@sun-asterisk/sungen 3.2.20-beta.1 → 3.2.20-beta.2

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (63) hide show
  1. package/dist/cli/commands/audit.d.ts.map +1 -1
  2. package/dist/cli/commands/audit.js +8 -0
  3. package/dist/cli/commands/audit.js.map +1 -1
  4. package/dist/cli/commands/delivery.d.ts.map +1 -1
  5. package/dist/cli/commands/delivery.js +7 -0
  6. package/dist/cli/commands/delivery.js.map +1 -1
  7. package/dist/exporters/matrix/export.d.ts.map +1 -1
  8. package/dist/exporters/matrix/export.js +11 -0
  9. package/dist/exporters/matrix/export.js.map +1 -1
  10. package/dist/exporters/matrix/render-xlsx.d.ts.map +1 -1
  11. package/dist/exporters/matrix/render-xlsx.js +15 -0
  12. package/dist/exporters/matrix/render-xlsx.js.map +1 -1
  13. package/dist/exporters/matrix/types.d.ts +2 -0
  14. package/dist/exporters/matrix/types.d.ts.map +1 -1
  15. package/dist/exporters/matrix/types.js.map +1 -1
  16. package/dist/exporters/playwright-report-parser.d.ts.map +1 -1
  17. package/dist/exporters/playwright-report-parser.js +1 -0
  18. package/dist/exporters/playwright-report-parser.js.map +1 -1
  19. package/dist/exporters/types.d.ts +2 -0
  20. package/dist/exporters/types.d.ts.map +1 -1
  21. package/dist/harness/audit.d.ts +2 -0
  22. package/dist/harness/audit.d.ts.map +1 -1
  23. package/dist/harness/audit.js +72 -9
  24. package/dist/harness/audit.js.map +1 -1
  25. package/dist/harness/flow-contract.d.ts +71 -0
  26. package/dist/harness/flow-contract.d.ts.map +1 -0
  27. package/dist/harness/flow-contract.js +235 -0
  28. package/dist/harness/flow-contract.js.map +1 -0
  29. package/dist/harness/flow-plan.d.ts +3 -0
  30. package/dist/harness/flow-plan.d.ts.map +1 -1
  31. package/dist/harness/flow-plan.js +6 -2
  32. package/dist/harness/flow-plan.js.map +1 -1
  33. package/dist/harness/parse.d.ts +5 -0
  34. package/dist/harness/parse.d.ts.map +1 -1
  35. package/dist/harness/parse.js +29 -1
  36. package/dist/harness/parse.js.map +1 -1
  37. package/dist/harness/perf.d.ts +40 -0
  38. package/dist/harness/perf.d.ts.map +1 -0
  39. package/dist/harness/perf.js +136 -0
  40. package/dist/harness/perf.js.map +1 -0
  41. package/dist/harness/sensors.d.ts.map +1 -1
  42. package/dist/harness/sensors.js +13 -1
  43. package/dist/harness/sensors.js.map +1 -1
  44. package/dist/orchestrator/templates/ai-src/commands/add-flow.md +41 -3
  45. package/dist/orchestrator/templates/ai-src/skills/sungen-tc-generation/SKILL.md +35 -16
  46. package/dist/orchestrator/templates/qa-context.md +14 -1
  47. package/package.json +3 -3
  48. package/src/cli/commands/audit.ts +8 -0
  49. package/src/cli/commands/delivery.ts +6 -0
  50. package/src/exporters/matrix/export.ts +11 -0
  51. package/src/exporters/matrix/render-xlsx.ts +15 -0
  52. package/src/exporters/matrix/types.ts +2 -0
  53. package/src/exporters/playwright-report-parser.ts +2 -0
  54. package/src/exporters/types.ts +2 -0
  55. package/src/harness/audit.ts +75 -10
  56. package/src/harness/flow-contract.ts +229 -0
  57. package/src/harness/flow-plan.ts +10 -3
  58. package/src/harness/parse.ts +31 -1
  59. package/src/harness/perf.ts +112 -0
  60. package/src/harness/sensors.ts +13 -1
  61. package/src/orchestrator/templates/ai-src/commands/add-flow.md +41 -3
  62. package/src/orchestrator/templates/ai-src/skills/sungen-tc-generation/SKILL.md +35 -16
  63. package/src/orchestrator/templates/qa-context.md +14 -1
@@ -71,6 +71,9 @@ export interface FlowPlan {
71
71
  legs: LegPlan[];
72
72
  byReason: Record<string, number>;
73
73
  capabilityManual: number;
74
+ /** @manual whose reason is "cross-screen → automate via flow" (class XS) — inside a flow this
75
+ * usually means the scenario should simply BE automated here. (#569) */
76
+ crossScreenManual: number;
74
77
  judgmentManual: number;
75
78
  contracts: Contract[];
76
79
  readiness: 'ready' | 'not-ready';
@@ -87,14 +90,18 @@ export function buildFlowPlan(cwd: string, flow: string): FlowPlan {
87
90
  // Legs = distinct screen namespaces.
