: run any other test in between (a `-g` debug run) and the list now describes that run. If you did, re-run the full wave instead.
+- **Nothing left red → no retry.** On an empty list Playwright prints "No tests found", exits 1 and still writes a zero-test report — that is not a failure of the unit, and it is why a retry never writes the per-target file.
+- **A locale run (`SUNGEN_ENV=<env>`) writes `test-results/sungen-wave.<env>.json`** — the config inserts the env before `.json`; it is still a scratch file, nothing reads it.
+- **A retry does not show whether your fix broke a test that was green.** Phase 3's full first run catches a Phase 2 fix that did; Phase 4 catches a Phase 3 one.
+
+---
+
 ## Phase 1: Smoke Check (catch fundamentals)
 
 Run **up to 5 scenarios** — pick the first `@high` scenarios in the feature file.
 
 ```bash
-npx playwright test --grep "VP-.*-001|VP-.*-002|VP-.*-003|VP-.*-004|VP-.*-005" --reporter=line
+PLAYWRIGHT_JSON_OUTPUT_NAME=test-results/sungen-wave.json \
+  npx playwright test specs/generated/<name>/ --grep "VP-.*-001|VP-.*-002|VP-.*-003|VP-.*-004|VP-.*-005"
 ```
 
 **Purpose:** Detect broken fundamentals before running 50+ tests:
@@ -248,7 +269,15 @@ npx playwright test --grep "VP-.*-001|VP-.*-002|VP-.*-003|VP-.*-004|VP-.*-005" -
 Run all `@high` scenarios:
 
 ```bash
-npx playwright test --grep "@high" --reporter=line
+PLAYWRIGHT_JSON_OUTPUT_NAME=test-results/sungen-wave.json \
+  npx playwright test specs/generated/<name>/ --grep "@high"
+```
+
+Fix retry — only the scenarios this wave left red:
+
+```bash
+PLAYWRIGHT_JSON_OUTPUT_NAME=test-results/sungen-wave.json \
+  npx playwright test specs/generated/<name>/ --grep "@high" --last-failed
 ```
 
 If your Playwright config doesn't support tag grep, use scenario name grep from the feature file — collect VP IDs of `@high` scenarios.
@@ -261,12 +290,24 @@ Max 2 fix attempts in this phase.
 
 ## Phase 3: Full Run (@normal + @low)
 
-Run remaining scenarios:
+Run remaining scenarios — the whole unit, into its per-target result file:
 
 ```bash
-npx playwright test --reporter=line
+PLAYWRIGHT_JSON_OUTPUT_NAME=specs/generated/<name>/<name>-test-result.json \
+  npx playwright test specs/generated/<name>/<name>.spec.ts
 ```
 
+A multi-feature screen (one result file per feature) or a locale run (`SUNGEN_ENV`): use the loop in `run-test` § "Playwright command guidelines".
+
+Fix retry — the failures only, into the wave file (the per-target file keeps this phase's full run until Phase 4 rewrites it):
+
+```bash
+PLAYWRIGHT_JSON_OUTPUT_NAME=test-results/sungen-wave.json \
+  npx playwright test specs/generated/<name>/<name>.spec.ts --last-failed
+```
+
+**Multi-feature screen — no `--last-failed` here.** Phase 3 ran one Playwright run per feature spec, so the failure list holds only the LAST spec's failures; a retry would skip every earlier feature's red tests without a word. Re-run the per-spec loop instead.
+
 Many selectors already fixed from Phase 2 (shared elements). Only diagnose **new** failures — selectors that only appear in lower-priority scenarios.
 
