@sun-asterisk/sungen 3.2.21-beta.1 → 3.2.21

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.`);
   }
@@ -556,7 +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;

package/src/orchestrator/templates/ai-src/commands/add-flow.md

@@ -86,52 +86,15 @@ qa/flows/${input:flow}/
     └── ui/                        # Screenshots, mockups
 ```
 
-### 1a. Define the flow's BOUNDARY, then its screens
+### 1a. Identify the screens in the flow
 
-A flow is the **smallest complete business action chain**: one clear trigger ending in ONE
-observable, valuable outcome. Before asking for screens, walk this checklist with the user —
-if 1, 3 or 8 fails, propose SPLITTING into separate flows:
-
-1. Exactly **one business goal**? (cart correctness + category filtering = two flows)
-2. A clear **trigger** and precondition?
-3. **One observable final outcome**? (a final assertion you can write in one sentence)
-4. Is that outcome **valuable to the actor**? (an order placed, a password reset — not "a page rendered")
-5. Is **every step necessary** for that outcome?
-6. Are all steps at the **same business abstraction**?
-7. Are optional/error branches **phases of this goal** (ER/EH), not new goals?
-8. Does **no segment** form an independently valuable flow on its own?
-9. Can you write **a single clear final assertion**?
-10. Can you name it "**Verb + outcome**"? (`place-order`, `reset-password` — not `cart-and-filter`)
-
-Then ask: "Which screens does this flow visit, in order? (e.g., login → dashboard → award-form → confirmation)"
+Ask the user: "Which screens does this flow visit, in order? (e.g., login → dashboard → award-form → confirmation)"
 
 Record the screen list — you will need it for:
 - Filling `spec.md` (Step 3)
 - Suggesting `[Screen:Element]` namespace prefixes
 - Capturing visuals per screen (Step 2)
 
-### 1b. Author the Flow Contract (`requirements/flow-contract.yaml`)
-
-Write the answers down as the flow's contract — `sungen audit` scores the flow **against it**
-(the `flowCoverage` axis: HP/ER/EH journey phases; `FLOW-OUTCOME-UNPROVEN` when no automated
-scenario asserts data on the outcome screen; `FLOW-SCOPE-CREEP` when scenarios never touch it):
-
-```yaml
-goal: "Place an order for a product added from home"   # Verb + outcome
-actor: user
-trigger: "Add a product to the cart from the home featured list"
-precondition: "A registered account; an empty cart"
-outcome:
-  screen: checkout            # the [Screen:...] namespace carrying the final proof
-  assertion: "The confirmation shows the order number and the paid total"
-value: "The customer has paid; the shop has a new order"
-phases: [HP, ER, EH]          # journey phases (default); add UI only if the flow owns UI states
-stateful: cart                # the mutated collection, if any — enables regression-depth dims
-```
-
-**A filled contract is an INPUT to generation — never an output.** Like `test-viewpoint.md`,
-generation must not rewrite it to match what was generated; disagree → propose the diff and ask.
-
 ### 2. Capture visual source
 
 **Mobile path** (`platform: mobile`):
@@ -224,8 +187,7 @@ If user picks `/sungen:create-test`, **you MUST use the Skill tool** to invoke i
 - Test data namespaced by phase: `login.email`, `submission.nominee`
 - `@flow` tag required at feature level
 - `Background:` should only contain the starting navigation — the URL path (web) or the `--reach` nav recipe (mobile)
-- Each scenario = one phase of the journey; ids are `FL-<PHASE>-NNN` (`HP`/`ER`/`EH`, optional `UI`)
-- One flow = ONE business goal with ONE observable outcome (`requirements/flow-contract.yaml`) — a segment with its own value is its own flow
+- Each scenario = one phase of the journey
 {{#cap parallel-subagents}}
 - Mobile flows are tagged `@platform:mobile` and run via `/sungen:run-test <flow>` (WebdriverIO, not Playwright)
 {{/cap}}

package/src/orchestrator/templates/ai-src/commands/create-test.md

@@ -6,16 +6,6 @@ order: 20
 claude-tools: "Read, Grep, Bash, Glob, Write, AskUserQuestion, Skill, mcp__playwright__browser_navigate, mcp__playwright__browser_snapshot, mcp__playwright__browser_take_screenshot"
 copilot-tools: "[vscode, execute, read, agent, edit, search, web, browser, todo, 'playwright/*']"
 codex-trigger: "Run when the user asks to CREATE, generate, write, or author test cases / a .feature file for a screen or flow. Step 2 (after add-screen/add-flow, before run-test). Do NOT use for executing, running, or compiling existing tests."
----
