@sun-asterisk/sungen 3.2.21-beta.2 → 3.2.21

package/src/exporters/matrix/render-xlsx.ts

@@ -224,7 +224,7 @@ function addMatrixSheet(wb: ExcelJS.Workbook, model: MatrixModel, sungenVersion?
         // parent's steps + these read as one continuous procedure.
         numbered(item.triggerDeltas[v.ref] ?? [], item.trigger.length + 1),
         expectedWithVerification(v.oracle, v.verification),
-        v.mode === 'manual' ? `Manual${v.manualReason ? ` (${v.manualReason})` : ''}` : (v.mocked ? 'Auto · Mocked' : 'Auto'),
+        v.mode === 'manual' ? `Manual${v.manualReason ? ` (${v.manualReason})` : ''}` : 'Auto',
         v.traces.join(', '),
         variantState(v),
         isoDate(v.result?.startTime),
@@ -315,50 +315,6 @@ function addCoverageSheet(wb: ExcelJS.Workbook, model: MatrixModel, sungenVersio
 
   let rowIdx = 7;
 
-  // --- Performance budget (#569) — advisory: config (qa/perf.yaml) vs measured runs.
-  if (model.perf) {
-    const p = model.perf;
-    dataCell(ws.getRow(rowIdx).getCell(2), 'Performance budget', { bold: true });
-    rowIdx++;
-    const r = ws.getRow(rowIdx);
-    dataCell(r.getCell(2), `p${p.percentile} ${p.metric.replace('_ms', ' time')}`);
-    dataCell(r.getCell(3), p.pass ? 'PASS' : 'OVER BUDGET', { bold: true });
-    dataCell(r.getCell(4), `measured ${p.measuredMs}ms`);
-    dataCell(r.getCell(5), `budget ${p.budgetMs}ms`);
-    dataCell(r.getCell(6), `${p.samples} runs`);
-    if (!p.pass && p.slowest.length) dataCell(r.getCell(8), `slowest: ${p.slowest.map((x) => `${x.ms}ms ${x.title.slice(0, 40)}`).join(' · ')}`);
-    rowIdx += 2;
-  }
-
-  // --- External Dependency Tracking (#580, ISTQB system-test practice): every case that
-  // ran against a MOCK is listed centrally. Its Result stays in the canonical vocabulary
-  // (the roll-up formulas depend on it); THIS table is where "Pass (Mocked) ≠ Pass (Real)"
-  // is visible — a case verified only against mocks is not Done until re-verified real.
-  {
-    const mockedVars = model.items.flatMap((it) => it.variants.filter((v) => v.mocked).map((v) => ({ it, v })));
-    if (mockedVars.length > 0) {
-      dataCell(ws.getRow(rowIdx).getCell(2), 'External Dependency Tracking — mock-verified cases (Pass here = Pass (Mocked), not Pass (Real))', { bold: true });
-      rowIdx++;
-      const hd = ws.getRow(rowIdx);
-      dataCell(hd.getCell(2), 'Variant');
-      dataCell(hd.getCell(3), 'Mocked dependency');
-      dataCell(hd.getCell(4), 'Delivery item');
-      dataCell(hd.getCell(5), 'Result');
-      dataCell(hd.getCell(8), 'Real-environment re-verify');
-      rowIdx++;
-      for (const { it, v } of mockedVars) {
-        const r = ws.getRow(rowIdx);
-        dataCell(r.getCell(2), v.ref);
-        dataCell(r.getCell(3), (v.mockDeps ?? []).join(', ') || '(scenario-level @mock)');
-        dataCell(r.getCell(4), it.id);
-        dataCell(r.getCell(5), variantState(v));
-        dataCell(r.getCell(8), '');
-        rowIdx++;
-      }
-      rowIdx++;
-    }
-  }
-
   // --- Requirement coverage (review §6): every requirement id gets ONE explicit status.
   if (model.requirements.length > 0) {
     dataCell(ws.getRow(rowIdx).getCell(2), 'Requirement coverage — every id has an explicit status', { bold: true });

package/src/exporters/matrix/types.ts

@@ -92,12 +92,6 @@ export interface CoverageVariant {
   /** 'auto' | 'manual'; manual keeps its @manual:Mx reason. */
   mode: 'auto' | 'manual';
   manualReason?: string;
-  /** The scenario runs against MOCKED network responses (@mock / `[x] mock is active`) —
-   *  a pass verifies the handling logic, not the real dependency (Pass (Mocked) ≠ Pass (Real),
-   *  #580: ISTQB external-dependency tracking). */
-  mocked?: boolean;
-  /** Named mocked dependencies (`[name] mock is active` refs), for the tracking table. */
-  mockDeps?: string[];
   layers: MatrixLayer[];
   /** @spec:* requirement traces. */
   traces: string[];
@@ -231,8 +225,6 @@ export interface MatrixModel {
   requirements: RequirementCoverage[];
   findings: MatrixFinding[];
   manifest: MatrixManifest;
-  /** Perf verdict (#569) — present when qa/perf.yaml sets a budget AND the run produced durations. */
-  perf?: import('../../harness/perf').PerfVerdict | null;
 }
 