 Max 1 fix attempt. If `@low` scenarios still fail after fix → **report and move on**, don't loop.
@@ -275,12 +316,15 @@ Max 1 fix attempt. If `@low` scenarios still fail after fix → **report and mov
 
 ## Phase 4: Regression
 
-One final full run to confirm all phases together:
+One final full run to confirm all phases together — this per-target file is the result `sungen delivery` reads:
 
 ```bash
-npx playwright test --reporter=line
+PLAYWRIGHT_JSON_OUTPUT_NAME=specs/generated/<name>/<name>-test-result.json \
+  npx playwright test specs/generated/<name>/<name>.spec.ts
 ```
 
+Same multi-feature / locale form as Phase 3.
+
 Report results. Do NOT enter another fix loop here.
 
 ---
@@ -458,13 +502,13 @@ user detail:
 
 ## Attempt Budget Summary
 
-| Phase | What runs | Max fix attempts | On failure after max |
-|---|---|---|---|
-| 0. Pre-gen | Playwright MCP snapshot → write selectors.yaml | 1 snapshot | Ask user — skip or retry navigation |
-| 0.5. Auth | Manual login in MCP browser → `browser_storage_state` → `specs/.auth/<role>.json` | 1 login | Ask user — retry login or fall back to `sungen makeauth` |
-| 1. Smoke | First 5 @high | 2 | Ask user — fundamentals broken |
-| 2. Priority | All @high | 2 | Report failures, continue to Phase 3 |
-| 3. Full | All tests | 1 | Report @low/@normal failures, continue |
-| 4. Regression | All tests | 0 | Report final results |
+| Phase | What runs | Fix retry runs | Max fix attempts | On failure after max |
+|---|---|---|---|---|
+| 0. Pre-gen | Playwright MCP snapshot → write selectors.yaml | — | 1 snapshot | Ask user — skip or retry navigation |
+| 0.5. Auth | Manual login in MCP browser → `browser_storage_state` → `specs/.auth/<role>.json` | — | 1 login | Ask user — retry login or fall back to `sungen makeauth` |
+| 1. Smoke | First 5 @high | the same 5 | 2 | Ask user — fundamentals broken |
+| 2. Priority | All @high | `--last-failed` | 2 | Report failures, continue to Phase 3 |
+| 3. Full | All tests | `--last-failed` (multi-feature screen: the per-spec loop again) | 1 | Report @low/@normal failures, continue |
+| 4. Regression | All tests | — | 0 | Report final results |
 
 **Total worst case: 5 fix attempts** (2+2+1), not unbounded loops. Phases 0 and 0.5 don't count toward fix budget.

package/src/orchestrator/templates/ai-src/skills/sungen-selector-keys/SKILL.md

@@ -128,7 +128,10 @@ back to the feature's own path (or `/<ref>/`) and warns `SG-W013`: a test that r
 
 If the label is used by only ONE type, the base key (no suffix) is fine — suffix only on collision.
 
-Type aliases (normalized automatically):
+Type aliases (normalized automatically). The suffix and the step may use **different spellings of
+the same type** — `confirm--modal` resolves for `[Confirm] dialog` and `confirm--dialog` resolves
+for `[Confirm] modal`. Prefer the normalized suffix for consistency; either is found.
+
 | Gherkin type | Normalized suffix |
 |---|---|
 | title, label, caption, message | `text` |
@@ -143,6 +146,15 @@ Type aliases (normalized automatically):
 | column | `columnheader` |
 | list-item | `listitem` |
 
+When the step names a type from this table, only a key whose suffix is a spelling of that type is
+used — `[Confirm] dialog` will not fall back to an unrelated `confirm--popup` entry. It reports the
+keys it looked for, so the fix is to add one of them.
+
+A bare `[Confirm]` keeps the older, looser fallback and matches any `confirm--*` entry, because with
+no type there is nothing to match a suffix against. So does a `[Ref]` followed by a word that is not
+an element type at all — `Then User see [Quick Panel] is hidden` reads `is` as that word. This is
+what keeps state-clause steps working on a label whose keys are all suffixed.
+
 ## Nth-Suffixed Keys
 
