@dzhechkov/p-replicator 1.12.0 → 1.13.2

package/README.md

@@ -14,7 +14,7 @@
 [![Tests](https://img.shields.io/badge/tests-440%20pass%20%7C%203%20skip-22c55e?style=flat-square)](#test-infrastructure)
 [![Skills](https://img.shields.io/badge/skills-10-8b5cf6?style=flat-square)](#skills-reference)
 [![Commands](https://img.shields.io/badge/slash%20commands-11-f59e0b?style=flat-square)](#commands-reference)
-[![Hooks](https://img.shields.io/badge/hooks-10-3b82f6?style=flat-square)](#hooks-system)
+[![Hooks](https://img.shields.io/badge/hooks-24-3b82f6?style=flat-square)](#hooks-system)
 [![SPARC Pipeline](https://img.shields.io/badge/SPARC-5%20phases-ec4899?style=flat-square)](#pipeline-overview--replicate-phases)
 
 <br>
@@ -101,12 +101,29 @@
 
 ## 📖 What is p-replicator?
 
-`@dzhechkov/p-replicator` installs a ready-made `.claude/` toolkit into any project: **11 slash commands**, **10 skills**, **4 agents**, **9 rules**, **10 hook utilities**, and a `settings.json` with pre-configured hooks and a multi-line statusline dashboard.
+`@dzhechkov/p-replicator` installs a ready-made `.claude/` toolkit into any project: **11 slash commands**, **10 skills**, **4 agents**, **13 rules**, **24 hook utilities**, and a `settings.json` with pre-configured hooks and a multi-line statusline dashboard.
 
-The flagship `/replicate` command takes a project through a **5-phase pipeline**:
+The flagship `/replicate` command takes a project through a **6-phase pipeline**:
 
 ```
 Phase 0 (optional)    Product Discovery     reverse-engineering of similar companies
+Phase 0.5 (ALWAYS)    Source Product Profile  capture the LOOK of the product being replicated
+
+> **Third-party design analyses enter as HYPOTHESES (v1.13).** A ready-made DESIGN.md from a
+> library like styles.refero.design is a claim, not a measurement: the profile header now carries
+> `**Происхождение:**` (closed list: `прокликано | сторонний-разбор | вручную | не снято`), and a
+> `сторонний-разбор` row REQUIRES the analysis source + its own capture date (read from the
+> source's markup, e.g. `extractedAt` — never from a tool's paraphrase). Hypothesis rows never
+> reach `Specification.md` without a DATED live confirmation; the gate is
+> `node .claude/hooks/check-look-origin.cjs .` (exit `0` verified · `1` a promoted hypothesis is
+> NAMED · `2` the check did not run — never "all clear").
+>
+> **Quoted evidence is now byte-verifiable (v1.13).** In `goap-research-ed25519`, every quote
+> carries a closed acquisition method (`raw-fetch | tool-summary | search-listing | manual`), and
+> the "verbatim" verdict can ONLY be produced by `verify_verbatim` comparing the quote against a
+> captured source excerpt (sha256-bound, self-gitignoring store) — a report cannot attest its own
+> quotes, and a tool paraphrase is categorically capped below "verbatim". A quoted phrase absent
+> from its stored source is a named violation.
 Phase 1               Planning              11 SPARC documents (PRD, Architecture, Pseudocode, ...)
 Phase 2               Validation            5-agent swarm validates against INVEST + SMART
 Phase 3               Toolkit Generation    project-specific agents, rules, skills
@@ -239,6 +256,24 @@ EOF
 
 Full recipe: see [User Guide](./README/eng/02_user_guide.md#feature-workflow-in-an-existing-project-mode-2) ([RU](./README/ru/02_user_guide.md#feature-workflow-в-существующем-проекте-mode-2)).
 
+### Three traceability gates
+
+Byte-level formats of the three reports (`validation-report.md`, `05_completion.md`, `review-report.md`) live in `templates/.claude/skills/requirements-validator/references/feature-report-contracts.md`; `/feature` names them per phase, the gates below enforce them. The shared `brutal-honesty-review` skill stays byte-identical to its canonical copy — the specification contract is a `/feature` obligation, not a change to the generic reviewer.
+
+`/feature` refuses silent traceability loss at validation, implementation, and review. The
+requirements score also has a `Traceability = 0` blocking floor: without a `## Criterion scenarios`
+table mapping every AC id to a named scenario, a high average cannot advance the story.