-## ⛔ HARD RULE — the run's LAST action is the next-step hand-back
-
-A create-test run is NOT finished when the files are written or the audit prints. The final
-action of EVERY run — success, partial, or aborted — is the next-step hand-back
-({{#cap parallel-subagents}}an `AskUserQuestion` offering the next actions{{/cap}}{{^cap parallel-subagents}}a numbered list of next-action choices{{/cap}};
-see "Finish — always hand the next step back" at the end of this file). Ending with a prose
-summary and no choices is a broken run: the operator is left guessing. This holds no matter
-how long the generation/repair loop ran.
-
 ---
 {{#cap parallel-subagents}}
 ## ⛔ HARD RULE — No Figma MCP when PAT data exists

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

@@ -605,40 +605,22 @@ error:
 
 > **Auto-detect**: if path is `qa/flows/<name>/` → use this section. Skip Steps 1–4 above.
 
-**Read `requirements/flow-contract.yaml` FIRST — it is the flow's boundary and the yardstick
-`sungen audit` scores the flow against** (`flowCoverage` axis = journey phases HP/ER/EH automated;
-`FLOW-OUTCOME-UNPROVEN`; `FLOW-SCOPE-CREEP`). No contract yet → author it with the user via the
-boundary checklist in `add-flow` (one business goal · clear trigger · ONE observable outcome
-valuable to the actor · name = "Verb + outcome"), THEN generate. **A filled contract is an INPUT —
-never rewrite it to match your output** (same rule as `test-viewpoint.md`).
-
 | Aspect | Screen | Flow |
 |---|---|---|
-| Section focus | UI patterns per section | Journey phases toward ONE declared outcome |
+| Section focus | UI patterns per section | Journey phases across screens |
 | Selector format | `[Element]` | `[Screen:Element]` (namespaced) |
 | Test data keys | `{{variable}}` | `{{phase.variable}}` |
 | Feature tag | `@auto` / `@smoke` etc. | `@flow` (required) |
-| Scenario ids | `VP-<CATEGORY>-NNN` | `FL-<PHASE>-NNN` — phases: `HP` (happy path), `ER` (error recovery), `EH` (guards), `UI` (journey UI states, optional) |
+| Viewpoints | VP-UI/VAL/LOGIC/SEC per section | VP-LOGIC (transitions), VP-SEC (auth persistence), VP-VAL (cross-screen data) |
 
-**Scenarios to generate — every phase demanded by the contract, automated:**
+**Scenarios to generate:**
 
-| Phase | What to test | Scoring |
-|---|---|---|
-| `FL-HP` happy path | The complete journey ending in the contract's `outcome.assertion` — an AUTOMATED **data** assertion on `outcome.screen` (an order number, a summed total — not just "page visible"). This scenario is WHY the flow exists: it is the regression proof after a lib/framework upgrade. | uncovered → `flowCoverage` drops + `FLOW-OUTCOME-UNPROVEN` |
-| `FL-ER` error recovery | Invalid input mid-flow → error shown → fix → the journey still completes. Validation must not trap the journey. | uncovered → `flowCoverage` drops |
-| `FL-EH` guards | Direct URL access without the precondition · browser back · refresh · expired context — each ends in a safe observable state. | uncovered → `flowCoverage` drops |
-| Cross-screen handoff | After every screen transition, assert the CARRIED state on the new screen (the added product's name in the cart, the email echoed on the sent screen). | blind tails cap `businessDepth` (`FLOW-HANDOFF-SHALLOW`) |
-| Stateful regression (when `stateful:` declared) | Count/quantity proof · teardown (remove → empty) · multi-source add. | missing dims cap `businessDepth` (`FLOW-DEPTH`) |
-
-**Boundary discipline while generating:** every scenario must serve the contract's goal. A scenario
-that never touches `outcome.screen` and is not a guard (`EH`) or error-recovery (`ER`) belongs in a
-DIFFERENT flow — propose the split instead of writing it here (`FLOW-SCOPE-CREEP` will flag it).
-Auth persistence across transitions is part of `EH` unless the project declares it its own phase.
-
-**Manual in flows**: always `@manual:Mx` with the reason code — bare `@manual` is flagged
-(`MANUAL-CODE-MISSING`) because the capability planner cannot route it. Typical flow deferrals:
-inbox/mail oracle → `M5`, network-request count → `M3`, context expiry control → `M7`. A
-cross-screen scenario inside the flow's own goal is NOT manual — automate it here.
+| Category | What to test |
+|---|---|
+| Happy path | Complete flow end-to-end with valid data |
+| Auth persistence | Auth state maintained across screen transitions |
+| Error recovery | Invalid input mid-flow → fix → continue |
+| Cross-screen data | Data entered on screen A visible on screen B |
 