88
91
  const legMap = new Map<string, { scenarios: Set<string>; refs: Set<string>; automated: boolean }>();
89
92
  const byReason: Record<string, number> = {};
90
- let capabilityManual = 0, judgmentManual = 0;
93
+ let capabilityManual = 0, judgmentManual = 0, crossScreenManual = 0;
91
94
 
92
95
  for (const sc of scenarios) {
93
96
  if (sc.manual) {
94
97
  const { code } = inferReasonCode(sc.tags, sc.reason);
95
98
  byReason[code] = (byReason[code] || 0) + 1;
96
99
  const cls = MANUAL_REASONS[code]?.cls;
97
- if (cls === 'capability') capabilityManual++; else if (cls === 'keep') judgmentManual++;
100
+ // XS ("cross-screen automate via flow") is a THIRD class; it used to be silently
101
+ // dropped from both counters, understating the plan's manual load. (#569)
102
+ if (cls === 'capability') capabilityManual++;
103
+ else if (cls === 'keep') judgmentManual++;
104
+ else if (cls === 'flow') crossScreenManual++;
98
105
  }
99
106
  for (const r of sc.refs) {
100
107
  const leg = r.screen.toLowerCase();
@@ -133,5 +140,5 @@ export function buildFlowPlan(cwd: string, flow: string): FlowPlan {
133
140
  }
134
141
  if (readiness === 'ready') plan.unshift('Selectors present for every automated leg — ready to compile + run.');
135
142
 
136
- return { flow, total: scenarios.length, legs, byReason, capabilityManual, judgmentManual, contracts, readiness, missingLegs, plan };
143
+ return { flow, total: scenarios.length, legs, byReason, capabilityManual, judgmentManual, crossScreenManual, contracts, readiness, missingLegs, plan };
137
144
  }
@@ -38,6 +38,8 @@ export interface ScenarioInfo {
38
38
  requiresCaps?: string[]; // @requires:<cap> — automation-ready but needs an opt-in driver (TQ-11)
39
39
  deferredToFlow?: boolean; // @deferred:flow — owned by a flow, not automated on this screen (H6)
40
40
  ownedByFlow?: string; // @owned-by:<flow> — the flow that owns this deferred scenario (H6)
41
+ /** Ordered steps with their resolved bucket (And/But inherit) — flow handoff analysis (#569). */
42
+ steps?: Array<{ bucket: 'given' | 'when' | 'then'; text: string }>;
41
43
  }
42
44
 
43
45
  /** Format-tolerant: is this token an ID (project's scheme), not a prose word?
