create-ai-project 1.20.4 → 1.20.6

package/CHANGELOG.md

@@ -5,6 +5,50 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [1.20.6] - 2026-04-10
+
+### Added
+
+- **quality-fixer: Incomplete implementation detection gate** (agents) — New blocking Step 1 reviews uncommitted diffs for stub/placeholder code (TODO markers, hardcoded returns, empty bodies) before any quality checks run. Returns `stub_detected` status with structured response; orchestrator routes back to task-executor for completion. Applied to both quality-fixer and quality-fixer-frontend
+- **quality-fixer: stub_detected branch handling** (commands/skills) — All consuming commands (build, front-build, implement, add-integration-tests) and orchestration guide updated with `stub_detected` routing in the 4-step task cycle
+- **design-sync: Match Basis Rules with confidence levels** (agents) — Replaced "Same Concept Criteria" with structured match rules. High confidence (exact_string, explicit_alias) produces confirmed conflicts; medium confidence (same_endpoint_role, same_integration_role, same_ac_slot) produces candidate conflicts with required structural evidence. Output split into CONFIRMED_CONFLICTS and CANDIDATE_CONFLICTS sections
+- **code-verifier: Literal identifier referential integrity** (agents) — New evidence rule verifies that concrete identifiers in documents (paths, endpoints, type names, config keys) have corresponding codebase definitions. Single authoritative definition sufficient for high confidence (exception to 2-source rule)
+- **technical-designer: Dependency Inventory for update mode** (agents) — Required step before document modification: extract literal identifiers from update scope, verify against codebase and Accepted ADRs, output structured inventory with conflict flags. Applied to both technical-designer and technical-designer-frontend
+- **investigator: Execution path mapping and node-by-node fault check** (agents) — Replaced hypothesis model with failure point model. New Step 3 traces execution paths per symptom (pathMap); new Step 4 checks each node for faults with explicit criteria and checkStatus (supported/weakened/blocked/not_reached). Multiple failure points preserved across layers
+- **verifier: Path coverage check and independent failure point evaluation** (agents) — Replaced ACH (alternative hypothesis generation + single-winner selection) with coverage-based verification. New Step 4 checks pathMap completeness; Step 6 evaluates each failure point independently with failurePointRelationships. Coverage assessment (sufficient/partial/insufficient) replaces confidence (high/medium/low)
+- **solver: Coverage-based recommendation strategy** (agents) — Accepts confirmedFailurePoints and coverageAssessment instead of causes and confidence. Failure points with blocked/not_reached status routed to residualRisks. Bash tool added
+- **diagnose: Coverage gate before solver** (commands) — New Step 4 blocks solver invocation until coverageAssessment=sufficient. Insufficient/partial loops back to investigator (max 2 iterations) before offering user choice to proceed
+- **acceptance-test-generator: Reserved E2E slot for user-facing journeys** (agents) — Multi-step user journey definition added with 3 conditions and user-facing vs service-internal classification. Budget enforcement reserves 1 E2E slot for user-facing journeys regardless of ROI threshold
+- **integration-e2e-testing: Multi-Step User Journey definition** (skills) — System-type-specific boundary definitions (Web/Mobile/CLI/API), classification scope for reserved slot vs normal ROI path vs E2E Gap Check
+- **work-planner: E2E Gap Check** (agents) — Warns when no E2E skeletons provided and Design Doc contains user-facing multi-step journey. Skipped when e2eAbsenceReason is provided from upstream
+- **documentation-criteria: UI Spec in skill description** (skills) — Added UI Spec to skill description trigger for improved routing precision
+
+### Changed
+
+- **acceptance-test-generator: ROI calculation fix** (agents) — Removed cost division from ROI formula (was mathematically broken — E2E could never reach threshold). Raw score (0-120) now used for same-type ranking. Push-down analysis preserves E2E candidates that are part of multi-step journeys
+- **acceptance-test-generator: Nullable E2E contract** (agents/commands/skills) — generatedFiles.e2e is now nullable with e2eAbsenceReason. Contract propagated through plan, front-plan, implement, and orchestration guide. 3 Generation Report patterns (e2e emitted / e2e null / both null)
+- **design-sync: sync_status contract alignment** (skills) — Orchestration guide updated from synced/conflicts_found to NO_CONFLICTS/CONFLICTS_FOUND matching design-sync actual output
+- **quality-fixer: blocked condition logic** (agents) — Fixed AND/OR logic inversion in Japanese version. All 3 conditions must be simultaneously met (multiple valid fixes AND business judgment required AND all confirmation methods exhausted)
+- **quality-fixer: Task-scoped stub detection** (agents) — git diff HEAD scoped to task-relevant files when orchestrator provides file list, preventing false positives from unrelated changes in multi-change branches
+- **PRD template: Sequential AC IDs** (skills) — AC format changed from flat `AC:` to sequential IDs (AC-001, AC-002, ...) for downstream traceability. Template shows 1-requirement-multiple-AC pattern
+- **UI Spec template: Semantic design tokens** (skills) — Flat token table replaced with semantic sub-sections (Color Roles, Typography Hierarchy, Spacing Scale, Elevation, Border Radius). Responsive Behavior table replaces flat breakpoint list. State x Display Matrix expanded with concrete component references
+
+## [1.20.5] - 2026-04-06
+
+### Added
+
+- **work-planner: Design-to-Plan Traceability** (agents) — New Step 5 scans all DD sections and extracts technical requirements into five categories (impl-target, connection-switching, contract-change, verification, prerequisite). Each item is mapped to a covering task in a Traceability table. Items without a covering task are marked as `gap` with mandatory justification; the plan remains in draft status until all gaps are user-confirmed
+- **plan-template: Design-to-Plan Traceability section** (skills) — Traceability table template with sample rows for both `covered` and `gap` entries, category value definitions, and gap status definitions
+- **design-template: Interface Change Matrix** (skills) — New section after Change Impact Map documenting existing-to-new interface mappings with conversion requirements and compatibility methods
+- **design-template: Verification Method column** (skills) — Integration Points List expanded with Verification Method column to explicitly capture how each switching point is verified
+- **orchestration-guide: Gap handling for work-planner handoff** (skills) — Orchestrator must present gap entries to user, keep plan in draft until confirmed, and block downstream agents (task-decomposer, etc.) until all gaps are resolved. Unjustified gaps are errors returned to work-planner
+
+### Changed
+
+- **work-planner: Output Policy gap exception** (agents) — Added explicit exception to the immediate-output policy: plans containing `gap` entries are output as draft and require user confirmation before finalizing
+- **work-planner: Category table simplified** (agents) — Removed Task Requirement column (redundant with Traceability table's Category column). Category values now use short codes (impl-target, etc.) matching the Traceability table directly
+- **orchestration-guide: technical-designer → work-planner handoff deduplicated** (skills) — Replaced inline category descriptions with reference to work-planner Step 5, establishing work-planner as single source of truth for category definitions
+
 ## [1.20.4] - 2026-04-05
 