 /** Complexity warning threshold (rules draft Gate I; raised for compact grouping) — echoed in the manifest. */

package/src/exporters/playwright-report-parser.ts

@@ -54,7 +54,6 @@ interface PlaywrightSpec {
     projectId?: string;
     results?: Array<{
       status?: string;           // 'passed' | 'failed' | 'timedOut' | 'skipped' | 'interrupted'
-      duration?: number;         // ms — Playwright emits this on every result (#569 perf budgets)
       startTime?: string;
       error?: { message?: string; stack?: string };
       attachments?: Array<{ name?: string; path?: string; contentType?: string }>;
@@ -121,7 +120,6 @@ export function loadPlaywrightReport(reportPath: string): Map<string, Playwright
     result.set(stripped, {
       testTitle: stripped,
       status,
-      duration: typeof res?.duration === 'number' ? res.duration : undefined,
       startTime: res?.startTime,
       error: errorMsg,
       tracePath: trace,

package/src/exporters/types.ts

@@ -121,8 +121,6 @@ export interface SpecFileData {
 export interface PlaywrightResult {
   testTitle: string;
   status: 'passed' | 'failed' | 'skipped' | 'timedOut' | 'interrupted' | 'unknown';
-  /** Wall-clock run time in ms (from Playwright's results[].duration) — perf budgets (#569). */
-  duration?: number;
   startTime?: string;
   error?: string;
   tracePath?: string;

package/src/generators/test-generator/diagnostics.ts

@@ -35,17 +35,12 @@
  *             nothing for the query half to assert, so the step checks the PATH only. Not an
  *             error (the value may legitimately vary per environment), but the step asserts
  *             strictly less than it appears to
- *   SG-W015 — `see [X] card/item with {{n}}` where the value is a bare number: this compiles
- *             to a TEXT assertion (toHaveText('3')), NOT a count — an author reaching for
- *             "3 cards are shown" gets an assertion that can only pass if the card's text
- *             IS the digit. Count semantics exist on row/table ("see [T] table with {{n}}",
- *             "see [Ref] row with {{n}}") and lists ("should have count") only
  *
  * Diagnostics never block generation by themselves; `sungen generate --strict`
  * turns any collected diagnostic into a non-zero exit.
  */
 
-export type DiagnosticCode = 'SG-W001' | 'SG-W002' | 'SG-W003' | 'SG-W005' | 'SG-W006' | 'SG-W007' | 'SG-W008' | 'SG-W009' | 'SG-W010' | 'SG-W011' | 'SG-W012' | 'SG-W013' | 'SG-W014' | 'SG-W015';
+export type DiagnosticCode = 'SG-W001' | 'SG-W002' | 'SG-W003' | 'SG-W005' | 'SG-W006' | 'SG-W007' | 'SG-W008' | 'SG-W009' | 'SG-W010' | 'SG-W011' | 'SG-W012' | 'SG-W013' | 'SG-W014';
 
 export interface Diagnostic {
   code: DiagnosticCode;

package/src/generators/test-generator/patterns/index.ts

@@ -73,40 +73,34 @@ export class PatternRegistry {
    * Prefers resolver (returns template data) over generator (renders directly).
    */
   generateStep(step: ParsedStep, context: PatternContext): MappedStep | null {
-    // Matchers see only the STEP; some decisions need context (a resolver reading the
-    // test-data value, e.g. row-count's numeric check). A resolver may therefore return
-    // null to say "not mine after all" — the search continues with the next matching
-    // pattern in priority order instead of crashing or silently dropping the step.
-    const matchStep: ParsedStep = { ...step, text: maskRefsForMatching(step.text) };
-    for (const pattern of this.patterns) {
-      if (!this.matchesPattern(matchStep, pattern.matcher)) continue;
-
-      // Prefer resolver (framework-agnostic) over generator (legacy)
-      if (pattern.resolver) {
-        const resolved = pattern.resolver(step, context);
-        if (!resolved) continue; // resolver declined — fall through
-
-        // Auto-inject parent scoping if step has parentRef
-        // Skip for table-* patterns — they resolve the table name internally from step text
-        if (step.parentRef && step.parentType && !pattern.name.startsWith('table-')) {
-          resolved.data.parentLocator = PatternRegistry.resolveParentLocator(
-            step.parentRef, step.parentType, context
-          );
-        }
-
-        const code = context.templateEngine.renderStep(resolved.templateName, resolved.data);
-        return {
-          code,
-          comment: resolved.comment,
-        };
-      }
+    const pattern = this.findPattern(step);
+    if (!pattern) {
+      return null;
+    }
 
-      if (pattern.generator) {
-        return pattern.generator(step, context);
+    // Prefer resolver (framework-agnostic) over generator (legacy)
+    if (pattern.resolver) {
+      const resolved = pattern.resolver(step, context);
+
+      // Auto-inject parent scoping if step has parentRef
+      // Skip for table-* patterns — they resolve the table name internally from step text
+      if (step.parentRef && step.parentType && !pattern.name.startsWith('table-')) {
+        resolved.data.parentLocator = PatternRegistry.resolveParentLocator(
+          step.parentRef, step.parentType, context
+        );
       }
 
-      return null;
+      const code = context.templateEngine.renderStep(resolved.templateName, resolved.data);
+      return {
+        code,
+        comment: resolved.comment,
+      };
     }
+
+    if (pattern.generator) {
+      return pattern.generator(step, context);
+    }
+
     return null;
   }
 

package/src/generators/test-generator/step-mapper.ts

@@ -175,28 +175,6 @@ export class StepMapper {
       this.lastPrimaryKeyword = keyword;
     }
 