 When targeting a specific occurrence by index, add `--N` suffix:

package/src/orchestrator/templates/ai-src/skills/sungen-tc-generation/SKILL.md

@@ -154,7 +154,9 @@ For append: read highest `VP-<CAT>-<NNN>`, continue from next number. Never modi
 Auto-detected by `create-test` before invoking this skill:
 - `spec.md` — primary, always read if present
 - `spec_figma.md` — Figma supplement (PAT flow already done — just read the file)
-- `ui/*.png` — visual context
+- `ui/*.{png,jpg,jpeg,webp,gif}` — design images. **Every one is read**, and turned into the Design Inventory below — not just "visual context".
+- `ui/figma-meta.md` — Figma MCP capture (`sungen-capture` mode figma-mcp): `## Components` and `## Text Content` feed the Design Inventory; `## Text Content` also supplies exact labels for `test-data.yaml`.
+- `ui-inventory.md` — the Design Inventory (written by create-test, see § Design Inventory). When it already exists, keep its ids and append new items; never renumber.
 - `test-viewpoint.md` — test conditions checklist; scan entirely regardless of format.
   Format varies across projects: hierarchical table (Level 1–4 columns), flat checklist,
   prose, or mixed. Regardless of format:
@@ -181,6 +183,40 @@ This means: do not test other screens' UI layout or navigation. It does NOT mean
 
 ---
 
+## Design Inventory (UI items)
+
+**When:** the unit holds any design source — images in `requirements/ui/`, `ui/figma-meta.md`, or `spec_figma.md`. Without one, skip this section.
+
+**Why:** a design read "for visual context" leaves no trace. An item that exists only on the design (a column `spec.md` never names) or contradicts the spec is then missed by the author AND the reviewer, and "deliberately deferred" looks exactly like "never seen". `sungen audit` checks this file (`UI-INVENTORY-MISSING`, `UI-INVENTORY-INCOMPLETE`, `UI-ITEM-UNCOVERED`).
+
+**Reading order:** `spec.md` (authoritative) → `spec_figma.md` → `ui/figma-meta.md` (`## Components`, `## Text Content`) → **every** image in `ui/`. Many images: group them (primary screen vs states/variants) but read them all — an unread state screen is an un-inventoried one.
+
+**Write `requirements/ui-inventory.md`** — one row per component, table column, button/link, field, state (empty/error/loading/disabled), and design annotation:
+
+```markdown
+# UI Inventory — <screen>
+images read: 3/3 · figma-meta.md: read · spec_figma.md: absent
+
+| ID | Item | Kind | Source | Seen in |
+|---|---|---|---|---|
+| UI-001 | Job List table | table | spec+design | ui/list.png |
+| UI-002 | Contract plan | column | spec+design | ui/list.png |
+| UI-014 | Salary range | column | design-only | ui/list.png |
+| UI-015 | Status badge "Closed" | column | conflict — spec says "Expired" | ui/list.png |
+```
+
+- **Source:** `spec+design` (both name it) · `design-only` (only the design shows it) · `conflict` (they disagree — say how in the cell). `spec.md` stays authoritative; a conflict is reported, never silently resolved in either direction.
+- **IDs** `UI-NNN` are append-only, like scenario ids: re-runs keep existing rows and continue numbering.
+- `images read: N/N` counts the images you actually opened against the images in the folder.
+
+**Account for every `design-only` / `conflict` item in the `.feature`** — either:
+- **cover it**: tag the scenario that proves it `@ui:UI-014`, or
+- **defer it with a reason** in a feature comment: `# UI-DEFERRED: UI-002..013 — per-column checks are owned by job-detail`. A deferral without a reason defers nothing.
+
+`spec+design` items are covered through the spec (`@spec:`), and citing their ids is optional.
+
+---
+
 ## Generation Workflow
 