+
+| Check | Command | Expected pass | Example named gap |
+|---|---|---|---|
+| Machine-key traceability and the scoring floor's input | `bash "$CHECK_PIPELINE_GAPS" "${CLAUDE_PROJECT_DIR:-.}" --traceability` | `PASS traceability: 3 AC ids mapped to 3 scenarios` | `GAP AC-checkout-2: no named scenario in 02_pseudocode.md` |
+| Validation report is bound to the current specification bytes | `bash "$CHECK_PIPELINE_GAPS" "${CLAUDE_PROJECT_DIR:-.}" --report-revision` | `PASS report-revision: validation-report.md matches 01_specification.md` | `GAP report-revision: validation=4f27c19c9b20 specification=8ab35d68e021` |
+| Every criterion names a real test file and title | `bash "$CHECK_PIPELINE_GAPS" "${CLAUDE_PROJECT_DIR:-.}" --completion` | `PASS completion: 3 AC ids covered by tests` | `GAP AC-checkout-2: test title not found in tests/checkout.test.js` |
+| Review discloses its family and answers the exact specification revision | `node "${CLAUDE_PROJECT_DIR:-.}/.claude/hooks/check-review-contract.cjs" "${CLAUDE_PROJECT_DIR:-.}" "checkout"` | `PASS review-contract feature=checkout AC-ids=3 rows=3` | `GAP AC-checkout-2 has no Spec conformance row` |
+
+All four commands use the same fail-closed convention: exit `0` passes, exit `1` prints every named
+gap, and exit `2` means the inputs could not be established. Neither non-zero result is a warning.
+
 ---
 
 ## 📦 Installation
@@ -263,7 +298,7 @@ This creates:
 - `.claude/commands/` — 11 slash commands (`/replicate`, `/run`, `/feature`, ...)
 - `.claude/agents/` — 4 pipeline agents
 - `.claude/rules/` — 9 governance rules
-- `.claude/hooks/` — 10 cross-platform Node scripts
+- `.claude/hooks/` — 24 cross-platform Node scripts
 - `.claude/settings.json` — hooks + statusline configuration
 - `.p-replicator.json` — install manifest
 
@@ -289,8 +324,8 @@ This creates:
 | **Skills** | 10 | 92 files (880K+ chars — MEASURED via `wc -c` over `templates/.claude/skills`: 880,679). Modular architecture: foundation → composite → master orchestrator |
 | **Commands** | 11 | `/replicate`, `/harvest`, `/start`, `/plan`, `/feature`, `/go`, `/run`, `/next`, `/myinsights`, `/docs`, `/deploy` |
 | **Agents** | 4 | `replicate-coordinator`, `product-discoverer`, `doc-validator`, `harvest-coordinator` |
-| **Rules** | 9 | `cost-of-detection-ladder`, `replicate-pipeline`, `skill-interface-protocol`, `git-workflow`, `insights-capture`, `feature-lifecycle`, `docker-ports`, `swarm-file-evidence`, `honest-configuration` |
-| **Hooks** | 10 | Ten cross-platform Node utilities, including the harvest insight writer |
+| **Rules** | 13 | `cost-of-detection-ladder`, `replicate-pipeline`, `skill-interface-protocol`, `git-workflow`, `insights-capture`, `feature-lifecycle`, `docker-ports`, `swarm-file-evidence`, `honest-configuration`, `embeddable-widget`, `incoming-webhooks` |
+| **Hooks** | 15 | Fifteen cross-platform Node utilities: 4 wired to events, 11 invoked deliberately (the harvest insight writer, the status line, and seven `check-*` gates) |
 | **Settings** | 1 | `settings.json` with statusLine + SessionStart + UserPromptSubmit + Stop hooks pre-configured |
 
 After running `/replicate`, the toolkit also generates **project-specific** artifacts:
@@ -427,6 +462,102 @@ A conscious rejection **with a reason written down** passes; a silent drop does
 **What it does NOT prove:** that anything was built, or that copying a competitor's growth move is
 lawful. It proves an obligation was carried forward, nothing more.
 
+### `check-look-trace.cjs` — the source product's LOOK was not captured and dropped
+
+```bash
+node .claude/hooks/check-look-trace.cjs .