-    // SG-W015 — `see [X] card with {{n}}` where n is a bare number compiles to a TEXT
-    // assertion (toHaveText('3')), not a count. An author reaching for "3 cards are shown"
-    // gets an assertion that can only pass if the card's TEXT is the digit — a wrong test
-    // with no error. Count semantics live on row/table ("see [T] table with {{n}}",
-    // "see [Ref] row with {{n}}") and lists ("should have count"); on the container types
-    // below, a bare-number text expectation is almost never what the author meant. A text
-    // element asserting a numeric readout ("[Row Quantity] text with 4") stays silent.
-    const COUNT_MISREAD_TYPES = new Set(['card', 'item', 'items', 'cards', 'option', 'options']);
-    if (this.diagnostics && step.dataRef && step.selectorRef &&
-        COUNT_MISREAD_TYPES.has((step.elementType || '').toLowerCase()) &&
-        /\bsees?\s+\[/.test(step.text) && /\bwith\b/.test(step.text)) {
-      // peekBaseValue reads the literal from test-data even in runtime-data mode,
-      // where resolveData returns an opaque marker (never numeric-looking).
-      let resolvedValue: string | undefined;
-      try { resolvedValue = this.dataResolver.peekBaseValue(step.dataRef, this.featureName); } catch { /* unresolved at compile time → cannot judge */ }
-      if (resolvedValue !== undefined && /^\d+$/.test(resolvedValue.trim())) {
-        this.diagnostics.add('SG-W015',
-          `\`see [${step.selectorRef}] ${step.elementType} with {{${step.dataRef}}}\` compiles to a TEXT assertion toHaveText('${resolvedValue.trim()}'), NOT a count of ${step.elementType}s. For a count use \`see [Ref] row with {{n}}\` / \`see [T] table with {{n}}\` (or a list "should have count"); if '${resolvedValue.trim()}' really is the ${step.elementType}'s text, assert it on a text/label element instead`,
-          { feature: this.featureName, step: step.text });
-      }
-    }
-
     // SG-W007 — a two-step-form browser-alert handler (`click [OK] alert`) written AFTER
     // an action step registers its page.once('dialog') listener too late: a dialog
     // triggered by the earlier action auto-dismisses with no error (issue #472).

package/src/harness/audit.ts

@@ -10,11 +10,10 @@ import * as path from 'path';
 import * as fs from 'fs';
 import { loadUnitScenarios, readUnitFeatureText, parseViewpointOverview, ScenarioInfo, ViewpointEntry } from './parse';
 import { checkViewpointBaseline, ViewpointBaseline } from './viewpoint-baseline';
-import { flowQuality, statefulDepthFor, FlowQualityResult } from './flow-contract';
 import { featureFilesFor } from './unit-paths';
 import {
   loadCatalog, viewpointGate, assertionDepth, dataThemesFor, depthThresholdFor, coverageBalance, duplicateClusters, traceability, claimProof, taxonomyLint,
-  automatableManual, flowCoveredThemes, flowRegressionDepth, oracleStrength, declaredPageType,
+  automatableManual, flowCoveredThemes, flowRegressionDepth, oracleStrength,
   GateResult, DepthResult, BalanceResult, DuplicateResult, TraceResult, ClaimProofResult, TaxonomyResult, Catalog, AutomatableManualResult, FlowDepthResult, OracleStrengthResult,
 } from './sensors';
 import { loadFlowScenarios } from './flow-check';
@@ -49,7 +48,6 @@ export interface AuditReport {
   oracle: OracleStrengthResult; // H4 — facet-oracle strength (weak name-substring vs title/detail/API/DB)
   ledger: LedgerResult;         // atomic viewpoint-item coverage (per-bullet status)
   viewpointBaseline: ViewpointBaseline; // is the yardstick still the accepted one? (#557)
-  flowQuality?: FlowQualityResult;      // #569 — flow contract verification (flows only)
   calibration: {                // #8 — multi-axis score so a high overall can't hide a weak axis
     axes: Record<string, number>;
     weakest: { axis: string; value: number };
@@ -156,28 +154,15 @@ export function runAudit(screenDir: string, screenName: string): AuditReport {
   // back to the in-core functions if the capability provides none.
   // A capability gate may need project context (the API gate resolves endpoint methods from the
   // catalog) + the focus depth threshold (so it scores depth with the SAME bar as the UI gate).
+  const capGate = scoringCap?.gateProvider as
+    ((i: { scenarios: ScenarioInfo[]; viewpoints: ViewpointEntry[]; catalog: Catalog; focus: typeof intent.focus; cwd: string; screenName: string; threshold: number; businessCriticalMethods?: string[] }) => { gate: GateResult; depth: DepthResult }) | undefined;
+  const provided = capGate?.({ scenarios, viewpoints, catalog, focus: intent.focus, cwd: projectRootFromScreenDir(screenDir), screenName: catalogScreenName, threshold: depthThresholdFor(intent.focus), businessCriticalMethods: intent.businessCriticalMethods });
   // Is the declaration this suite is measured against still the one a human accepted?
   const viewpointBaseline = checkViewpointBaseline(
     projectRootFromScreenDir(screenDir), catalogScreenName, viewpointPath);
   const viewpointText = fs.existsSync(viewpointPath) ? readTextFile(viewpointPath) : '';
-  // #569 — a flow WITH a contract is measured against ITS declaration, not a screen's.
-  // Keyword page-type detection force-fit flows into screen checklists (a registration
-  // flow judged as a `form` screen scored coverage 0%), so for contract-holding flows the
-  // page-type machinery only applies when the project DECLARED a page type explicitly.
-  // The stripped catalog feeds the CAPABILITY gate too — driver-ui provides the gate on
-  // real projects, and handing it the unstripped catalog re-introduced the force-fit.
-  const isUiFlowUnit = /^flows\//.test(catalogScreenName);
-  const flowQ: FlowQualityResult = isUiFlowUnit
-    ? flowQuality(screenDir, scenarios)
-    : { hasContract: false, errors: [], outcomeProven: false, outcomeManualOnly: false, offGoal: [], offGoalRatio: 0, offGoalCategories: [], phases: [], phaseRatio: 1, handoffs: { total: 0, asserted: 0, ratio: 1 } };
-  const gateCatalog = (flowQ.hasContract && !declaredPageType(viewpointText, catalog))
-    ? { ...catalog, page_types: {} } as Catalog
-    : catalog;
-  const capGate = scoringCap?.gateProvider as
-    ((i: { scenarios: ScenarioInfo[]; viewpoints: ViewpointEntry[]; catalog: Catalog; focus: typeof intent.focus; cwd: string; screenName: string; threshold: number; businessCriticalMethods?: string[] }) => { gate: GateResult; depth: DepthResult }) | undefined;
-  const provided = capGate?.({ scenarios, viewpoints, catalog: gateCatalog, focus: intent.focus, cwd: projectRootFromScreenDir(screenDir), screenName: catalogScreenName, threshold: depthThresholdFor(intent.focus), businessCriticalMethods: intent.businessCriticalMethods });
-  const gate = provided?.gate ?? viewpointGate(scenarios, viewpoints, gateCatalog, platform === 'mobile', viewpointText);
-  const depth = provided?.depth ?? assertionDepth(scenarios, dataThemesFor(gateCatalog, gate.pageType), intent.focus);
+  const gate = provided?.gate ?? viewpointGate(scenarios, viewpoints, catalog, platform === 'mobile', viewpointText);
+  const depth = provided?.depth ?? assertionDepth(scenarios, dataThemesFor(catalog, gate.pageType), intent.focus);
 