 ### Step 1 — Orient: patterns + risk
@@ -219,6 +255,10 @@ Viewpoint items:     [TV-01 – <exact wording from test-viewpoint row 1>]
                      [TV-02 – <exact wording from test-viewpoint row 2>]
                      → one entry per row/bullet in test-viewpoint.md, regardless of format
                      → leave empty if no test-viewpoint.md present
+UI items:            [UI-014 – Salary range column (design-only) → covered by VP-TAB-012]
+                     [UI-002..013 – per-column checks → deferred: owned by job-detail]
+                     → one entry per Design Inventory row: covered / deferred-with-reason / uncovered
+                     → leave empty if the unit has no design source
 Spec gaps:           [G1 – behavior undocumented → verify with dev]
 ```
 

package/src/orchestrator/templates/ai-src/skills/sungen-tc-review/SKILL.md

@@ -152,6 +152,7 @@ Build a mapping table: for each applicable group, does the feature have a matchi
 | **Account Status** | Deleted / blocked / inactive account login | VP-LOGIC | corresponding account data → error message |
 | **Security** | SQL Injection; XSS; Data Integrity (edit dropdown/remove `disabled` via DevTools → server rejects). SQL on LIKE/search fields → 2 TCs per the `sungen-viewpoint` SQL 2-layer rule. | VP-SEC | `with {{xss_*}}`/`{{sql_*}}`; data-integrity usually `@manual` |
 | **Cross-surface outcomes** | Admin action → outcome on user-facing surface (portal, mobile, widget); spec defines display condition on another URL | VP-LOGIC-CS | `@high @manual` per surface per business rule — at minimum 1 per cross-surface rule; use `@auth:role` + `@extend` if surface reachable in same test run |
+| **Design items** *(when `requirements/ui/*`, `ui/figma-meta.md` or `spec_figma.md` exist)* | Every component / column / button / state the design shows is in `ui-inventory.md`; each `design-only` / `conflict` item is covered or deferred with a reason. Read the images yourself — the inventory can be incomplete | UI-NNN | `@ui:UI-NNN` on the proving scenario, or `# UI-DEFERRED: UI-NNN — <reason>`; rules in `sungen-tc-generation` § Design Inventory |
 
 **Mobile units add four categories** (`@platform:mobile/android/ios` only). Each maps to a theme the
 `sungen audit` mobile catalog scores, so a gap here is a scored gap, not a stylistic note — the
@@ -374,6 +375,12 @@ If `selectors/<name>.yaml` exists, count lines matching `@needs-live-verify` (pr
 | Behavior Handling / Account Status | ... | ... |
 | Security (XSS/SQLi) | ... | ... |
 | Cross-surface outcomes | ... | ... |
+| Design items | n/a / ✓ / ✗ | ... |
+
+### Design ↔ Feature (when the unit holds design sources)
+- Images read: <n>/<n> · inventory items: <n> (covered <n> · deferred <n> · uncovered <n>)
+- Not inventoried: <item> (<image>) …
+- Mislabelled / unproven: UI-NNN …
 
 ### ✗ Must-fix
 1. [DIMENSION] VP-XXX issue description

package/src/orchestrator/templates/playwright.config.ts

@@ -105,8 +105,10 @@ export default defineConfig({
     /* Per-action timeout (click, fill, etc.) */
     actionTimeout: 10_000,
 
-    /* Collect trace when retrying the failed test. See https://playwright.dev/docs/trace-viewer */
-    trace: 'on-first-retry',
+    /* Trace for a failed test. CI: recorded on the first retry (CI retries). Local: no retries, so
+       'on-first-retry' would never record one — keep the trace of every failed test instead.
+       See https://playwright.dev/docs/trace-viewer */
+    trace: process.env.CI ? 'on-first-retry' : 'retain-on-failure',
 
     /**
      * Capture screenshot after each test failure.

package/src/orchestrator/templates/specs-locale-fixture.ts

@@ -42,18 +42,30 @@ interface LocaleInjection {
 
