@longph2102/v-flow 1.5.0

package/AGENTS.md

@@ -0,0 +1,265 @@
+# AGENTS.md — V Flow Instructions for AI Agents
+
+> This file provides instructions for AI coding agents (**GitHub Copilot**, **Antigravity**, Cursor, etc.)
+> working within a project that uses **V Flow** — Agentic AI Software Development Process.
+
+---
+
+## 🧠 Tầng 1: Project Overview & Rules (Auto-injected)
+
+<!-- AUTO-INJECT: context.md START -->
+> ⚠️ **Project Context**: Nội dung `context.md` (~50-100 dòng overview bao gồm tech stack, architecture, no-go zones) sẽ được V Flow CLI tự động inject vào đây.
+<!-- AUTO-INJECT: context.md END -->
+
+<!-- AUTO-INJECT: constitution.md START -->
+> ⚠️ **Project Rules**: Nội dung `constitution.md` (Development rules, formatting, policies) sẽ được V Flow CLI tự động inject vào đây.
+<!-- AUTO-INJECT: constitution.md END -->
+
+---
+
+## 📦 Tầng 2: Module Cards (On-Demand Context)
+
+> Khi làm việc với một feature cụ thể, Agent PHẢI chủ động đọc các file liên quan trong thư mục `.v-flow/modules/<domain>/SKILL.md` (Bao gồm các file: `use-cases.md`, `api-specs.md`, `tech-context.md`, `db.md`, `state-diagram.md`, `permissions.md`, `ui-flows.md`) để lấy context nghiệp vụ sâu hơn.
+
+---
+
+## Project Structure
+
+This project follows the **V Flow** Spec-Driven Development process.
+All development artifacts are stored in `.v-flow/`:
+
+```
+.v-flow/
+├── context.md               ← U.0a: Project Overview (~50-100 dòng, auto-loaded)
+├── constitution.md          ← Project principles & rules
+├── agent-config.md          ← Agent pipeline configuration
+├── modules/                 ← U.0b: Module Cards (Skills format)
+│   └── domain-xxx/
+│       ├── SKILL.md         ← Entry point + frontmatter
+│       ├── use-cases.md     ← Business rules (BA đọc)
+│       ├── api-specs.md     ← API endpoints (Dev đọc)
+│       ├── ui-flows.md      ← Màn hình, UX, Field Mapping
+│       ├── tech-context.md  ← Patterns, dependencies (Dev đọc)
+│       ├── cross-service.md ← Tích hợp liên service
+│       ├── db.md            ← Database, schema, DDL
+│       ├── state-diagram.md ← State machine, transitions
+│       ├── permissions.md   ← Roles, VPS, policies
+│       └── raw/             ← U.0d (Optional — NotebookLM raw data)
+│           ├── sources/     ← Source fulltext files
+│           ├── domain-map.json ← Mind map export
+│           └── overview.md  ← Briefing doc
+├── knowledge/               ← Domain knowledge base
+│   ├── domain-glossary.md
+│   ├── architecture-decisions/
+│   ├── patterns/
+│   ├── golden-tests/
+│   └── lessons-learned/
+├── features/                ← Feature artifacts
+│   └── NNN-feature-name/
+│       ├── spec.md          ← S.1: Specification — SINGLE SOURCE OF TRUTH
+│       ├── .flow-state.yaml ← Machine-readable pipeline state
+│       ├── _session.md      ← Working memory
+│       ├── plan.md          ← B.0a: Technical Plan
+│       ├── tasks.md         ← B.0a: Task Breakdown
+│       ├── test-cases.md    ← B.0b: Test Cases
+│       ├── validation-report.md ← R.0: Master Check
+│       └── metrics-report.md    ← M.0: Metrics
+└── incidents/               ← O.0: Operations
+```
+
+## Critical Rules
+
+1. **Tuân thủ tuyệt đối Project Context & Constitution** được auto-inject ở trên (Tầng 1). Không hành động nếu chưa rõ no-go zones.