+```
+
+```
+❌ часть обязательств по облику потеряна (1 из 2):
+   • FR-LOOK-002 (путь)
+   Каждое надо либо перенести в docs/Specification.md, либо отклонить С ПРИЧИНОЙ — молча уронить нельзя.
+```
+
+When a project reproduces an existing product, that product's LOOK — palette, typography, density,
+screen layout, the order of steps — is the substance of the task. **Phase 0.5 (Source Product
+Profile)** asks that question in every run, including the `--from-docs` entry that skips Phase 0
+entirely, and writes `docs/source-product-profile.md`. This checker asks whether its
+`FR-LOOK-nnn` rows survived into `docs/Specification.md`. A conscious rejection **with a reason
+written down** passes; a silent drop does not.
+
+Exit `2` covers five states, and none of them is a pass: no profile (the phase did not run), no
+Specification, an untouched template table, an EMPTY `путь` axis that was never declared — and the
+two legitimate non-captures, `ИСТОЧНИКА НЕТ` (the project replicates nothing) and `НЕ ИЗМЕРЕНО` (a
+source was NAMED but could not be captured, with a reason from the closed list `no-browser-mcp` |
+`no-browser` | `unreachable` | `auth-required` | `out-of-scope` | `bot-protected` | `timeout` |
+`robots-disallowed`). Per the shipped `honest-configuration` rule (CFG-I4), an unreachable source of
+truth yields UNKNOWN, never a plausible value — a palette invented for a product nobody looked at is
+exactly that plausible value.
+
+**Two axes, one identifier family, separate answers.** `облик` (what is seen) and `путь` (the order
+of screens) are a COLUMN of `FR-LOOK-nnn`, never a second namespace — but they fail apart, so each
+declares its own status: `**Статус съёмки:**` and `**Статус съёмки (путь):**`. The path declaration
+is demanded only when the axis carries no rows; rows are its answer. A proven loss outranks an
+unanswered axis — exit `1` beats exit `2`.
+
+The `облик` capture is done by the canonical `clone-website` skill from the separate
+[`@dzhechkov/skills-website-cloner`](https://www.npmjs.com/package/@dzhechkov/skills-website-cloner)
+package, run recon-only. It is **called, never vendored** (ADR-0001): unavailable prerequisites make
+the outcome `НЕ ИЗМЕРЕНО` with a named reason, they never stall the pipeline.
+
+### `capture-source-path.cjs` — the `путь` axis, clicked through in a browser
+
+```bash
+node .claude/hooks/capture-source-path.cjs https://example.test/ --max-pages 5 --delay-ms 1500
+```
+
+```
+ℹ️  robots.txt прочитан, стартовый путь разрешён
+ℹ️  шаг 2 — экран входа. Останавливаемся: за вход не ходим, это законная последняя точка пути.
+
+✅ СНЯТ: ось «путь», 2 экрана(ов). Доказательства: .p-replicator/source-path-capture
+
+| FR-LOOK-005 | Шаг 1 из 2; стартовый экран; полей формы: 0; заметных призывов: 1 из 2; уровней заголовков: 2 | путь | / | прокликано …, step-01.aria.txt | ЧЕРНОВИК |
+| FR-LOOK-006 | Шаг 2 из 2; достижим одним действием с предыдущего; полей формы: 2; …; это экран входа | путь | /signup | прокликано …, step-02.aria.txt | ЧЕРНОВИК |
+| FR-LOOK-007 | Последовательность экранов до конца снятого пути: 2 экрана(ов), / → /signup; путь упирается в экран входа | путь | / | прокликано …, capture.json | ЧЕРНОВИК |
+
+Замеренные закономерности СТАРТОВОГО экрана (материал для оси «облик», строк не выпускаем):
+  шаг сетки отступов: 4px (доля 0.91) · кеглей: 3 · начертаний: 3 · радиусов: 1 · переменных в :root: 3
+```
+
+`clone-website` captures what is visible at ONE url; a person's route through a product —
+registration, onboarding, first value, paywall — was captured by nothing. This walks it and emits
+rows in the SAME `FR-LOOK-nnn` family, continuing the profile's numbering, on the `путь` axis.
+
+Three outcomes, each with its own exit code: `0` captured · `1` the source opened but has no onward
+step (a one-screen product — a legitimate `ИСТОЧНИКА НЕТ` for this axis, and a PROVEN negative, not
+blindness) · `2` `НЕ ИЗМЕРЕНО` with a reason from the closed list, printed as the exact profile
+lines to paste.