   // TQ-4 — deferral-aware coverage credit: an inherently cross-screen theme (cart / detail /
   // filter correctness) belongs in a FLOW, not on the screen. When a flow deeply covers a screen
@@ -215,15 +200,8 @@ export function runAudit(screenDir: string, screenName: string): AuditReport {
   // collection, the regression dimensions (count/quantity proof · teardown · multi-source) cap the
   // businessDepth headroom: it can reach 1.0 only when all three are exercised, so a present-but-
   // shallow flow can't claim a perfect score (floor 0.5 — assertion depth still dominates).
-  const isUiFlow = isUiFlowUnit;
-  let flowDepth = isUiFlow ? flowRegressionDepth(scenarios) : { stateful: false, countProof: false, teardown: false, multiSource: false, ratio: 1, missing: [] } as FlowDepthResult;
-  // Contract-declared statefulness generalizes the cart-hardcoded vocabulary: a flow that
-  // mutates ANY named collection (order, application, submission …) gets the same three
-  // regression dims. The legacy cart detection stays for contract-less flows (snapshots).
-  if (isUiFlow && !flowDepth.stateful && flowQ.contract?.stateful) {
-    const d = statefulDepthFor(flowQ.contract.stateful, scenarios);
-    flowDepth = { stateful: true, countProof: d.countProof, teardown: d.teardown, multiSource: d.multiSource, ratio: d.ratio, missing: d.missing };
-  }
+  const isUiFlow = /^flows\//.test(catalogScreenName);
+  const flowDepth = isUiFlow ? flowRegressionDepth(scenarios) : { stateful: false, countProof: false, teardown: false, multiSource: false, ratio: 1, missing: [] } as FlowDepthResult;
   const FLOW_DEPTH_FLOOR = 0.5;
   // H4 — oracle strength: a weak facet oracle (name-substring "proves" category/brand membership)
   // caps businessDepth the same way (floor 0.5). A suite with no facet claim, or a strong oracle, is
@@ -233,14 +211,10 @@ export function runAudit(screenDir: string, screenName: string): AuditReport {
 