 let cached: LocaleInjection | null | undefined;
 
+/**
+ * Absent config → null, and locale injection is simply off (the documented no-op).
+ *
+ * A config that EXISTS but does not parse is a different thing and must not read the same way: the
+ * header above tells you to hand-edit this file, and a trailing comma would otherwise turn a
+ * `SUNGEN_ENV=ja` run into a base-locale run that asserts base-locale strings — green, and wrong.
+ * We warn rather than throw, because the injection is genuinely optional and failing every test on
+ * it would be a larger change than the fault warrants; the line names the file and the consequence.
+ */
 function loadLocaleConfig(): LocaleInjection | null {
   if (cached !== undefined) return cached;
+  const cfgPath = path.join(__dirname, 'locale-config.json');
+  if (!fs.existsSync(cfgPath)) {
+    cached = null;
+    return null;
+  }
   try {
-    const cfgPath = path.join(__dirname, 'locale-config.json');
-    if (!fs.existsSync(cfgPath)) {
-      cached = null;
-      return null;
-    }
-    const parsed = JSON.parse(fs.readFileSync(cfgPath, 'utf-8')) as LocaleInjection;
-    cached = parsed;
-    return parsed;
-  } catch {
+    cached = JSON.parse(fs.readFileSync(cfgPath, 'utf-8')) as LocaleInjection;
+    return cached;
+  } catch (e) {
+    console.error(
+      `⚠️  ${cfgPath}: unreadable (${e instanceof Error ? e.message.split('\n')[0] : String(e)}) — ` +
+      `locale state is NOT injected, so this SUNGEN_ENV run behaves as the base locale.`,
+    );
     cached = null;
     return null;
   }

package/src/orchestrator/templates/specs-url-assert.ts

@@ -46,6 +46,16 @@ export interface UrlExpectation {
   declared?: string;
   /** The test-data key behind `declared`, so a failure can name what an author must go fix. */
   dataRef?: string;
+  /**
+   * The page selector declared `query: any` — the query string is not part of this page's identity,
+   * so SILENCE from the step stops meaning "there must be no query" (#748).
+   *
+   * Emitted only for a page whose entry says so. The default stays strict: a bare page means a bare
+   * page, which is the #464 assertion and the reason a filter regression cannot read as green.
+   * It is the no-`dataRef` branch alone that this relaxes — a step that declares `with {{v}}` still
+   * asserts every param it names, on this page as on any other.
+   */
+  ignoreQuery?: boolean;
 }
 
 /**
@@ -113,8 +123,14 @@ function locationMatches(pathname: RegExp, hostname: RegExp | undefined, u: URL)
  *                      tracking/UTM param the test never cared about cannot turn it red.
  *                      A declared fragment is asserted too; a declared value carrying no query
  *                      at all constrains the path only (the compiler warns: SG-W014).
+ *
+ * `query: any` on the page's selector entry overrides the FIRST rule only (#748): silence then
+ * means "the query is not mine to judge" instead of "there must be none". It is the escape hatch
+ * for a page whose query is generated at run time — `/admins/detail?id=<the row just clicked>` —
+ * which the strict rule makes unassertable and `with {{v}}` cannot express, since that asserts a
+ * literal. The second rule is untouched: declare params and they are still all checked.
  */
-export function urlMatches({ pathname, hostname, declared, dataRef }: UrlExpectation): (u: URL) => boolean {
+export function urlMatches({ pathname, hostname, declared, dataRef, ignoreQuery }: UrlExpectation): (u: URL) => boolean {
   return (u: URL): boolean => {
     if (!locationMatches(pathname, hostname, u)) return false;
 
@@ -123,7 +139,13 @@ export function urlMatches({ pathname, hostname, declared, dataRef }: UrlExpecta