@@ -91,6 +93,29 @@ export function parseViewpointOverview(filePath: string): ViewpointEntry[] {
91
93
  }
92
94
  }
93
95
 
96
+ // 1b) Flow-style declarations (#569). Flow viewpoint files commonly declare per-item
97
+ // ids at the END of a bullet ("… → **FL-HP-001**") under phase section headers
98
+ // ("## FL-HP — Happy Path"). Neither matched the table/group passes, so every flow
99
+ // audit collapsed to taxonomy=0% / traceability n-a — scenarios correctly tagged
100
+ // FL-HP-001 were reported as unmapped. Both forms are additive here.
101
+ for (const raw of lines) {
102
+ const line = raw.trim();
103
+ const section = line.match(/^##\s+([A-Z]{2,}(?:-[A-Z0-9]{2,})*)\s+[—–-]\s*(.*)$/);
104
+ if (section && isViewpointId(section[1] + '-0')) {
105
+ const id = section[1].toUpperCase();
106
+ if (!entries.has(id)) entries.set(id, { id, priority: 'Unknown', reason: section[2] ?? '' });
107
+ }
108
+ if (/^[-*+]\s/.test(line)) {
109
+ const arrow = line.match(/(?:→|->)\s*\*{0,2}([A-Z][A-Z0-9]*(?:-[A-Z0-9]+)*-\d+[a-zA-Z]?)\*{0,2}\s*$/);
110
+ if (arrow) {
111
+ const id = arrow[1].toUpperCase();
112
+ if (!entries.has(id)) {
113
+ entries.set(id, { id, priority: 'Unknown', reason: line.replace(/\s*(?:→|->).*$/, '').replace(/^[-*+]\s+/, '') });
114
+ }
115
+ }
116
+ }
117
+ }
118
+
94
119
  // 2) Viewpoint Grouping: ### Required / ### Recommended / ### Optional → bullet list
95
120
  let group: ViewpointEntry['group'] | undefined;
96
121
  for (const raw of lines) {
@@ -144,7 +169,9 @@ function classifyScenario(sc: ParsedScenario): ScenarioInfo {
144
169
  // Category is everything between `VP-` and the final `-<sequence>` — INCLUDING hyphens, so
145
170
  // compound categories (VP-LIST-DISPLAY-01, VP-ADD-TO-CART-03, VP-PRODUCT-DISCOVERY-02) parse,
146
171
  // not just single-word ones. A single-word category (VP-CART-001) still works. (H1)
147
- const codeMatch = sc.name.match(/\bVP-([A-Z]+(?:-[A-Z]+)*)-\d+/i);
172
+ // Flows use journey-phase ids (FL-HP-001 / FL-ER-002) — the VP- anchor rejected them, so every
173
+ // flow scenario had NO category and the whole suite bucketed `other` (taxonomy=0%, #569).
174
+ const codeMatch = sc.name.match(/\b(?:VP|FL)-([A-Z]+(?:-[A-Z]+)*)-\d+/i);
148
175
  const vpCode = codeMatch ? codeMatch[0].toUpperCase() : undefined;
149
176
  const category = codeMatch ? codeMatch[1].toUpperCase() : undefined;
150
177
  // Project-scheme ID: the leading token of the title (VP0-001 / MS-HP-001 / VP-LIST-001).
@@ -158,12 +185,14 @@ function classifyScenario(sc: ParsedScenario): ScenarioInfo {
158
185
  const skeletonParts: string[] = [];
159
186
  const textParts: string[] = [sc.name];
160
187
  const stepTextParts: string[] = [];
188
+ const orderedSteps: Array<{ bucket: 'given' | 'when' | 'then'; text: string }> = [];
161
189
 
162
190
  for (const step of sc.steps as ParsedStep[]) {
163
191
  const kw = step.keyword.trim();
164
192
  if (kw === 'Given' || kw === 'When' || kw === 'Then') last = kw;
165
193
  textParts.push(step.text);
166
194
  stepTextParts.push(step.text);
195
+ orderedSteps.push({ bucket: (kw === 'And' || kw === 'But' ? last : kw).toLowerCase() as 'given' | 'when' | 'then', text: step.text });
167
196
  // normalized skeleton: keep [refs] (distinct targets = distinct tests),
168
197
  // but neutralize {{vars}} and quoted values so EP/data families collapse.