+2. **Chủ động đọc Module Cards (Tầng 2)** cho các module liên quan — `.v-flow/modules/domain-xxx/SKILL.md`
+3. **`spec.md` is Single Source of Truth** — if plan/tasks conflict with spec, spec wins
+4. **Check `.flow-state.yaml`** before starting any phase — know where you are in the pipeline
+5. **Read `_session.md`** for working memory — decisions made, open questions, concerns from Master Check
+6. **Check spec Status block** (Section 7) before implementing — all artifacts must be ✅
+7. **If spec changes** — run `/v.sync` before continuing implementation
+8. **Follow the pipeline order** — Spec → Plan + Tests → Master Check → Implement
+9. **Never implement without Master Check = PASS (or PASS WITH CONCERNS)** — check `validation-report.md` *(ngoại lệ: prebuilt flows `quickfix`/`bugfix`/`hotfix` cố ý bỏ R.0 — đánh đổi tốc độ, chỉ dùng cho thay đổi nhỏ/khẩn)*
+10. **Follow Routing Rules** — `validation-report.md` contains `fail_reason` + `route_decision`; don't guess
+11. **Concerns carry-forward** — resolve all [W-xxx] concerns during B.1; verify during R.1
+12. **Update `.flow-state.yaml`** at end of each phase — record verdict, routing, phase transition
+13. **Append to `_session.md`** at end of each phase — execution trace entry
+14. **Use EARS notation** for requirements: Ubiquitous, Event-driven, State-driven, Optional
+15. **Map everything to requirements** — tasks, tests, and code must trace back to REQ-xxx
+16. **TDD** — write tests before implementation code
+17. **Log mỗi ranh giới phase** — cùng lúc với việc append `_session.md` (rule 13), gọi `v-flow log` để ghi vào nhật ký tổng `.v-flow/logs/activity.jsonl` (xem "Activity Logging" bên dưới)
+
+## Activity Logging
+
+Mọi hoạt động "dùng v-flow" được gom về **một nhật ký cục bộ** tại `.v-flow/logs/`:
+- `activity.jsonl` — mỗi dòng 1 JSON (máy đọc / `jq` / grep)
+- `activity.log` — bản dễ đọc cho người
+
+Lệnh CLI tự ghi mỗi lần chạy. **Agent phải chủ động ghi** các mốc của mình bằng `v-flow log` (best-effort — nếu lệnh không có thì bỏ qua, đừng để chặn pipeline):
+
+```bash
+# Bắt đầu một phase
+v-flow log --feature <feature> --phase S.1 --actor ba-agent --event start "Bắt đầu viết spec"
+# Quyết định / mốc quan trọng
+v-flow log --feature <feature> --phase R.0 --actor master-check-agent --event decision "Verdict=PASS_WITH_CONCERNS" --data '{"verdict":"PASS_WITH_CONCERNS","concerns":2}'
+# Kết thúc phase
+v-flow log --feature <feature> --phase B.1 --actor implement-agent --event end "12/12 tasks done"
+```
+
+Ghi tối thiểu ở: **start** và **end** mỗi phase, mỗi **routing decision**, và mỗi **lỗi/blocker**. Xem lại nhanh: `v-flow log --tail 30`. Tắt toàn bộ: đặt `VFLOW_NO_LOG=1`.
+
+### Tự động bắt log qua lifecycle hooks (đa-provider, deterministic)
+
+Ngoài cách agent chủ động gọi `v-flow log` (best-effort), có thể cài hook để **mọi tool-use / session** được ghi tự động — chạy được trên Claude Code, Cursor, GitHub Copilot, OpenAI Codex, Antigravity:
+
+```bash
+v-flow hooks install        # rải hook vào config từng tool (merge, không clobber)
+v-flow hooks status         # xem tool nào đã cài
+v-flow hooks uninstall      # gỡ hook của v-flow (giữ hook của bạn)
+```
+
+Mỗi hook gọi `v-flow log --hook <provider>` — đọc payload tool đẩy qua stdin, chuẩn hoá rồi append vào cùng `.v-flow/logs/`. Hook luôn exit 0, không bao giờ chặn agent. Yêu cầu: `v-flow` có trên PATH (IDE GUI không kế thừa PATH → cài bằng `v-flow hooks install --abs`).
+
+Safety-net tool-agnostic: `v-flow watch` giám sát `.v-flow/**` và log mọi thay đổi file — bắt được cả tool không có hook lẫn sửa tay.