     // undefined, and they mean opposite things: the first is a real assertion (a bare page must be
     // bare), the second is broken test-data. Only the generator knows which, and it says so by
     // emitting `dataRef` exactly when the step carried a `with {{v}}`.
-    if (!dataRef) return u.search === '';
+    // `query: any` on the page entry (#748): the query is not part of this page's identity, so
+    // silence from the step is not an assertion about it. Without this there is NO way to express
+    // "this path, whatever the query" — `with {{v}}` asserts a LITERAL value, and a detail page
+    // keyed by `?id=<the row just clicked>` has no literal to declare. Checked before the bare-page
+    // rule, never instead of the declared-value rule below: a step that names params still asserts
+    // them here exactly as anywhere else.
+    if (!dataRef) return ignoreQuery ? true : u.search === '';
 
     // Silence is the whole bug class this assertion exists to kill, so an unusable declared value
     // is an ERROR, not a quiet verdict. Returning false would read as "the app is wrong"; these
@@ -183,6 +205,26 @@ interface SettleOpts { settleMs?: number; networkIdleMs?: number }
  * 3000) bounds step 1. Both are runtime knobs, so a slow environment is a config change, not a
  * regenerate.
  */
+/**
+ * A millisecond budget from the environment. `Number()` alone turned `"1s"` / `"abc"` into NaN and
+ * `""` into 0 (`??` keeps an empty string): the hold loop then never ran, so a late redirect passed
+ * unseen, and a 0 / NaN networkidle timeout means "no timeout" to Playwright — the step waited out
+ * the whole test (#856). Unset or empty means the default; a malformed value falls back to it and
+ * says so, once per variable. `allowZero` for a knob where 0 is a real choice (no hold).
+ */
+const warnedEnv = new Set<string>();
+function msEnv(name: string, fallback: number, allowZero = false): number {
+  const raw = process.env[name]?.trim();
+  if (!raw) return fallback;
+  const value = Number(raw);
+  if (Number.isFinite(value) && (value > 0 || (allowZero && value === 0))) return value;
+  if (!warnedEnv.has(name)) {
+    warnedEnv.add(name);
+    console.warn(`[sungen] ${name}=${JSON.stringify(raw)} is not a ${allowZero ? 'non-negative' : 'positive'} number of milliseconds — using the default ${fallback}.`);
+  }
+  return fallback;
+}
+
 async function settleAndHold(
   page: SettleablePage,
   matches: (u: URL) => boolean,
@@ -190,8 +232,9 @@ async function settleAndHold(
   label: string,
   consequence: string,
 ): Promise<void> {
-  const settleMs = opts.settleMs ?? Number(process.env.SUNGEN_URL_SETTLE_MS ?? 1000);
-  const networkIdleMs = opts.networkIdleMs ?? Number(process.env.SUNGEN_URL_NETWORKIDLE_MS ?? 3000);
+  // 0 holds nothing (the operator's call); 0 for networkidle would mean "no timeout", so it is refused.
+  const settleMs = opts.settleMs ?? msEnv('SUNGEN_URL_SETTLE_MS', 1000, true);
+  const networkIdleMs = opts.networkIdleMs ?? msEnv('SUNGEN_URL_NETWORKIDLE_MS', 3000);
   const arrivedAt = page.url();
 
   await page.waitForLoadState('networkidle', { timeout: networkIdleMs }).catch(() => undefined);

package/src/orchestrator/templates/wdio.conf.ts

@@ -48,6 +48,14 @@ function evidenceSlug(s: string): string {
     .slice(0, 80) || 'unnamed';
 }
 
+/**
+ * True only in the launcher process. `@wdio/runner` re-parses this very config inside EVERY worker
+ * it spawns, so anything logged at module scope prints once per spec unless it is gated. WDIO's own
+ * "am I a worker" test is this env var (`@wdio/local-runner` sets it when spawning; `@wdio/utils`
+ * reads it) — reuse it rather than inventing a second signal.
+ */
+const IS_LAUNCHER = !process.env.WDIO_WORKER_ID;