   // Sub-scores
   const coverage = gate.coverageRatio;
-  // #569 — handoff integrity: a flow's depth is also whether each cross-screen transition
-  // is followed by an assertion on the carried state (same floor discipline as the other caps).
-  const HANDOFF_FLOOR = 0.5;
   const businessDepth = Math.min(
     depth.bcDepthRatio,
     flowDepth.stateful ? FLOW_DEPTH_FLOOR + (1 - FLOW_DEPTH_FLOOR) * flowDepth.ratio : 1,
     oracle.weak.length ? ORACLE_FLOOR + (1 - ORACLE_FLOOR) * oracle.ratio : 1,
-    flowQ.hasContract && flowQ.handoffs.total > 0 ? HANDOFF_FLOOR + (1 - HANDOFF_FLOOR) * flowQ.handoffs.ratio : 1,
   );
   // When the taxonomy drifted (most scenarios unclassified), the balance axis is unreliable — cap it
   // at 0.5 instead of awarding a vacuous 1.0 so a stale taxonomy fails loudly, not silently (H1).
@@ -272,15 +246,8 @@ export function runAudit(screenDir: string, screenName: string): AuditReport {
   //     what a create-test run did, silently dropping the performance viewpoint on
   //     the way. While the declaration is unconfirmed, neither axis is evidence.
   const viewpointMoved = viewpointBaseline.status === 'changed';
-  //  4. A flow with a contract is measured on ITS coverage: journey phases (HP/ER/EH
-  //     automated) — not a screen's page-type themes. Same weight, same criticality,
-  //     so a flow can finally EARN the coverage axis instead of always losing it to
-  //     PAGE-TYPE-UNDETERMINED (or worse, being judged against a form checklist).
-  const flowScored = flowQ.hasContract;
   const axisDefs: Array<{ key: string; value: number; weight: number; applicable: boolean; critical: boolean }> = [
-    flowScored
-      ? { key: 'flowCoverage', value: flowQ.phaseRatio, weight: 0.22, applicable: true, critical: true }
-      : { key: 'coverage', value: coverage, weight: 0.22, applicable: !!gate.pageType && gate.themesTotal > 0, critical: true },
+    { key: 'coverage', value: coverage, weight: 0.22, applicable: !!gate.pageType && gate.themesTotal > 0, critical: true },
     { key: 'specFR', value: specRatio, weight: 0.15, applicable: spec.hasSpec && spec.frTotal > 0, critical: true },
     { key: 'atomicLedger', value: ledger.ratio, weight: 0.13, applicable: ledger.hasViewpoint && ledger.total > 0 && !viewpointMoved, critical: true },
     { key: 'businessDepth', value: businessDepth, weight: 0.20, applicable: true, critical: true },
@@ -330,40 +297,9 @@ export function runAudit(screenDir: string, screenName: string): AuditReport {
     ].filter(Boolean).join(', ');
     findings.push(`VIEWPOINT-BASELINE-CHANGED: test-viewpoint.md no longer declares what it did when this unit was last accepted — ${moved || 'the declared ids were reordered or replaced'}. atomicLedger + traceability measure the suite AGAINST this file, so they are excluded from the score until the change is confirmed: a generator that rewrites the declaration scores both 100% by construction, and a viewpoint dropped from the file stops being missing from anything. Review the diff (a removed id means that coverage is now unclaimed), then run \`sungen audit --screen ${screenName} --accept-viewpoint\`.`);
   }
-  // On a contract flow the page type is INTENTIONALLY not applied — the contract is the checklist.
-  if (gate.pageTypeSource === 'undetermined' && !flowScored) {
+  if (gate.pageTypeSource === 'undetermined') {
     findings.push(`PAGE-TYPE-UNDETERMINED: no page type fits this screen with enough confidence (best ${gate.pageTypeEvidence?.hits ?? 0} keyword hit(s)), so NO critical themes were demanded and the coverage axis is excluded from the score — declare it in test-viewpoint.md (\`page-type: <id>\`) to have the theme checklist applied.`);
   }
-  // ---- #569: flow-quality findings -----------------------------------------
-  if (isUiFlow && !flowQ.hasContract) {
-    const why = flowQ.errors.length ? ` (${flowQ.errors.join('; ')})` : '';
-    findings.push(`FLOW-CONTRACT-MISSING: this flow declares no contract${why} — without \`requirements/flow-contract.yaml\` (goal · trigger · outcome.screen · phases) the flow cannot be scored on ITS coverage (journey phases) and falls back to screen machinery. Author it with the boundary checklist: one business goal, a clear trigger, ONE observable outcome valuable to the actor, name = "Verb + outcome"; split anything that forms an independently valuable flow.`);
-  }
-  if (flowQ.hasContract) {
-    if (!flowQ.outcomeProven) {
-      findings.push(flowQ.outcomeManualOnly
-        ? `FLOW-OUTCOME-UNPROVEN: the goal "${flowQ.contract!.goal}" is covered only by @manual scenarios — the flow's own outcome (\`${flowQ.contract!.outcome.screen}\`) has no AUTOMATED data assertion, so a regression run cannot prove the journey still works. Automate the happy-path proof first; it is the reason this flow exists.`
-        : `FLOW-OUTCOME-UNPROVEN: no scenario asserts data on the outcome screen \`${flowQ.contract!.outcome.screen}\` — the flow never proves its own goal ("${flowQ.contract!.goal}"). Add the happy-path scenario that ends in the declared final assertion.`);
-    }
-    if (flowQ.offGoalRatio > 0.35) {
-      findings.push(`FLOW-SCOPE-CREEP: ${flowQ.offGoal.length}/${scenarios.length} scenarios never touch the outcome screen \`${flowQ.contract!.outcome.screen}\` and are not guards/error-recovery (categories: ${flowQ.offGoalCategories.join(', ')}) — evidence of a SECOND business goal in this flow. Split them into their own flow (one flow = one goal, one observable outcome).`);
-    }
-    for (const ph of flowQ.phases.filter((p) => !p.covered || !p.automated)) {
-      findings.push(`FLOW-PHASE-${ph.phase}-MISSING: journey phase ${ph.phase} (${ph.phase === 'HP' ? 'happy path proving the outcome' : ph.phase === 'ER' ? 'error recovery — validation must not trap the journey' : 'guards — direct access / back / refresh'}) is ${ph.covered ? 'covered only by @manual' : 'not covered'} → it does not count toward flowCoverage until automated.`);
-    }
-    if (flowQ.handoffs.total > 0 && flowQ.handoffs.ratio < 0.7) {
-      findings.push(`FLOW-HANDOFF-SHALLOW: only ${flowQ.handoffs.asserted}/${flowQ.handoffs.total} cross-screen transitions are followed by an assertion in the new screen — carried state (the thing a flow exists to verify) is unproven at the handoff. businessDepth is capped until asserted.`);
-    }
-  }
-  if (isUiFlow) {
-    // Flows defer more than screens, so an UNCODED @manual is costlier here: the capability
-    // planner cannot route it to the driver that would automate it (password-reset: 11/22
-    // manual, most of them M5-inbox — invisible while tagged bare).
-    const bareManual = (featureText.match(/@manual(?![:\w])/g) || []).length;
-    if (bareManual > 0) {
-      findings.push(`MANUAL-CODE-MISSING: ${bareManual} @manual tag(s) carry no reason code — tag @manual:M1..M9 (M5 = external dependency e.g. inbox, M3 = mock/network, M7 = environment) so the capability planner can route what a driver would automate.`);
-    }
-  }
   for (const c of flowCredits) {
     findings.push(`COVERED-VIA-FLOW: critical theme "${c.theme}" is not covered on this screen but is deeply covered by flow "${c.flow}" — cross-screen depth correctly owned by the flow, so it is credited (not a screen gap). Verify with \`sungen flow-check\`.`);
   }
@@ -440,13 +376,6 @@ export function runAudit(screenDir: string, screenName: string): AuditReport {
     const shown = spec.inferredOnly.slice(0, 8).join(', ');
     findings.push(`SPEC-TRACE-IMPLICIT: ${spec.inferredOnly.length} requirement(s) are covered by keyword inference only (${shown}${spec.inferredOnly.length > 8 ? ', …' : ''}) — add \`@spec:<id>\` to the scenario that proves each so the trace survives edits and delivery can report it.`);
   }
-  // Safety net for the parser (#572): the feature CITES requirement ids the spec parser
-  // could not find a declaration for. Silence here would repeat the original failure —
-  // "nothing found" quietly excluding the specFR axis while the evidence sits in the file.
-  if (spec.hasSpec && spec.frTotal === 0 && /@spec:/i.test(featureText)) {
-    const cited = [...new Set([...featureText.matchAll(/@spec:([A-Za-z0-9.-]+)/g)].map((m) => m[1].toUpperCase()))];
-    findings.push(`SPEC-IDS-UNPARSED: the feature cites ${cited.length} requirement id(s) via @spec: (${cited.slice(0, 5).join(', ')}${cited.length > 5 ? ', …' : ''}) but requirements/spec.md yielded ZERO parseable requirement declarations — the spec likely declares ids in a scheme the parser cannot read. Declare each requirement on its own line as \`<ID>: <clause>\` (bullet/bold/backtick all fine; the id must end in a number) so the MUST-coverage gate and the specFR axis can run.`);
-  }
   if (taxonomyMismatch) {
     findings.push(`VP-TAXONOMY-MISMATCH: only ${(trace.mappedRatio * 100).toFixed(0)}% of scenarios use the viewpoint IDs declared in test-viewpoint.md — scenarios invented a generic VP-<CAT> scheme. Re-tag to the project's viewpoint IDs so the coverage matrix is accurate.`);
   }
@@ -500,18 +429,6 @@ export function runAudit(screenDir: string, screenName: string): AuditReport {
   for (const d of ownership.duplicates.slice(0, 6)) {
     findings.push(`DUPLICATE-FLOW-OWNERSHIP: "${d.scenario}" has the same shape as a scenario in flow "${d.flow}" — keep one owner (screen-local vs flow); the other should only reference/set up.`);
   }
-  // #580 — the flow→screen half of the ownership rule (ISTQB system-test design:
-  // representative, not exhaustive). A flow's cases take ONE representative input per
-  // branch — one valid value on the main flow, one error value that triggers the
-  // exception flow. An EP/data-family cluster (≥3 same-shape scenarios differing only
-  // in data) inside a FLOW is the screen's validation matrix being re-run where it
-  // does not belong: the screen already proved the field rejects each class; the flow
-  // only has to prove the journey survives one of them. Advisory — never a gate.
-  if (isUiFlow) {
-    for (const c of duplicates.clusters.filter((x) => !x.sameDataLikely && x.scenarios.length >= 3).slice(0, 4)) {
-      findings.push(`FLOW-EXHAUSTIVE-INPUTS: ${c.scenarios.length} same-shape scenarios differ only in data (${c.scenarios.slice(0, 3).join(' | ')}${c.scenarios.length > 3 ? ' …' : ''}) — a system-test flow uses REPRESENTATIVE inputs (one valid + one that triggers the exception branch); the full EP/boundary matrix belongs to the owning screen (or a @cases dataset there). Keep one representative per branch here.`);
-    }
-  }
   for (const u of unsourced.slice(0, 6)) {
     findings.push(`UNSOURCEABLE-SCENARIO: "${u}" doesn't trace to any FR / viewpoint item — link it to a source, or tag it @exploration (not part of the official suite).`);
   }
@@ -568,7 +485,6 @@ export function runAudit(screenDir: string, screenName: string): AuditReport {
     scenarioCount: scenarios.length,
     gate, depth, claim, taxonomy, balance, duplicates, trace, spec,
     taxonomyMismatch, downstream, manualOracle: manualOracleResult, automatableManual: autoManual, flowDepth, oracle, ledger, viewpointBaseline, calibration,
-    ...(isUiFlow ? { flowQuality: flowQ } : {}),
     score: {
       overall: Math.round(overall * 10) / 10,
       coverage: Math.round(coverage * 100) / 100,