+
+## Available Slash Commands
+
+| Command | Phase | Purpose |
+|---------|-------|---------|
+| `/v.help` | Any | **Help Agent — gợi ý bước tiếp theo dựa trên trạng thái dự án** |
+| `/v.understand` | U.0 | **Understand Agent — Multi-Layer Understanding (`--all`, `--module`, `--check`, `--refresh`)** |
+| `/v.import-ba-docs` | U.0d | **Import BA Docs Agent — Import tài liệu BA từ NotebookLM** |
+| `/v.constitution` | S.0a | Create/update project constitution (interactive) |
+| `/v.constitute` | S.0a | **Constitute Agent — AI scan → tự động draft `constitution.md` v2** |
+| `/v.specify` | S.1 | BA Agent — create spec + init `.flow-state.yaml` + `_session.md` (`--dual` for Dual BA loop) |
+| `/v.ba-critic` | S.1 | **BA Critic Agent — Adversarial spec review 1 round (dùng sau `--dual`)** |
+| `/v.prototype` | S.1.5 | **Prototype Agent — Throwaway prototype để validate ambiguity (`--logic`, `--ui`, `--question`)** |
+| `/v.ba-to-ptyc` | S.1.6 | **BA to PTYC Agent — spec.md → PTYC.docx (Phân tích Yêu cầu chuẩn ISO), optional Doc Export** |
+| `/v.plan` | B.0a | Plan Agent — create technical plan + tasks (`--with-security`, `--with-ux`, `--with-api`) |
+| `/v.test` | B.0b | Test Agent — create test cases from spec |
+| `/v.check` | R.0 | Master Check — cross-validate Spec ↔ Plan ↔ Tests + explicit routing |
+| `/v.implement` | B.1 | Implement Agent — TDD coding, concerns handling (`--yolo` for auto-mode) |
+| `/v.review` | R.1 | Adversarial Code Review — ≥ 3 issues + routing rules |
+| `/v.sync` | Any | **Sync Agent — propagate spec changes to artifacts** |
+| `/v.sprint` | B.1 | Sprint Planning Agent — tạo sprint-status.md |
+| `/v.retrospective` | R.1 | Retrospective Agent — tổng kết sau epic/sprint |
+| `/v.metrics` | M.0 | Collect agent performance metrics |
+| `/v.operations` | O.0 | **Operations Agent — rà soát vận hành định kỳ / phân loại incident → route flow khắc phục** |
+| `/v.status` | Any | **📊 Dashboard — pipeline state, routing decision, concerns** |
+| `/v.quickfix` | S.1→B.1→R.1 | **Quick Fix Flow — bug fix nhỏ, skip full pipeline** |
+| `/v.bugfix` | S.1→B.0→B.1→R.1 | **Bugfix Flow — root cause analysis + 3-behavior model + 3-tier TDD (Kiro-inspired)** |
+| `/v.hotfix` | O.0→B.1→R.1 | **Hotfix Flow — P1/P2 production incident** |
+| `/v.rewind` | Any | **Time Travel: Rewind — rollback spec/pipeline về version cũ** |
+| `/v.fork` | Any | **Time Travel: Fork — branch feature để thử hướng khác** |
+
+## Pipeline Flow
+
+```
+Constitution → Agent Setup → BA Agent [→ BA Critic (opt-in `--dual`)] → [Prototype Agent (opt S.1.5)] → Plan Agent ⧑ Test Agent → Master Check → Implement → Verify → Metrics
+```
+
+**Với Legacy Project**, thêm bước đầu tiên:
+```
+init → /v.understand --all (context.md) → /v.understand --module (Module Cards) → /v.constitute (constitution.md v2) → ... (giống trên)
+Understand (context.md + Module Cards) → Constitution → ... (giống trên)
+```
+
+**Prebuilt Flows** (skip full pipeline):
+```
+Quick Fix:  S.1 nhanh → B.1 → R.1          ← bug fix nhỏ, đơn giản (<5 files)
+Bugfix:     S.1 → B.0(opt) → B.1 → R.1     ← bug cần root cause analysis + regression prevention
+Hotfix:     O.0 → B.1 → R.1 → O.0          ← production incident P1/P2
+```
+
+**Nhóm Phase** (U → S → B → R → M → O):
+- **U** Understand: `U.0` — Multi-Layer Understanding (context.md + Module Cards + Graph)
+- **S** Specify: `S.0a` Foundation, `S.0b` Agent Setup (thủ công — cấu hình `agent-config.md`, không có agent riêng), `S.1` Specification, `S.1.5` Prototype (optional)
+- **B** Build: `B.0` Plan+Test (+Optional: Security/UX/API), `B.1` Implement
+- **R** Review: `R.0` Master Check (explicit routing), `R.1` Verify
+- **M** Metrics: `M.0`
+- **O** Operations: `O.0`
+
+## Agent Roles
+
+When activated via slash commands, adopt the specific agent persona:
+- **Help Agent**: Navigator — đọc `.flow-state.yaml`, gợi ý bước tiếp theo dựa trên trạng thái pipeline
+- **Understand Agent**: Multi-Layer Scanner — generate context.md (Tầng 1) + Module Cards (Tầng 2), đánh dấu `[⚠️ AI-inferred]`; tự kiểm card bằng CLI: `v-flow validate --module-cards <m>` (lint) + `score-card <m>` (điểm objective) + `dev-quiz <m>` (answer key từ graph)