+
 /**
  * Auto-select MOBILE specs and classify each by its target OS (dual-target routing).
  *
@@ -82,7 +90,12 @@ function classifyMobileSpecs(root: string): Array<{ file: string; platform: stri
     .filter((full) => {
       try {
         return fs.readFileSync(full, 'utf-8').includes('@wdio/globals');
-      } catch {
+      } catch (e) {
+        // A spec listed by readdir that cannot be read cannot be classified, so it is excluded.
+        // Excluding it SILENTLY is the fault: WDIO only fails the run when the spec list is
+        // EMPTY ("No specs found to run"), so dropping one spec out of ten leaves a green suite
+        // containing a test that never executed. Name it, and let the count below show the shortfall.
+        if (IS_LAUNCHER) console.warn(`[sungen] could not read ${full} — EXCLUDED from this run: ${(e as Error).message}`);
         return false;
       }
     })
@@ -98,6 +111,16 @@ const androidSpecs = specEntries.filter((e) => e.platform !== 'ios').map((e) =>
 const iosSpecs = specEntries.filter((e) => e.platform !== 'android').map((e) => e.file);
 const mobileSpecs = specEntries.map((e) => e.file); // union — top-level fallback
 
+// What DISCOVERY found — a mobile suite is only trustworthy if you can see how many specs it
+// resolved: WDIO fails loudly on an EMPTY list but says nothing when the list is merely short,
+// which is exactly what a dropped spec looks like.
+//
+// Two things this number is not. It is not an error on 0: WDIO already exits 1 there
+// ("No specs found to run"), and a `--shard` that legitimately receives none must still return 0.
+// And under `--shard` it is the PRE-SHARD total, because ConfigParser slices `specs` down to the
+// shard's portion only after this file has finished evaluating.
+if (IS_LAUNCHER) console.log(`[sungen] specs: ${specEntries.length} mobile spec(s) — android ${androidSpecs.length}, ios ${iosSpecs.length}`);
+
 // App under test — env overrides the init-configured defaults.
 const APP_PACKAGE = process.env.APP_PACKAGE || '__APP_PACKAGE__';
 const APP_ACTIVITY = process.env.APP_ACTIVITY || '__APP_ACTIVITY__';

package/src/orchestrator/trace-mode-install.ts

@@ -0,0 +1,94 @@
+/**
+ * Move an existing project's `use.trace` off the old default so local runs keep a trace.
+ *
+ * The old template paired `trace: 'on-first-retry'` with `retries: process.env.CI ? 2 : 0`: off CI
+ * there is no retry, so there is never a "first retry" and a QA fixing a test on their machine never
+ * got a trace. The template now records `on-first-retry` on CI and `retain-on-failure` locally.
+ *
+ * `playwright.config.ts` is written once at `sungen init` and never re-synced (see
+ * report-sanitize-reporter-install.ts, the same problem for the sanitize reporter), so `sungen update`
+ * calls this. It rewrites the trace line ONLY while it is still the old default — any other value is
+ * a choice the user made, and a wrong edit to a user's config is worse than a missing trace. The same
+ * goes for a `trace` inside `projects: [...]`: that is a per-project override, not the template default.
+ */
+import * as fs from 'fs';
+import * as path from 'path';
+
+/** The template's trace value: cheap on CI (retries exist there), a trace for every local failure. */
+export const TRACE_MODE_EXPRESSION = "process.env.CI ? 'on-first-retry' : 'retain-on-failure'";
+
+/**
+ *   'patched' — the old default was rewritten now
+ *   'present' — already the template expression (idempotent)
+ *   'custom'  — another value, no trace line, more than one old-default line, or one inside
+ *               `projects: [...]`; nothing written
+ *   'absent'  — no playwright.config.ts (a mobile-only project, or not a sungen project)