+
+**Playwright is an external prerequisite** — this package ships ZERO dependencies, so the module is
+resolved on the machine at run time (`PLAYWRIGHT_MODULE`, the project, `npm root -g`) and its
+absence is the honest outcome `no-browser`, never a crash:
+`npm i -D playwright && npx playwright install chromium`.
+
+**What it captures, and what it refuses to.** REGULARITIES — the spacing step, the type scale, how
+many hierarchy levels, how many form fields, how many screens to first value — never VALUES to carry
+over, such as this exact purple. Only computed styles and semantic roles (aria, form types,
+accessible names); **never class names** like `sx-ds2y8i`, which bundlers change on every build of
+someone else's site. Third-party CSS and DOM are copyrighted code — a basis for measurement, never
+material to copy — so they are not stored by default, and the evidence directory ships its own
+`.gitignore`. `robots.txt` is read before crawling more than one page, and a refusal is the outcome
+`robots-disallowed`. Authentication and any circumvention are refused outright: a login screen is
+recorded as the legitimate last step and the crawl stops there. One thread, a pause between pages
+(minimum 250 ms), a hard page cap of 12 — exceeding either is clamped out loud.
+
+**When to run it:** inside Phase 0.5, once the source product is named.
+
+**What it does NOT prove:** the designer's intent. It reports "padding 8px ×137", never "a 4pt
+step is the design language"; the reading is yours.
+
+**When to run it:** at the Phase 1 → Phase 2 boundary, beside `check-growth-trace.cjs`. Run at the
+end of Phase 0.5 itself it answers `2` by construction — nothing has been promoted yet — and there
+it only tells you the profile parses.
+
+**What it does NOT prove:** that the interface resembles the source. It proves an obligation was
+carried forward. Resemblance is proven by visual comparison, not by matching identifiers.
+
 ### `check-pipeline-gaps.sh` — FR/NFR/AC keys are joins, not decoration
 
 The package owns the checker at `scripts/check-pipeline-gaps.sh`. `/feature` resolves that packaged
@@ -479,7 +610,7 @@ npx @dzhechkov/p-replicator verify
 
 The command checks:
 
-- **Pre-shipped contract** (must-have): 10 skills + 11 commands + 4 agents + 9 rules + settings.json + 10 hooks
+- **Pre-shipped contract** (must-have): 10 skills + 11 commands + 4 agents + 13 rules + settings.json + 25 hooks
 - **Post-/replicate hints** (advisory): CLAUDE.md, project-specific agents, feature-roadmap.json, security rules, etc.
 
 **Exit codes:**
@@ -852,7 +983,7 @@ P-Replicator V1.5.x ● user │ Sonnet 4.7
 🚀 Pipeline   /<cmd> ▓▓▓░░░░ 50%  │ Phase: VALIDATE (2/4)  │ Last: /replicate
 🎯 Roadmap    [●●●○○○○○] mvp 3/8   │ Done 5/12  │ ▶ auth-jwt  │ Domain: banking
 📊 SPARC      ●11/11  │ 🟢 78/100  │ Plans ●3  │ ADRs ●2  │ Harvest 2026-05-05
-🛠️ Toolkit    Skills ●10/10 │ Cmds ●11/11 │ Agents ●4+3 │ Rules ●9+2 │ Hooks ●9/9
+🛠️ Toolkit    Skills ●10/10 │ Cmds ●11/11 │ Agents ●4+3 │ Rules ●13+2 │ Hooks ●17/17
 💡 Insights   ●12 (2026-05-06) │ Tests 85/85 ✓ │ MCP ●1/1 │ Settings ✓ │ 🧬 Keysarium ✓
 ```
 
@@ -888,7 +1019,7 @@ For internals, see [admin guide](./README/eng/03_admin_guide.md#statusline--dash
 
 ## 🪝 Hooks System
 
-`p-replicator` ships **10 cross-platform Node utilities** in `.claude/hooks/`. The Bash traceability
+`p-replicator` ships **11 cross-platform Node utilities** in `.claude/hooks/`. The Bash traceability
 gate described above remains a package script and is resolved directly by `/feature`; it is not a
 Claude Code hook.