169
198
  const skel = step.text
@@ -199,6 +228,7 @@ function classifyScenario(sc: ParsedScenario): ScenarioInfo {
199
228
  stepSkeleton: skeletonParts.join(' | '),
200
229
  haystack: textParts.join(' ').toLowerCase(),
201
230
  stepsText: stepTextParts.join(' ').toLowerCase(),
231
+ steps: orderedSteps,
202
232
  vpId,
203
233
  casesDataset,
204
234
  queryRefs: queryRefs.size ? [...queryRefs] : undefined,
@@ -0,0 +1,112 @@
1
+ /**
2
+ * Performance budgets — config + percentile math + the per-unit verdict. (#569)
3
+ *
4
+ * A flow's regression value includes "still fast enough": after a lib/framework
5
+ * upgrade the main journeys must not only pass but hold their response-time
6
+ * budget. Sungen had no perf concept at all — the Playwright JSON parser even
7
+ * dropped the `duration` field Playwright already emits on every result.
8
+ *
9
+ * Scope discipline:
10
+ * - This is config + measurement + report over runs sungen already makes.
11
+ * Real load tests stay @manual:M8 → a dedicated tool.
12
+ * - The AUDIT never reads it: the quality score is documented as a pure
13
+ * function of the design artifacts ("reads no test-results, live page, or
14
+ * clock"). Perf reports where runs are already read — `sungen delivery`
15
+ * and the dashboard. Advisory: a blown budget never fails the design gate.
16
+ *
17
+ * Config: qa/perf.yaml
18
+ * percentile: p75 # default p75 — "≥75% of runs meet the budget"
19
+ * defaults:
20
+ * scenario_ms: 30000 # whole-scenario wall clock (Playwright duration)
21
+ * page_load_ms: 3000 # Phase B — needs per-transition runtime timing
22
+ * transition_ms: 2000 # Phase B
23
+ * units:
24
+ * place-order: { scenario_ms: 20000 }
25
+ */
26
+ import * as fs from 'fs';
27
+ import * as path from 'path';
28
+ import { parse as parseYaml } from 'yaml';
29
+ import { readTextFile } from './read-text';
30
+
31
+ export interface PerfConfig {
32
+ /** 0..100 — e.g. 75 for p75. */
33
+ percentile: number;
34
+ defaults: Record<string, number>;
35
+ units: Record<string, Record<string, number>>;
36
+ }
37
+
38
+ export interface PerfVerdict {
39
+ unit: string;
40
+ metric: string; // 'scenario_ms' today; page_load_ms/transition_ms in Phase B
41
+ percentile: number; // 75
42
+ budgetMs: number;
43
+ measuredMs: number; // the pXX of the observed durations
44
+ samples: number;
45
+ pass: boolean;
46
+ /** Titles of the slowest offenders (only when failing), for the report. */
47
+ slowest: Array<{ title: string; ms: number }>;
48
+ }
49
+
50
+ export function perfConfigPath(projectRoot: string): string {
51
+ return path.join(projectRoot, 'qa', 'perf.yaml');
52
+ }
53
+
54
+ /** Absent file → null (perf reporting is opt-in; nothing changes until configured). */
55
+ export function loadPerfConfig(projectRoot: string): PerfConfig | null {
56
+ const p = perfConfigPath(projectRoot);
57
+ if (!fs.existsSync(p)) return null;
58
+ let raw: Record<string, unknown>;
59
+ try { raw = parseYaml(readTextFile(p)) as Record<string, unknown>; } catch { return null; }
60
+ if (!raw || typeof raw !== 'object') return null;
61
+ const pctRaw = String(raw.percentile ?? 'p75').toLowerCase().replace(/^p/, '');
62
+ const percentile = Math.min(100, Math.max(1, Number(pctRaw) || 75));
63
+ const num = (o: unknown): Record<string, number> => {
64
+ const out: Record<string, number> = {};
65
+ if (o && typeof o === 'object') {
66
+ for (const [k, v] of Object.entries(o as Record<string, unknown>)) {
67
+ const n = Number(v);
68
+ if (Number.isFinite(n) && n > 0) out[k] = n;
69
+ }
70
+ }
71
+ return out;
72
+ };
73
+ const units: Record<string, Record<string, number>> = {};
74
+ if (raw.units && typeof raw.units === 'object') {
75
+ for (const [u, o] of Object.entries(raw.units as Record<string, unknown>)) units[u] = num(o);
76
+ }
77
+ return { percentile, defaults: num(raw.defaults), units };
78
+ }
79
+
80
+ /**
81
+ * Nearest-rank percentile (ceil), the standard "≥pXX of samples meet the budget"
82
+ * reading: p75 of [a…] is the value at ceil(0.75·n) in the sorted list. One
83
+ * sample → that sample. Deterministic, no interpolation.
