@sun-asterisk/sungen 3.2.16-beta.3 → 3.2.16-beta.4

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

@@ -36,7 +36,8 @@ oracle-shape/precondition mismatches are review-level, silenced once approved an
 D executability (precondition · condition+
 data · trigger · oracle all renderable; every `{{var}}` resolves; **no template token may survive
 into a rendered cell** — test-data cross-references are resolved for display) · E drift
-(fingerprint mismatch → back to review) · G review state (proposed groups block the official
+(scenario fingerprint mismatch → back to review; the map's OWN reviewed
+wording/grouping is fingerprinted as `__map__` too, so post-approval edits re-open review) · G review state (proposed groups block the official
 render; `--preview` renders a DRAFT watermark) · R requirement coverage (spec FR/TR/NFR ids with
 no trace and no `requirements:` status → warning) · W wording lint (map intent/oracle containing
 tokens, `[Selector]` refs, DSL phrasing, or generator labels → warning).
@@ -51,17 +52,33 @@ Precondition, `Action:` → Action, `Observable:` → Expected Result, `Oracle:`
 mid-flow assertions inline as `Verify: …`; only the final Then block is the Expected Result.
 Empty test values render as `(empty)`.
 
+**Output format**: **XLSX only by default** (one artifact). `--format csv` writes the flat CSV
+instead; `--format both` writes both.
+
 **Workbook**: `Testcases` sheet — parent rows + outline-level-1 variant sub-rows for **every**
 item (single-variant included: the sub-row carries the source VP-id, resolved data, and the
-result/evidence entry). Collapse outline for the customer view, expand to execute. Variant Result
-cells have a dropdown (Passed/Failed/Blocked/Pending/N/A) and the parent Result is a **live Excel
-formula** over its children (failed→blocked→pending→partial→passed, e.g. `2/3 Passed · 1 Failed`)
-— a parent can never contradict its variants, even after manual edits. ID + Target columns are
-frozen; dates are ISO (`2026-08-04`). `Coverage` sheet — requirement coverage table (every FR/TR/
-NFR id with an explicit status), target × category grid with explicit `—` gaps, dispositions,
-manifest. CSV is flat with a `Level` column (`item`/`variant`) + a requirement-coverage appendix.
+result/evidence entry). Sub-rows are **delta-only**: knowledge common to the whole item
+(precondition, trigger) is written ONCE on the parent; a child repeats only what distinguishes it
+(condition/data, precondition delta, trigger when it differs) plus its own precise oracle and
+execution fields. The parent never carries placeholder text — when triggers differ the cell is
+simply empty and the sub-rows carry them.
+
+Parent preconditions are the **intersection** of the variants' — variant-specific setup stays on
+the variant, so no variant inherits a wrong start state. Variant Result cells have a dropdown
+(Passed/Failed/Blocked/Pending/N/A) and the parent Result is a **live Excel formula** over its
+children (failed→blocked→pending→partial→passed; **`N/A` leaves the denominator**, all-`N/A` →
+`N/A`). The summary band counts results with live formulas too. Evidence and Defect ID are
+separate columns; ID + Target frozen; ISO dates; landscape print with repeated header rows.
+
+`Coverage` sheet — requirement coverage table (every FR/TR/NFR id with an explicit status),
+target × category grid with explicit `—` gaps, dispositions, manifest. CSV mirrors the same model
+flat with a `Level` column (`item`/`variant`) + a requirement-coverage appendix.
 `delivery_item_count` ≠ progress — variants are the execution metric.
 
+**Data fidelity**: invisible test data is made visible, never normalized — `''` → `(empty)`,
+whitespace-only → `(5 spaces)`, padded → `" value "` quoted verbatim. A trim/collapse here would
+silently break the whitespace tests it describes.
+
 **Authoring guidance the matrix rewards** (create-test side): payload/provider matrices (SQLi
 payload lists, OAuth provider sets) belong in `@cases` datasets so each case is an atomic,
 independently-reportable variant; keep dataset `case:` labels short and stable (`CHK-EMAIL-I1`),