 ```gherkin
 @flow @auth:user
@@ -648,20 +630,19 @@ Feature: Award Submission Flow
     Given User is on [Login] page
 
   @high
-  Scenario: FL-HP-001 A signed-in user's nomination is submitted and confirmed
+  Scenario: User logs in successfully
     When User fill [Login:Email] field with {{login.email}}
     And User fill [Login:Password] field with {{login.password}}
     And User click [Login:Submit] button
     Then User see [Dashboard] page
-    When User click [Dashboard:Awards] link
-    And User fill [Awards:Nominee] field with {{submission.nominee}}
-    And User click [Awards:Submit] button
-    Then User see [Awards:Success Message] text with {{success_message}}
 
   @high
-  Scenario: FL-EH-001 Direct access to the award form without login redirects to login
-    When User go to [Awards] page
-    Then User see [Login] page
+  Scenario: User submits nomination
+    When User click [Dashboard:Awards] link
+    Then User see [Awards] page
+    When User fill [Awards:Nominee] field with {{submission.nominee}}
+    And User click [Awards:Submit] button
+    Then User see {{success_message}} message
 ```
 
 ```yaml

package/src/orchestrator/templates/qa-context.md

@@ -45,21 +45,8 @@ Example:
 
 ## Testing Strategy
 
-Machine-readable intent — `sungen audit` reads these keys (Intent Profile). Values here are
-live even when the surrounding text changes; an invalid value silently falls back to the default.
-
-focus: functional
-<!-- focus: functional | e-commerce | security | smoke — drives the audit's depth threshold -->
-
-risk_tier: normal
-<!-- risk_tier: high | normal | low -->
-
-To silence driver suggestions in audit findings, add a line: capability_suggestions with value off.
-
 **Focus areas** — what to cover thoroughly:
-<!-- Prose for humans; the parseable value is the `focus:` key above.
-     Response-time budgets are NOT a focus value — declare them in qa/perf.yaml
-     (percentile + scenario_ms budgets; reported by `sungen delivery`). -->
+<!-- List from: functional, security, ui, accessibility, performance -->
 <!-- Example: functional, security -->
 
 **Mandatory coverage:**

package/dist/harness/flow-contract.d.ts

@@ -1,71 +0,0 @@
-import { ScenarioInfo } from './parse';
-export interface FlowContract {
-    goal: string;
-    actor?: string;
-    trigger?: string;
-    precondition?: string;
-    outcome: {
-        screen: string;
-        assertion?: string;
-    };
-    value?: string;
-    /** Journey phases this flow declares. Default [HP, ER, EH]; UI is allowed but
-     *  never demanded (presentation is the balance axis's business, not coverage's). */
-    phases: string[];
-    /** The mutated collection (cart, order, application …) — enables regression dims. */
-    stateful?: string;
-    budgets?: Record<string, number>;
-}
-export interface FlowQualityResult {
-    hasContract: boolean;
-    contract?: FlowContract;
-    /** Parse/shape errors — a broken contract is reported, never silently ignored. */
-    errors: string[];
-    outcomeProven: boolean;
-    /** Manual-only proof: the goal is covered but not by automation. */
-    outcomeManualOnly: boolean;
-    /** Scenario names that never touch the outcome screen (guard/error phases excluded). */
-    offGoal: string[];
-    offGoalRatio: number;
-    /** Off-goal categories, for the split suggestion ("VP-FILTER-* looks like its own flow"). */
-    offGoalCategories: string[];
-    phases: {
-        phase: string;
-        covered: boolean;
-        automated: boolean;
-    }[];
-    /** Covered-and-automated phases / declared phases (UI excluded) — the flow coverage axis. */
-    phaseRatio: number;
-    /** Cross-namespace transitions followed by an assertion / all transitions. */
-    handoffs: {
-        total: number;
-        asserted: number;
-        ratio: number;
-    };
-}
-export declare function flowContractPath(unitDir: string): string;
-/** Load + validate. Returns null when absent; a present-but-broken file returns errors. */
-export declare function loadFlowContract(unitDir: string): {
-    contract: FlowContract | null;
-    errors: string[];
-};
-/** Phase of a scenario: its declared phase token (FL-HP-001 / VP-FLOW-ER-02 / MS-EH-005)
- *  when present, else vocabulary detection. */
-export declare function phaseOf(s: ScenarioInfo, declared: string[]): string | null;
-/**
- * Verify the suite against the contract. Deterministic; a flow without a contract
- * returns hasContract:false and neutral values (the audit reports the checklist).
- */
-export declare function flowQuality(unitDir: string, scenarios: ScenarioInfo[]): FlowQualityResult;
-/**
- * Generalized stateful regression depth: the contract names the mutated collection,
- * so the three dims (count-proof · teardown · multi-source) stop being cart-only.
- */
-export declare function statefulDepthFor(collection: string, scenarios: ScenarioInfo[]): {
-    countProof: boolean;
-    teardown: boolean;
-    multiSource: boolean;
-    missing: string[];
-    ratio: number;
-};
-//# sourceMappingURL=flow-contract.d.ts.map