+- **Import BA Docs Agent**: BA Docs Importer — trích xuất tài liệu BA từ NotebookLM, generate raw materials + AI drafts cho Module Cards
+- **Constitute Agent**: Constitution Drafter — infer từ code, để trống rõ ràng, không lý tưởng hóa
+- **BA Agent**: Business Analyst — focus on WHAT, not HOW; init `.flow-state.yaml` + `_session.md`. Dual BA mode: cập nhật spec sau mỗi round review. Auto-scan ambiguity và gợi ý `/v.prototype`
+- **BA Critic Agent**: Adversarial Spec Critic — fresh context mỗi round, phải tìm issues, không LGTM
+- **Prototype Agent**: Throwaway Builder — tạo LOGIC TUI hoặc UI variants để resolve ambiguity, capture verdict vào `prototype-notes.md`, carry-forward qua `_session.md`
+- **Plan Agent**: Technical Architect — design, don't implement; update `_session.md`
+- **Test Agent**: QA Engineer — comprehensive coverage
+- **Master Check Agent**: Quality Gate — strict, 3-level verdict (PASS/PASS WITH CONCERNS/FAIL), explicit `fail_reason` + `route_decision`
+- **Implement Agent**: Developer — TDD, follow the plan exactly, resolve all concerns [W-xxx]
+- **Bugfix Analyst Agent**: Bug Diagnostician — 3-behavior model (Current/Expected/Unchanged), root cause analysis, 3-tier test plan (Red/Green/Protect). Kiro-inspired surgical fix approach
+- **Review Agent**: Adversarial Reviewer — find ≥3 issues, never "LGTM", verify concerns resolved
+- **Status Agent**: Observer — cross-check `.flow-state.yaml` vs reality, render visual dashboard
+- **Sync Agent**: Synchronizer — propagate spec changes to plan/tasks/tests
+- **Sprint Planning Agent**: Scrum Master — create sprint-status.md and track dependencies
+- **Retrospective Agent**: Evaluator — tổng kết và cập nhật knowledge base sau epic/sprint
+- **Metrics Agent**: Analyst — đo lường performance và đề xuất cải tiến Agent Config
+- **Operations Agent**: SRE — monitor production, detect incidents, tạo báo cáo và trigger pipeline
+
+## New Concepts (LangGraph-inspired)
+
+| Concept | File | Purpose |
+|---------|------|---------|
+| **Auto-Inject Context** | `AGENTS.md` / `.cursorrules` | Tự động bơm Tầng 1 (`context.md` + `constitution.md`) cho GitHub Copilot, Antigravity & Cursor đọc trực tiếp |
+| **Multi-Layer Understanding** | `.v-flow/modules/` | Tầng 1: Auto-injected Overview. Tầng 2: Module Cards on-demand. Tầng 3: Code Graph |
+| **Typed State** | `.flow-state.yaml` | Machine-readable pipeline state — phase, artifacts, routing, concerns |
+| **Routing Rules** | Embedded in each workflow | Explicit next-step decision — no guessing |
+| **Working Memory** | `_session.md` | Short-term context: decisions, trace, open questions |
+| **Concerns Carry-forward** | `validation-report.md` → `_session.md` | Concerns from R.0 tracked through B.1 and verified at R.1 |
+| **Prebuilt Flows** | `quickfix`, `bugfix`, `hotfix` | Optimized pipelines for common scenarios (3 tiers: simple → structured → emergency) |
+| **Spec Change Policy** | `constitution.md` §9 | Double-texting: Enqueue/Interrupt/Rollback/Reject based on phase |
+| **Map-Reduce** | `/v.plan --with-*` | Optional agents (Security, UX, API) run in parallel at B.0 |
+| **Thread Isolation** | `/v.sprint` | Multi-feature dashboard with dependency tracking |
+| **Task Annotations** | `tasks.md` | `[ENTRYPOINT]`, `[PARALLEL]`, `[GATE]`, `[CRITICAL]`, `[RISKY]`, `[SPIKE]`, `[CLARIFY]`, `[DEPENDS: T-xxx]` scheduling |
+| **Streaming** | `_session.md` Live Progress | Real-time progress updates during agent execution |
+| **Time Travel** | `/v.rewind`, `/v.fork` | Rollback to previous versions, branch to explore alternatives |
+| **Command Pattern** | `agent-command-template.yaml` | Structured agent output = state_update + routing in one block |
+| **Subgraph Composition** | `module-template.yaml` | Reusable pipeline modules with encapsulated input/output |
+
+<!-- gitnexus:start -->
+# GitNexus — Code Intelligence
+
+This project is indexed by GitNexus as **ADLC** (14006 symbols, 15182 relationships, 70 execution flows). Use the GitNexus MCP tools to understand code, assess impact, and navigate safely.