package/src/harness/flow-plan.ts

@@ -71,9 +71,6 @@ export interface FlowPlan {
   legs: LegPlan[];
   byReason: Record<string, number>;
   capabilityManual: number;
-  /** @manual whose reason is "cross-screen → automate via flow" (class XS) — inside a flow this
-   *  usually means the scenario should simply BE automated here. (#569) */
-  crossScreenManual: number;
   judgmentManual: number;
   contracts: Contract[];
   readiness: 'ready' | 'not-ready';
@@ -90,18 +87,14 @@ export function buildFlowPlan(cwd: string, flow: string): FlowPlan {
   // Legs = distinct screen namespaces.
   const legMap = new Map<string, { scenarios: Set<string>; refs: Set<string>; automated: boolean }>();
   const byReason: Record<string, number> = {};
-  let capabilityManual = 0, judgmentManual = 0, crossScreenManual = 0;
+  let capabilityManual = 0, judgmentManual = 0;
 
   for (const sc of scenarios) {
     if (sc.manual) {
       const { code } = inferReasonCode(sc.tags, sc.reason);
       byReason[code] = (byReason[code] || 0) + 1;
       const cls = MANUAL_REASONS[code]?.cls;
-      // XS ("cross-screen → automate via flow") is a THIRD class; it used to be silently
-      // dropped from both counters, understating the plan's manual load. (#569)
-      if (cls === 'capability') capabilityManual++;
-      else if (cls === 'keep') judgmentManual++;
-      else if (cls === 'flow') crossScreenManual++;
+      if (cls === 'capability') capabilityManual++; else if (cls === 'keep') judgmentManual++;
     }
     for (const r of sc.refs) {
       const leg = r.screen.toLowerCase();
@@ -140,5 +133,5 @@ export function buildFlowPlan(cwd: string, flow: string): FlowPlan {
   }
   if (readiness === 'ready') plan.unshift('Selectors present for every automated leg — ready to compile + run.');
 