84
+ */
85
+ export function percentileOf(p: number, values: number[]): number {
86
+ if (values.length === 0) return 0;
87
+ const sorted = [...values].sort((a, b) => a - b);
88
+ const rank = Math.min(sorted.length, Math.max(1, Math.ceil((p / 100) * sorted.length)));
89
+ return sorted[rank - 1];
90
+ }
91
+
92
+ /** Budget for a metric on a unit: per-unit override, else defaults, else none. */
93
+ export function budgetFor(config: PerfConfig, unit: string, metric: string): number | undefined {
94
+ return config.units[unit]?.[metric] ?? config.defaults[metric];
95
+ }
96
+
97
+ /**
98
+ * The scenario_ms verdict for one unit's run. `durations` = per-test wall-clock ms
99
+ * (a @cases scenario contributes one sample per row-test — each is a real run).
100
+ */
101
+ export function perfVerdict(
102
+ config: PerfConfig,
103
+ unit: string,
104
+ samples: Array<{ title: string; ms: number }>,
105
+ ): PerfVerdict | null {
106
+ const budgetMs = budgetFor(config, unit, 'scenario_ms');
107
+ if (budgetMs === undefined || samples.length === 0) return null;
108
+ const measuredMs = percentileOf(config.percentile, samples.map((s) => s.ms));
109
+ const pass = measuredMs <= budgetMs;
110
+ const slowest = pass ? [] : [...samples].sort((a, b) => b.ms - a.ms).slice(0, 3);
111
+ return { unit, metric: 'scenario_ms', percentile: config.percentile, budgetMs, measuredMs, samples: samples.length, pass, slowest };
112
+ }
@@ -33,10 +33,22 @@ const BUCKET_ORDER: Array<[string, string[]]> = [
33
33
  ];
34
34
  const BUCKETS: Record<string, string[]> = Object.fromEntries(BUCKET_ORDER);
35
35
 
36
+ // Flow journey-phase categories (FL-HP-001, FL-ER-002 …). Matched on exact SEGMENTS,
37
+ // never by containment — 'SHOP'.includes('HP') is true, which is exactly the kind of
38
+ // false hit substring matching would produce for two-letter phase tokens. (#569)
39
+ const PHASE_BUCKETS: Record<string, string> = {
40
+ HP: 'business-core', // happy path = the business goal itself
41
+ ER: 'validation-security', // error recovery (validation must not trap the journey)
42
+ EH: 'validation-security', // guards & leakage (direct access, back, refresh)
43
+ };
44
+
36
45
  /** Classify a VP category into a balance bucket by keyword containment + precedence (H1). */
37
46
  export function bucketForCategory(category: string | undefined): string {
38
47
  const cat = (category || '').toUpperCase();
39
48
  if (!cat) return 'other';
49
+ for (const seg of cat.split('-')) {
50
+ if (PHASE_BUCKETS[seg]) return PHASE_BUCKETS[seg];
51
+ }
40
52
  for (const [bucket, kws] of BUCKET_ORDER) {
41
53
  if (kws.some((k) => cat.includes(k))) return bucket;
42
54
  }
@@ -351,7 +363,7 @@ export function flowRegressionDepth(scenarios: ScenarioInfo[]): FlowDepthResult
351
363
  // 1. Count/quantity proof — a row count or item quantity, not just presence of a row.