 
@@ -901,6 +1032,16 @@ Claude Code hook.
 | `statusline.cjs` | (statusLine config) | Multi-line dashboard above the prompt |
 | `state-update.cjs` | (utility) | Argv-driven helper for writing `.claude/.p-replicator-state.json` |
 | `write-insight.cjs` | (utility, invoked by `/harvest`) | Establish Markdown, return `created`/`appended`/`duplicate`, then best-effort project the same record through dz |
+| `check-ports.cjs` | (utility, invoke deliberately) | docker-ports Правило №0 against a real Compose config — exits 0/1/2 |
+| `check-docs-complete.cjs` | (utility, invoke deliberately) | Are the Phase-1 documents written and placeholder-free, before the Phase-2 swarm — exits 0/1/2 |
+| `check-growth-trace.cjs` | (utility, invoke deliberately) | Did the M5 `FR-GROWTH-nnn` seed reach `docs/Specification.md` — exits 0/1/2 |
+| `check-look-trace.cjs` | (utility, invoke deliberately) | Did the Phase-0.5 `FR-LOOK-nnn` source-look seed reach `docs/Specification.md` — exits 0/1/2 |
+| `capture-source-path.cjs` | (utility, invoke deliberately) | Phase-0.5 `путь` axis: click through the source product in a browser and emit `FR-LOOK-nnn` rows — exits 0/1/2 |
+
+The four `check-*` utilities and `capture-source-path.cjs` are **not wired to any event**, and that
+is deliberate: a hook of this package is non-blocking by contract (pinned by a test that requires
+exit 0), so a hook could only print — it could never refuse anything. Invoke them where their answer
+changes a decision.
 
 ### Cross-platform discipline
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@dzhechkov/p-replicator",
-  "version": "1.12.0",
+  "version": "1.13.2",
   "description": "P-Replicator — Claude Code toolkit for AI-assisted product development (Vibe Coding). Full /replicate pipeline, 10 modular skills (880K+ chars), /harvest knowledge extraction, swarm agents, quality gates, security patterns, cross-project learning.",
   "main": "src/cli.js",
   "bin": {
@@ -19,15 +19,6 @@
     ".dz-manifest.json",
     "sbom.json"
   ],
-  "scripts": {
-    "prepublishOnly": "node --test tests/snapshot/templates.test.js && node ../../../scripts/prepublish-gate.mjs",
-    "test": "node --test tests/unit/honest-failure-rules.test.js tests/unit/detection-ladder-contract.test.js tests/unit/detection-ladder-registry.test.js tests/unit/utils.test.js tests/unit/adr-scanner-contract.test.js tests/unit/sparc-reconciliation.test.js tests/unit/pipeline-file-ownership.test.js tests/unit/validation-gate-teeth.test.js tests/unit/hooks-project-anchored.test.js tests/unit/insights-dz-delegation.test.js tests/unit/statusline-two-roots.test.js tests/unit/external-dependency-check.test.js tests/unit/spec-pseudocode-traceability.test.js tests/unit/sync-templates-guard.test.js tests/unit/db-port-rule.test.js tests/unit/roadmap-one-schema.test.js tests/unit/hooks-report-failures.test.js tests/unit/growth-module-b2b-gate.test.js tests/unit/detect-parse-anchor.test.js tests/unit/statusline-honest-labels.test.js tests/unit/autocommit-deletion.test.js tests/unit/check-ports.test.js tests/unit/doctor-insight-flow.test.js tests/unit/adr-decision-coverage.test.js tests/unit/growth-requirements-bridge.test.js tests/unit/check-growth-trace.test.js tests/unit/growth-gate-conditional.test.js tests/unit/module-copy-identity.test.js tests/unit/growth-axes-and-compliance.test.js tests/unit/skill-paths-prebaked.test.js tests/unit/check-docs-complete.test.js tests/unit/absence-is-not-emptiness.test.js tests/unit/guard-forms.test.js tests/unit/generated-guard-templates.test.js tests/unit/shipped-suite-context.test.js tests/unit/assess-scripts.test.js tests/unit/insights-docs-tell-the-truth.test.js tests/unit/insights-writer.test.js tests/unit/check-pipeline-gaps.test.js tests/unit/traceability-machine-ids.test.js tests/unit/traceability-negative-fixture.test.js tests/e2e/lifecycle.test.js tests/e2e/packed-insights-writer.test.js tests/snapshot/templates.test.js",