-  return { flow, total: scenarios.length, legs, byReason, capabilityManual, judgmentManual, crossScreenManual, contracts, readiness, missingLegs, plan };
+  return { flow, total: scenarios.length, legs, byReason, capabilityManual, judgmentManual, contracts, readiness, missingLegs, plan };
 }

package/src/harness/parse.ts

@@ -38,8 +38,6 @@ export interface ScenarioInfo {
   requiresCaps?: string[];    // @requires:<cap> — automation-ready but needs an opt-in driver (TQ-11)
   deferredToFlow?: boolean;   // @deferred:flow — owned by a flow, not automated on this screen (H6)
   ownedByFlow?: string;       // @owned-by:<flow> — the flow that owns this deferred scenario (H6)
-  /** Ordered steps with their resolved bucket (And/But inherit) — flow handoff analysis (#569). */
-  steps?: Array<{ bucket: 'given' | 'when' | 'then'; text: string }>;
 }
 
 /** Format-tolerant: is this token an ID (project's scheme), not a prose word?
@@ -93,29 +91,6 @@ export function parseViewpointOverview(filePath: string): ViewpointEntry[] {
     }
   }
 
-  // 1b) Flow-style declarations (#569). Flow viewpoint files commonly declare per-item
-  // ids at the END of a bullet ("… → **FL-HP-001**") under phase section headers
-  // ("## FL-HP — Happy Path"). Neither matched the table/group passes, so every flow
-  // audit collapsed to taxonomy=0% / traceability n-a — scenarios correctly tagged
-  // FL-HP-001 were reported as unmapped. Both forms are additive here.
-  for (const raw of lines) {
-    const line = raw.trim();
-    const section = line.match(/^##\s+([A-Z]{2,}(?:-[A-Z0-9]{2,})*)\s+[—–-]\s*(.*)$/);
-    if (section && isViewpointId(section[1] + '-0')) {
-      const id = section[1].toUpperCase();
-      if (!entries.has(id)) entries.set(id, { id, priority: 'Unknown', reason: section[2] ?? '' });
-    }
-    if (/^[-*+]\s/.test(line)) {
-      const arrow = line.match(/(?:→|->)\s*\*{0,2}([A-Z][A-Z0-9]*(?:-[A-Z0-9]+)*-\d+[a-zA-Z]?)\*{0,2}\s*$/);
-      if (arrow) {
-        const id = arrow[1].toUpperCase();
-        if (!entries.has(id)) {
-          entries.set(id, { id, priority: 'Unknown', reason: line.replace(/\s*(?:→|->).*$/, '').replace(/^[-*+]\s+/, '') });
-        }
-      }
-    }
-  }
-
   // 2) Viewpoint Grouping: ### Required / ### Recommended / ### Optional → bullet list
   let group: ViewpointEntry['group'] | undefined;
   for (const raw of lines) {
@@ -169,9 +144,7 @@ function classifyScenario(sc: ParsedScenario): ScenarioInfo {
   // Category is everything between `VP-` and the final `-<sequence>` — INCLUDING hyphens, so
   // compound categories (VP-LIST-DISPLAY-01, VP-ADD-TO-CART-03, VP-PRODUCT-DISCOVERY-02) parse,
   // not just single-word ones. A single-word category (VP-CART-001) still works. (H1)
-  // Flows use journey-phase ids (FL-HP-001 / FL-ER-002) — the VP- anchor rejected them, so every
-  // flow scenario had NO category and the whole suite bucketed `other` (taxonomy=0%, #569).
-  const codeMatch = sc.name.match(/\b(?:VP|FL)-([A-Z]+(?:-[A-Z]+)*)-\d+/i);
+  const codeMatch = sc.name.match(/\bVP-([A-Z]+(?:-[A-Z]+)*)-\d+/i);
   const vpCode = codeMatch ? codeMatch[0].toUpperCase() : undefined;
   const category = codeMatch ? codeMatch[1].toUpperCase() : undefined;
   // Project-scheme ID: the leading token of the title (VP0-001 / MS-HP-001 / VP-LIST-001).
@@ -185,14 +158,12 @@ function classifyScenario(sc: ParsedScenario): ScenarioInfo {
   const skeletonParts: string[] = [];
   const textParts: string[] = [sc.name];
   const stepTextParts: string[] = [];
-  const orderedSteps: Array<{ bucket: 'given' | 'when' | 'then'; text: string }> = [];
 
   for (const step of sc.steps as ParsedStep[]) {
     const kw = step.keyword.trim();
     if (kw === 'Given' || kw === 'When' || kw === 'Then') last = kw;
     textParts.push(step.text);
     stepTextParts.push(step.text);
-    orderedSteps.push({ bucket: (kw === 'And' || kw === 'But' ? last : kw).toLowerCase() as 'given' | 'when' | 'then', text: step.text });
     // normalized skeleton: keep [refs] (distinct targets = distinct tests),
     // but neutralize {{vars}} and quoted values so EP/data families collapse.
     const skel = step.text
@@ -228,7 +199,6 @@ function classifyScenario(sc: ParsedScenario): ScenarioInfo {
     stepSkeleton: skeletonParts.join(' | '),
     haystack: textParts.join(' ').toLowerCase(),
     stepsText: stepTextParts.join(' ').toLowerCase(),
-    steps: orderedSteps,
     vpId,
     casesDataset,
     queryRefs: queryRefs.size ? [...queryRefs] : undefined,

package/src/harness/sensors.ts

@@ -33,22 +33,10 @@ const BUCKET_ORDER: Array<[string, string[]]> = [
 ];
 const BUCKETS: Record<string, string[]> = Object.fromEntries(BUCKET_ORDER);
 