package/dist/harness/flow-contract.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"flow-contract.d.ts","sourceRoot":"","sources":["../../src/harness/flow-contract.ts"],"names":[],"mappings":"AA0BA,OAAO,EAAE,YAAY,EAAE,MAAM,SAAS,CAAC;AAGvC,MAAM,WAAW,YAAY;IAC3B,IAAI,EAAE,MAAM,CAAC;IACb,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,OAAO,EAAE;QAAE,MAAM,EAAE,MAAM,CAAC;QAAC,SAAS,CAAC,EAAE,MAAM,CAAA;KAAE,CAAC;IAChD,KAAK,CAAC,EAAE,MAAM,CAAC;IACf;wFACoF;IACpF,MAAM,EAAE,MAAM,EAAE,CAAC;IACjB,qFAAqF;IACrF,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,OAAO,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;CAClC;AAED,MAAM,WAAW,iBAAiB;IAChC,WAAW,EAAE,OAAO,CAAC;IACrB,QAAQ,CAAC,EAAE,YAAY,CAAC;IACxB,kFAAkF;IAClF,MAAM,EAAE,MAAM,EAAE,CAAC;IACjB,aAAa,EAAE,OAAO,CAAC;IACvB,oEAAoE;IACpE,iBAAiB,EAAE,OAAO,CAAC;IAC3B,wFAAwF;IACxF,OAAO,EAAE,MAAM,EAAE,CAAC;IAClB,YAAY,EAAE,MAAM,CAAC;IACrB,6FAA6F;IAC7F,iBAAiB,EAAE,MAAM,EAAE,CAAC;IAC5B,MAAM,EAAE;QAAE,KAAK,EAAE,MAAM,CAAC;QAAC,OAAO,EAAE,OAAO,CAAC;QAAC,SAAS,EAAE,OAAO,CAAA;KAAE,EAAE,CAAC;IAClE,6FAA6F;IAC7F,UAAU,EAAE,MAAM,CAAC;IACnB,8EAA8E;IAC9E,QAAQ,EAAE;QAAE,KAAK,EAAE,MAAM,CAAC;QAAC,QAAQ,EAAE,MAAM,CAAC;QAAC,KAAK,EAAE,MAAM,CAAA;KAAE,CAAC;CAC9D;AAID,wBAAgB,gBAAgB,CAAC,OAAO,EAAE,MAAM,GAAG,MAAM,CAExD;AAED,2FAA2F;AAC3F,wBAAgB,gBAAgB,CAAC,OAAO,EAAE,MAAM,GAAG;IAAE,QAAQ,EAAE,YAAY,GAAG,IAAI,CAAC;IAAC,MAAM,EAAE,MAAM,EAAE,CAAA;CAAE,CAkCrG;AAaD;+CAC+C;AAC/C,wBAAgB,OAAO,CAAC,CAAC,EAAE,YAAY,EAAE,QAAQ,EAAE,MAAM,EAAE,GAAG,MAAM,GAAG,IAAI,CAU1E;AAED;;;GAGG;AACH,wBAAgB,WAAW,CAAC,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,YAAY,EAAE,GAAG,iBAAiB,CA2EzF;AAED;;;GAGG;AACH,wBAAgB,gBAAgB,CAAC,UAAU,EAAE,MAAM,EAAE,SAAS,EAAE,YAAY,EAAE,GAAG;IAAE,UAAU,EAAE,OAAO,CAAC;IAAC,QAAQ,EAAE,OAAO,CAAC;IAAC,WAAW,EAAE,OAAO,CAAC;IAAC,OAAO,EAAE,MAAM,EAAE,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,CAWlL"}