-    "test:unit": "node --test tests/unit/honest-failure-rules.test.js tests/unit/detection-ladder-contract.test.js tests/unit/detection-ladder-registry.test.js tests/unit/utils.test.js tests/unit/adr-scanner-contract.test.js tests/unit/sparc-reconciliation.test.js tests/unit/pipeline-file-ownership.test.js tests/unit/validation-gate-teeth.test.js tests/unit/hooks-project-anchored.test.js tests/unit/insights-dz-delegation.test.js tests/unit/statusline-two-roots.test.js tests/unit/external-dependency-check.test.js tests/unit/spec-pseudocode-traceability.test.js tests/unit/sync-templates-guard.test.js tests/unit/db-port-rule.test.js tests/unit/roadmap-one-schema.test.js tests/unit/hooks-report-failures.test.js tests/unit/growth-module-b2b-gate.test.js tests/unit/detect-parse-anchor.test.js tests/unit/statusline-honest-labels.test.js tests/unit/autocommit-deletion.test.js tests/unit/check-ports.test.js tests/unit/doctor-insight-flow.test.js tests/unit/adr-decision-coverage.test.js tests/unit/growth-requirements-bridge.test.js tests/unit/check-growth-trace.test.js tests/unit/growth-gate-conditional.test.js tests/unit/module-copy-identity.test.js tests/unit/growth-axes-and-compliance.test.js tests/unit/skill-paths-prebaked.test.js tests/unit/check-docs-complete.test.js tests/unit/absence-is-not-emptiness.test.js tests/unit/guard-forms.test.js tests/unit/generated-guard-templates.test.js tests/unit/shipped-suite-context.test.js tests/unit/assess-scripts.test.js tests/unit/insights-docs-tell-the-truth.test.js tests/unit/insights-writer.test.js tests/unit/check-pipeline-gaps.test.js tests/unit/traceability-machine-ids.test.js tests/unit/traceability-negative-fixture.test.js",
-    "test:e2e": "node --test tests/e2e/lifecycle.test.js tests/e2e/packed-insights-writer.test.js",
-    "test:snapshot": "node --test tests/snapshot/templates.test.js",
-    "test:doctor": "node bin/cli.js doctor",
-    "snapshot:baseline": "node tests/snapshot/update-baseline.js"
-  },
   "keywords": [
     "claude",
     "claude-code",
@@ -80,5 +71,13 @@
   },
   "publishConfig": {
     "access": "public"
+  },
+  "scripts": {
+    "test": "node --test tests/unit/honest-failure-rules.test.js tests/unit/generator-swarm-contract.test.js tests/unit/check-swarm-receipts.test.js tests/unit/check-embed-contract.test.js tests/unit/check-model-cost.test.js tests/unit/check-canon.test.js tests/unit/check-file-ownership.test.js tests/unit/check-source-version.test.js tests/unit/check-handoff-manifest.test.js tests/unit/check-external-deps.test.js tests/unit/check-metric-source.test.js tests/unit/negative-conclusion-gate.test.js tests/unit/check-webhook-contract.test.js tests/unit/check-job-contract.test.js tests/unit/detection-ladder-contract.test.js tests/unit/detection-ladder-registry.test.js tests/unit/utils.test.js tests/unit/adr-scanner-contract.test.js tests/unit/sparc-reconciliation.test.js tests/unit/pipeline-file-ownership.test.js tests/unit/validation-gate-teeth.test.js tests/unit/hooks-project-anchored.test.js tests/unit/insights-dz-delegation.test.js tests/unit/statusline-two-roots.test.js tests/unit/external-dependency-check.test.js tests/unit/spec-pseudocode-traceability.test.js tests/unit/sync-templates-guard.test.js tests/unit/db-port-rule.test.js tests/unit/roadmap-one-schema.test.js tests/unit/hooks-report-failures.test.js tests/unit/growth-module-b2b-gate.test.js tests/unit/detect-parse-anchor.test.js tests/unit/statusline-honest-labels.test.js tests/unit/autocommit-deletion.test.js tests/unit/check-ports.test.js tests/unit/doctor-insight-flow.test.js tests/unit/adr-decision-coverage.test.js tests/unit/growth-requirements-bridge.test.js tests/unit/check-growth-trace.test.js tests/unit/check-look-trace.test.js