+ */
+export type TraceModeOutcome = 'patched' | 'present' | 'custom' | 'absent';
+
+const escapeRegExp = (s: string): string => s.replace(/[.*+?^${}()|[\]\\]/g, '\\$&');
+
+// Both are anchored to the start of a line, so a commented-out copy (`// trace: …`) never counts.
+// OLD_DEFAULT keeps the indentation in $1 and leaves the trailing comma, comment and EOL alone.
+const OLD_DEFAULT = /^([ \t]*trace:[ \t]*)(['"])on-first-retry\2/gm;
+const PRESENT = new RegExp(`^[ \\t]*trace:[ \\t]*${escapeRegExp(TRACE_MODE_EXPRESSION)}`, 'm');
+
+// The old template's one-line comment, only when it sits directly above the (patched) trace line.
+// A comment the user wrote is theirs and stays.
+const OLD_COMMENT = new RegExp(
+  `^([ \\t]*)/\\* Collect trace when retrying the failed test\\. See https://playwright\\.dev/docs/trace-viewer \\*/(\\r?\\n)(?=[ \\t]*trace:[ \\t]*${escapeRegExp(TRACE_MODE_EXPRESSION)})`,
+  'm',
+);
+const newComment = (indent: string, eol: string): string =>
+  `${indent}/* Trace for a failed test. CI: recorded on the first retry (CI retries). Local: no retries, so${eol}` +
+  `${indent}   'on-first-retry' would never record one — keep the trace of every failed test instead.${eol}` +
+  `${indent}   See https://playwright.dev/docs/trace-viewer */${eol}`;
+
+/**
+ * [start, end] offsets of each live `projects: [...]` array, matched by bracket depth. Brackets inside
+ * a `//` or `/* *\/` comment or a quoted string are skipped: one stray `]` in a note would otherwise
+ * close the span early and expose a per-project override to the patch.
+ */
+function projectsSpans(src: string): Array<[number, number]> {
+  const spans: Array<[number, number]> = [];
+  for (const m of src.matchAll(/^[ \t]*projects\s*:\s*\[/gm)) {
+    const open = m.index! + m[0].length - 1;
+    let depth = 0;
+    for (let i = open; i < src.length; i++) {
+      const ch = src[i];
+      if (ch === '/' && src[i + 1] === '/') { const nl = src.indexOf('\n', i); i = nl < 0 ? src.length : nl; continue; }
+      if (ch === '/' && src[i + 1] === '*') { const end = src.indexOf('*/', i + 2); i = end < 0 ? src.length : end + 1; continue; }
+      if (ch === "'" || ch === '"' || ch === '`') {
+        for (i++; i < src.length && src[i] !== ch; i++) if (src[i] === '\\') i++;
+        continue;
+      }
+      if (ch === '[') depth++;
+      else if (ch === ']' && --depth === 0) { spans.push([open, i]); break; }
+    }
+  }
+  return spans;
+}
+
+export function patchTraceMode(cwd: string, dryRun: boolean = false): TraceModeOutcome {
+  const configPath = path.join(cwd, 'playwright.config.ts');
+  if (!fs.existsSync(configPath)) return 'absent';
+  const src = fs.readFileSync(configPath, 'utf8');
+  if (PRESENT.test(src)) return 'present';
+  const hits = [...src.matchAll(OLD_DEFAULT)];
+  if (hits.length !== 1) return 'custom'; // none: the user's own value · several: ambiguous, don't guess
+  const at = hits[0].index!;
+  if (projectsSpans(src).some(([s, e]) => at > s && at < e)) return 'custom'; // a per-project override
+  if (!dryRun) {
+    const patched = src
+      .replace(OLD_DEFAULT, `$1${TRACE_MODE_EXPRESSION}`)
+      .replace(OLD_COMMENT, (_m, indent: string, eol: string) => newComment(indent, eol));
+    fs.writeFileSync(configPath, patched);
+  }
+  return 'patched';
+}
+
+/** `sungen update` step: patch the trace mode and say so when it changed. */
+export function installTraceMode(cwd: string, dryRun: boolean = false): void {
+  if (patchTraceMode(cwd, dryRun) !== 'patched') return;
+  console.log(`${dryRun ? '📋 would update' : '✓ updated'} playwright.config.ts: local runs now keep a trace for every failed test (trace: ${TRACE_MODE_EXPRESSION})`);
+}