352
364
  const countProof = any(/\b(quantity|qty|two (?:rows|lines|cart)|row count|count column|number of items|one[_ ]row|two[_ ]rows|qty[_ ])/i);
353
365
  // 2. Teardown — removes the item and verifies the empty/zero state (the inverse operation).
354
- const teardown = any(/\b(remove|delete|clear)\b/i) && any(/\b(empty|no items|zero|removed|0 items)\b/i);
366
+ 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);
355
367
  // 3. Multi-source — the cart is fed from >1 source (the main list AND a recommended/related rail).
356
368
  const multiSource = any(/\b(recommended|related|you may also|suggest)\b/i) && addsToCart;
357
369
 
@@ -86,15 +86,52 @@ qa/flows/${input:flow}/
86
86
  └── ui/ # Screenshots, mockups
87
87
  ```
88
88
 
89
- ### 1a. Identify the screens in the flow
89
+ ### 1a. Define the flow's BOUNDARY, then its screens
90
90
 
91
- Ask the user: "Which screens does this flow visit, in order? (e.g., login dashboard → award-form → confirmation)"
91
+ A flow is the **smallest complete business action chain**: one clear trigger ending in ONE
92
+ observable, valuable outcome. Before asking for screens, walk this checklist with the user —
93
+ if 1, 3 or 8 fails, propose SPLITTING into separate flows:
94
+
95
+ 1. Exactly **one business goal**? (cart correctness + category filtering = two flows)
96
+ 2. A clear **trigger** and precondition?
97
+ 3. **One observable final outcome**? (a final assertion you can write in one sentence)
98
+ 4. Is that outcome **valuable to the actor**? (an order placed, a password reset — not "a page rendered")
99
+ 5. Is **every step necessary** for that outcome?
100
+ 6. Are all steps at the **same business abstraction**?
101
+ 7. Are optional/error branches **phases of this goal** (ER/EH), not new goals?
102
+ 8. Does **no segment** form an independently valuable flow on its own?
103
+ 9. Can you write **a single clear final assertion**?
104
+ 10. Can you name it "**Verb + outcome**"? (`place-order`, `reset-password` — not `cart-and-filter`)
105
+
106
+ Then ask: "Which screens does this flow visit, in order? (e.g., login → dashboard → award-form → confirmation)"
92
107
 
93
108
  Record the screen list — you will need it for:
94
109
  - Filling `spec.md` (Step 3)
95
110
  - Suggesting `[Screen:Element]` namespace prefixes
96
111
  - Capturing visuals per screen (Step 2)
97
112
 
113
+ ### 1b. Author the Flow Contract (`requirements/flow-contract.yaml`)
114
+
115
+ Write the answers down as the flow's contract — `sungen audit` scores the flow **against it**
116
+ (the `flowCoverage` axis: HP/ER/EH journey phases; `FLOW-OUTCOME-UNPROVEN` when no automated
117
+ scenario asserts data on the outcome screen; `FLOW-SCOPE-CREEP` when scenarios never touch it):
118
+
119
+ ```yaml
120
+ goal: "Place an order for a product added from home" # Verb + outcome
121
+ actor: user
122
+ trigger: "Add a product to the cart from the home featured list"
123
+ precondition: "A registered account; an empty cart"
124
+ outcome:
125
+ screen: checkout # the [Screen:...] namespace carrying the final proof
126
+ assertion: "The confirmation shows the order number and the paid total"
127
+ value: "The customer has paid; the shop has a new order"
128
+ phases: [HP, ER, EH] # journey phases (default); add UI only if the flow owns UI states
129
+ stateful: cart # the mutated collection, if any — enables regression-depth dims
130
+ ```
131
+
132
+ **A filled contract is an INPUT to generation — never an output.** Like `test-viewpoint.md`,
133
+ generation must not rewrite it to match what was generated; disagree → propose the diff and ask.