tests/unit/check-look-origin.test.js tests/unit/quote-provenance.test.js tests/unit/guard-honest-input-meta.test.js tests/unit/look-phase-contract.test.js tests/unit/capture-source-path.test.js tests/unit/growth-gate-conditional.test.js tests/unit/module-copy-identity.test.js tests/unit/growth-axes-and-compliance.test.js tests/unit/skill-paths-prebaked.test.js tests/unit/check-docs-complete.test.js tests/unit/check-review-contract.test.js tests/unit/absence-is-not-emptiness.test.js tests/unit/guard-forms.test.js tests/unit/generated-guard-templates.test.js tests/unit/shipped-suite-context.test.js tests/unit/assess-scripts.test.js tests/unit/insights-docs-tell-the-truth.test.js tests/unit/insights-writer.test.js tests/unit/check-pipeline-gaps.test.js tests/unit/traceability-machine-ids.test.js tests/unit/traceability-completion-gate.test.js tests/unit/traceability-negative-fixture.test.js tests/e2e/lifecycle.test.js tests/e2e/feature-contour.test.js tests/e2e/packed-insights-writer.test.js tests/snapshot/templates.test.js",
+    "test:unit": "node --test tests/unit/honest-failure-rules.test.js tests/unit/generator-swarm-contract.test.js tests/unit/check-swarm-receipts.test.js tests/unit/check-embed-contract.test.js tests/unit/check-model-cost.test.js tests/unit/check-canon.test.js tests/unit/check-file-ownership.test.js tests/unit/check-source-version.test.js tests/unit/check-handoff-manifest.test.js tests/unit/check-external-deps.test.js tests/unit/check-metric-source.test.js tests/unit/negative-conclusion-gate.test.js tests/unit/check-webhook-contract.test.js tests/unit/check-job-contract.test.js tests/unit/detection-ladder-contract.test.js tests/unit/detection-ladder-registry.test.js tests/unit/utils.test.js tests/unit/adr-scanner-contract.test.js tests/unit/sparc-reconciliation.test.js tests/unit/pipeline-file-ownership.test.js tests/unit/validation-gate-teeth.test.js tests/unit/hooks-project-anchored.test.js tests/unit/insights-dz-delegation.test.js tests/unit/statusline-two-roots.test.js tests/unit/external-dependency-check.test.js tests/unit/spec-pseudocode-traceability.test.js tests/unit/sync-templates-guard.test.js tests/unit/db-port-rule.test.js tests/unit/roadmap-one-schema.test.js tests/unit/hooks-report-failures.test.js tests/unit/growth-module-b2b-gate.test.js tests/unit/detect-parse-anchor.test.js tests/unit/statusline-honest-labels.test.js tests/unit/autocommit-deletion.test.js tests/unit/check-ports.test.js tests/unit/doctor-insight-flow.test.js tests/unit/adr-decision-coverage.test.js tests/unit/growth-requirements-bridge.test.js tests/unit/check-growth-trace.test.js tests/unit/check-look-trace.test.js tests/unit/check-look-origin.test.js tests/unit/quote-provenance.test.js tests/unit/guard-honest-input-meta.test.js tests/unit/look-phase-contract.test.js tests/unit/capture-source-path.test.js tests/unit/growth-gate-conditional.test.js tests/unit/module-copy-identity.test.js tests/unit/growth-axes-and-compliance.test.js tests/unit/skill-paths-prebaked.test.js tests/unit/check-docs-complete.test.js tests/unit/check-review-contract.test.js tests/unit/absence-is-not-emptiness.test.js tests/unit/guard-forms.test.js tests/unit/generated-guard-templates.test.js tests/unit/shipped-suite-context.test.js tests/unit/assess-scripts.test.js tests/unit/insights-docs-tell-the-truth.test.js tests/unit/insights-writer.test.js tests/unit/check-pipeline-gaps.test.js tests/unit/traceability-machine-ids.test.js tests/unit/traceability-completion-gate.test.js tests/unit/traceability-negative-fixture.test.js",
+    "test:e2e": "node --test tests/e2e/lifecycle.test.js tests/e2e/packed-insights-writer.test.js",
+    "test:snapshot": "node --test tests/snapshot/templates.test.js",
+    "test:doctor": "node bin/cli.js doctor",
+    "snapshot:baseline": "node tests/snapshot/update-baseline.js"
   }
-}
+}