-// Flow journey-phase categories (FL-HP-001, FL-ER-002 …). Matched on exact SEGMENTS,
-// never by containment — 'SHOP'.includes('HP') is true, which is exactly the kind of
-// false hit substring matching would produce for two-letter phase tokens. (#569)
-const PHASE_BUCKETS: Record<string, string> = {
-  HP: 'business-core',          // happy path = the business goal itself
-  ER: 'validation-security',    // error recovery (validation must not trap the journey)
-  EH: 'validation-security',    // guards & leakage (direct access, back, refresh)
-};
-
 /** Classify a VP category into a balance bucket by keyword containment + precedence (H1). */
 export function bucketForCategory(category: string | undefined): string {
   const cat = (category || '').toUpperCase();
   if (!cat) return 'other';
-  for (const seg of cat.split('-')) {
-    if (PHASE_BUCKETS[seg]) return PHASE_BUCKETS[seg];
-  }
   for (const [bucket, kws] of BUCKET_ORDER) {
     if (kws.some((k) => cat.includes(k))) return bucket;
   }
@@ -363,7 +351,7 @@ export function flowRegressionDepth(scenarios: ScenarioInfo[]): FlowDepthResult
   // 1. Count/quantity proof — a row count or item quantity, not just presence of a row.
   const countProof = any(/\b(quantity|qty|two (?:rows|lines|cart)|row count|count column|number of items|one[_ ]row|two[_ ]rows|qty[_ ])/i);
   // 2. Teardown — removes the item and verifies the empty/zero state (the inverse operation).
-  const teardown = any(/\b(remove|delete|clear)(?:s|d|ed|ing)?\b/i) && any(/\b(empty|emptied|no items|zero|removed|cleared|0 items)\b/i);
+  const teardown = any(/\b(remove|delete|clear)\b/i) && any(/\b(empty|no items|zero|removed|0 items)\b/i);
   // 3. Multi-source — the cart is fed from >1 source (the main list AND a recommended/related rail).
   const multiSource = any(/\b(recommended|related|you may also|suggest)\b/i) && addsToCart;
 

package/src/harness/spec-coverage.ts

@@ -66,43 +66,21 @@ export function parseSpecClauses(specPath: string): { frs: FrClause[]; valRows:
   if (!fs.existsSync(specPath)) return { frs: [], valRows: [] };
   const lines = readTextFile(specPath).split('\n');
 
-  // Requirement ids follow the PROJECT's scheme, not ours (#572): `**FR-1**:` is one
-  // convention among many — a real spec declared ~30 MUST clauses as `` `REQ-SRCH-001`: ``
-  // and the FR-locked pattern returned zero, so the MUST-coverage gate never ran and the
-  // specFR axis was excluded "for lack of evidence" that was sitting right there. Same
-  // silent-failure class as the CRLF parsers, same id-scheme-tolerance lesson as delivery.
-  //
-  // A declaration is: line-leading (optionally bulleted / bold / backticked) `<ID>:` where
-  // the id ends in a number. Table rows are EXCLUDED — traceability tables cite requirement
-  // ids without declaring them — and so are prefixes that are never requirements
-  // (test cases, viewpoints, known-defect records, data-factory checks, flows, delivery items).
-  const NON_REQUIREMENT_PREFIX = /^(TC|VP|KD|CHK|FL|DI)-/i;
   const frs: FrClause[] = [];
-  const seen = new Set<string>();
   for (const line of lines) {
-    if (/^\s*\|/.test(line)) continue;   // table row = citation, not declaration
-    const m = line.match(/^\s*(?:[-*+]\s+)?[*_`]*([A-Z][A-Z0-9]*(?:-[A-Z0-9]+)*-\d+[a-zA-Z]?)[*_`]*\s*:\s*(.+)$/);
-    if (!m || NON_REQUIREMENT_PREFIX.test(m[1])) continue;
-    const id = m[1].toUpperCase();
-    if (seen.has(id)) continue;          // first declaration wins
-    seen.add(id);
-    frs.push({ id, text: m[2].replace(/\*\*/g, '').trim(), modality: modalityOf(m[2]) });
+    const m = line.match(/\*\*FR-(\d+)\*\*\s*:\s*(.+)$/);
+    if (m) frs.push({ id: `FR-${m[1]}`, text: m[2].replace(/\*\*/g, '').trim(), modality: modalityOf(m[2]) });
   }
 
   // Validation Rules table: a row carries a Constraint, a Trigger cell, and (often) a code.
-  // A "Trigger" column alone is NOT enough to claim the table (#578): a screen-STATES table
-  // ("State ID | Trigger | URL/heading oracle | …") uses Trigger for the user action that
-  // enters the state, and reading it as validation rows invented three gate-relevant
-  // TRIGGER-UNCOVERED gaps on a real spec. The table must also name a Constraint/Rule/
-  // Validation column — the thing a validation row is ABOUT.
   const valRows: ValRow[] = [];
   let cTrigger = -1, cConstraint = -1, cCode = -1, inTable = false;
   for (const raw of lines) {
     const line = raw.trim();
-    if (line.startsWith('|') && /\btrigger\b/i.test(line) && /\b(constraint|rule|validation)\b/i.test(line) && cTrigger < 0) {
+    if (line.startsWith('|') && /\btrigger\b/i.test(line) && cTrigger < 0) {
       const cells = line.split('|').map((c) => c.trim());
       cTrigger = cells.findIndex((c) => /^trigger$/i.test(c));
-      cConstraint = cells.findIndex((c) => /constraint|rule|validation/i.test(c));
+      cConstraint = cells.findIndex((c) => /constraint/i.test(c));
       cCode = cells.findIndex((c) => /code/i.test(c));
       inTable = cTrigger >= 0;
       continue;