134
+
98
135
  ### 2. Capture visual source
99
136
 
100
137
  **Mobile path** (`platform: mobile`):
@@ -187,7 +224,8 @@ If user picks `/sungen:create-test`, **you MUST use the Skill tool** to invoke i
187
224
  - Test data namespaced by phase: `login.email`, `submission.nominee`
188
225
  - `@flow` tag required at feature level
189
226
  - `Background:` should only contain the starting navigation — the URL path (web) or the `--reach` nav recipe (mobile)
190
- - Each scenario = one phase of the journey
227
+ - Each scenario = one phase of the journey; ids are `FL-<PHASE>-NNN` (`HP`/`ER`/`EH`, optional `UI`)
228
+ - One flow = ONE business goal with ONE observable outcome (`requirements/flow-contract.yaml`) — a segment with its own value is its own flow
191
229
  {{#cap parallel-subagents}}
192
230
  - Mobile flows are tagged `@platform:mobile` and run via `/sungen:run-test <flow>` (WebdriverIO, not Playwright)
193
231
  {{/cap}}
@@ -605,22 +605,40 @@ error:
605
605
 
606
606
  > **Auto-detect**: if path is `qa/flows/<name>/` → use this section. Skip Steps 1–4 above.
607
607
 
608
+ **Read `requirements/flow-contract.yaml` FIRST — it is the flow's boundary and the yardstick
609
+ `sungen audit` scores the flow against** (`flowCoverage` axis = journey phases HP/ER/EH automated;
610
+ `FLOW-OUTCOME-UNPROVEN`; `FLOW-SCOPE-CREEP`). No contract yet → author it with the user via the
611
+ boundary checklist in `add-flow` (one business goal · clear trigger · ONE observable outcome
612
+ valuable to the actor · name = "Verb + outcome"), THEN generate. **A filled contract is an INPUT —
613
+ never rewrite it to match your output** (same rule as `test-viewpoint.md`).
614
+
608
615
  | Aspect | Screen | Flow |
609
616
  |---|---|---|
610
- | Section focus | UI patterns per section | Journey phases across screens |
617
+ | Section focus | UI patterns per section | Journey phases toward ONE declared outcome |
611
618
  | Selector format | `[Element]` | `[Screen:Element]` (namespaced) |
612
619
  | Test data keys | `{{variable}}` | `{{phase.variable}}` |
613
620
  | Feature tag | `@auto` / `@smoke` etc. | `@flow` (required) |
614
- | Viewpoints | VP-UI/VAL/LOGIC/SEC per section | VP-LOGIC (transitions), VP-SEC (auth persistence), VP-VAL (cross-screen data) |
621
+ | Scenario ids | `VP-<CATEGORY>-NNN` | `FL-<PHASE>-NNN` phases: `HP` (happy path), `ER` (error recovery), `EH` (guards), `UI` (journey UI states, optional) |
615
622
 
616
- **Scenarios to generate:**
623
+ **Scenarios to generate — every phase demanded by the contract, automated:**
617
624
 
618
- | Category | What to test |
619
- |---|---|
620
- | Happy path | Complete flow end-to-end with valid data |
621
- | Auth persistence | Auth state maintained across screen transitions |
622
- | Error recovery | Invalid input mid-flow fixcontinue |
623
- | Cross-screen data | Data entered on screen A visible on screen B |
625
+ | Phase | What to test | Scoring |
626
+ |---|---|---|
627
+ | `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` |
628
+ | `FL-ER` error recovery | Invalid input mid-flow error shown → fix → the journey still completes. Validation must not trap the journey. | uncovered → `flowCoverage` drops |
629
+ | `FL-EH` guards | Direct URL access without the precondition · browser back · refresh · expired context — each ends in a safe observable state. | uncovered `flowCoverage` drops |
630
+ | 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`) |
631
+ | Stateful regression (when `stateful:` declared) | Count/quantity proof · teardown (remove → empty) · multi-source add. | missing dims cap `businessDepth` (`FLOW-DEPTH`) |
632
+
633
+ **Boundary discipline while generating:** every scenario must serve the contract's goal. A scenario
634
+ that never touches `outcome.screen` and is not a guard (`EH`) or error-recovery (`ER`) belongs in a
635
+ DIFFERENT flow — propose the split instead of writing it here (`FLOW-SCOPE-CREEP` will flag it).