 ### Added

package/README.ja.md

@@ -137,7 +137,7 @@ Claude Codeで利用できる主要なコマンド
 | `/plan` | 設計書から作業計画書を作成 | 設計承認後 |
 | `/build` | 既存の計画から実行 | 作業の再開時 |
 | `/review` | コードの準拠性確認 | 実装完了後 |
-| `/diagnose` | 根本原因分析ワークフロー | デバッグ、トラブルシューティング |
+| `/diagnose` | 問題診断ワークフロー | デバッグ、トラブルシューティング |
 | `/reverse-engineer` | コードからPRD/Design Docを生成 | 既存システムのドキュメント化 |
 | `/create-skill` | 対話形式で新しいスキルを作成 | プロジェクト固有のルールを追加 |
 | `/refine-skill` | 品質レビュー付きでスキルを修正 | スキルの精度改善 |
@@ -175,10 +175,10 @@ Claude Codeで利用できる主要なコマンド
 | `coding-standards` | 汎用コーディング原則、アンチパターン、デバッグ |
 | `typescript-rules` | TypeScript型安全性、非同期パターン、リファクタリング |
 | `typescript-testing` | Vitest、TDD、カバレッジ要件 |
-| `documentation-criteria` | PRD、ADR、Design Doc基準 |
+| `documentation-criteria` | PRD、ADR、Design Doc、UI Spec基準 |
 | `technical-spec` | アーキテクチャ、環境設定、ビルドコマンド |
 | `implementation-approach` | 戦略パターン、タスク分解 |
-| `integration-e2e-testing` | 統合/E2Eテスト設計、ROIベース選択 |
+| `integration-e2e-testing` | 統合/E2Eテスト設計、ROIベース選択、ジャーニー定義 |
 | `frontend-typescript-rules` | Reactコンポーネント設計、Props駆動パターン |
 | `frontend-typescript-testing` | React Testing Library、MSW、Playwright E2E |
 | `frontend-technical-spec` | Reactアーキテクチャ、環境設定、データフロー |

package/README.md

@@ -137,7 +137,7 @@ Essential commands for Claude Code:
 | `/plan` | Create work plan from design | After design approval |
 | `/build` | Execute from existing plan | Resume work |
 | `/review` | Check code compliance | Post-implementation |
-| `/diagnose` | Root cause analysis workflow | Debugging, troubleshooting |
+| `/diagnose` | Problem diagnosis workflow | Debugging, troubleshooting |
 | `/reverse-engineer` | Generate PRD/Design Docs from code | Legacy system documentation |
 | `/create-skill` | Create a new skill through interactive dialog | Adding project-specific rules |
 | `/refine-skill` | Modify an existing skill with quality review | Improving skill accuracy |
@@ -175,10 +175,10 @@ This boilerplate provides the principles used in agentic implementation workflow
 | `coding-standards` | Universal coding principles, anti-patterns, debugging |
 | `typescript-rules` | TypeScript type safety, async patterns, refactoring |
 | `typescript-testing` | Vitest, TDD, coverage requirements |
-| `documentation-criteria` | PRD, ADR, Design Doc standards |
+| `documentation-criteria` | PRD, ADR, Design Doc, UI Spec standards |
 | `technical-spec` | Architecture, environment, build commands |
 | `implementation-approach` | Strategy patterns, task decomposition |
-| `integration-e2e-testing` | Integration/E2E test design, ROI-based selection |
+| `integration-e2e-testing` | Integration/E2E test design, ROI-based selection, journey definition |
 | `frontend-typescript-rules` | React component design, Props-driven patterns |
 | `frontend-typescript-testing` | React Testing Library, MSW, Playwright E2E |
 | `frontend-technical-spec` | React architecture, environment, data flow |

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "create-ai-project",
-  "version": "1.20.4",
+  "version": "1.20.6",
   "packageManager": "npm@10.8.2",
   "description": "TypeScript boilerplate with skills and sub-agents for Claude Code. Prevents context exhaustion through role-based task splitting.",
   "keywords": [