636
+ Auth persistence across transitions is part of `EH` unless the project declares it its own phase.
637
+
638
+ **Manual in flows**: always `@manual:Mx` with the reason code — bare `@manual` is flagged
639
+ (`MANUAL-CODE-MISSING`) because the capability planner cannot route it. Typical flow deferrals:
640
+ inbox/mail oracle → `M5`, network-request count → `M3`, context expiry control → `M7`. A
641
+ cross-screen scenario inside the flow's own goal is NOT manual — automate it here.
624
642
 
625
643
  ```gherkin
626
644
  @flow @auth:user
@@ -630,19 +648,20 @@ Feature: Award Submission Flow
630
648
  Given User is on [Login] page
631
649
 
632
650
  @high
633
- Scenario: User logs in successfully
651
+ Scenario: FL-HP-001 A signed-in user's nomination is submitted and confirmed
634
652
  When User fill [Login:Email] field with {{login.email}}
635
653
  And User fill [Login:Password] field with {{login.password}}
636
654
  And User click [Login:Submit] button
637
655
  Then User see [Dashboard] page
638
-
639
- @high
640
- Scenario: User submits nomination
641
656
  When User click [Dashboard:Awards] link
642
- Then User see [Awards] page
643
- When User fill [Awards:Nominee] field with {{submission.nominee}}
657
+ And User fill [Awards:Nominee] field with {{submission.nominee}}
644
658
  And User click [Awards:Submit] button
645
- Then User see {{success_message}} message
659
+ Then User see [Awards:Success Message] text with {{success_message}}
660
+
661
+ @high
662
+ Scenario: FL-EH-001 Direct access to the award form without login redirects to login
663
+ When User go to [Awards] page
664
+ Then User see [Login] page
646
665
  ```
647
666
 
648
667
  ```yaml
@@ -45,8 +45,21 @@ Example:
45
45
 
46
46
  ## Testing Strategy
47
47
 
48
+ Machine-readable intent — `sungen audit` reads these keys (Intent Profile). Values here are
49
+ live even when the surrounding text changes; an invalid value silently falls back to the default.
50
+
51
+ focus: functional
52
+ <!-- focus: functional | e-commerce | security | smoke — drives the audit's depth threshold -->
53
+
54
+ risk_tier: normal
55
+ <!-- risk_tier: high | normal | low -->
56
+
57
+ To silence driver suggestions in audit findings, add a line: capability_suggestions with value off.
58
+
48
59
  **Focus areas** — what to cover thoroughly:
49
- <!-- List from: functional, security, ui, accessibility, performance -->
60
+ <!-- Prose for humans; the parseable value is the `focus:` key above.
61
+ Response-time budgets are NOT a focus value — declare them in qa/perf.yaml
62
+ (percentile + scenario_ms budgets; reported by `sungen delivery`). -->
50
63
  <!-- Example: functional, security -->
51
64
 
52
65
  **Mandatory coverage:**