+
+> If any GitNexus tool warns the index is stale, run `npx gitnexus analyze` in terminal first.
+
+## Always Do
+
+- **MUST run impact analysis before editing any symbol.** Before modifying a function, class, or method, run `gitnexus_impact({target: "symbolName", direction: "upstream"})` and report the blast radius (direct callers, affected processes, risk level) to the user.
+- **MUST run `gitnexus_detect_changes()` before committing** to verify your changes only affect expected symbols and execution flows.
+- **MUST warn the user** if impact analysis returns HIGH or CRITICAL risk before proceeding with edits.
+- When exploring unfamiliar code, use `gitnexus_query({query: "concept"})` to find execution flows instead of grepping. It returns process-grouped results ranked by relevance.
+- When you need full context on a specific symbol — callers, callees, which execution flows it participates in — use `gitnexus_context({name: "symbolName"})`.
+
+## Never Do
+
+- NEVER edit a function, class, or method without first running `gitnexus_impact` on it.
+- NEVER ignore HIGH or CRITICAL risk warnings from impact analysis.
+- NEVER rename symbols with find-and-replace — use `gitnexus_rename` which understands the call graph.
+- NEVER commit changes without running `gitnexus_detect_changes()` to check affected scope.
+
+## Resources
+
+| Resource | Use for |
+|----------|---------|
+| `gitnexus://repo/ADLC/context` | Codebase overview, check index freshness |
+| `gitnexus://repo/ADLC/clusters` | All functional areas |
+| `gitnexus://repo/ADLC/processes` | All execution flows |
+| `gitnexus://repo/ADLC/process/{name}` | Step-by-step execution trace |
+
+## CLI
+
+| Task | Read this skill file |
+|------|---------------------|
+| Understand architecture / "How does X work?" | `.claude/skills/gitnexus/gitnexus-exploring/SKILL.md` |
+| Blast radius / "What breaks if I change X?" | `.claude/skills/gitnexus/gitnexus-impact-analysis/SKILL.md` |
+| Trace bugs / "Why is X failing?" | `.claude/skills/gitnexus/gitnexus-debugging/SKILL.md` |
+| Rename / extract / split / refactor | `.claude/skills/gitnexus/gitnexus-refactoring/SKILL.md` |
+| Tools, resources, schema reference | `.claude/skills/gitnexus/gitnexus-guide/SKILL.md` |
+| Index, status, clean, wiki CLI commands | `.claude/skills/gitnexus/gitnexus-cli/SKILL.md` |
+
+<!-- gitnexus:end -->

package/CHANGELOG.md

@@ -0,0 +1,318 @@
+# Changelog
+
+All notable changes to V-Flow 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).
+
+---
+
+## [1.5.0] — 2026-06-30
+
+### Added — BA → PTYC doc export (`/v.ba-to-ptyc`)
+- **`/v.ba-to-ptyc`** (agent `ba-to-ptyc-agent`, phase **S.1.6**, optional) — sinh tài liệu **Phân tích Yêu cầu (PTYC.docx)** chuẩn ISO từ `spec.md`. Cơ chế **Route B**: đổ dữ liệu vào template `.docx` có sẵn → giữ nguyên 100% format. Tương ứng bước 7a trong BA Skill Pipeline (BABOK v3).
+  - Template `templates/ptyc/PTYC.template.docx` (+ `ptyc.meta.example.yaml`); scripts `scripts/ptyc/` (`build_template.py`, `render_ptyc.py`, `render_mermaid.mjs` graceful, `ptyc-data.schema.json`); sample fixture `scripts/ptyc/sample/`.
+  - Map spec→PTYC: suy luận C/R/U/D, tách màn hình mức control, rút glossary; sơ đồ mermaid→PNG (placeholder nếu thiếu `mmdc`).
+  - Skills copy vào `skills/specify/ba-doc-generator`, `skills/specify/ba-bpmn-doc-gen`.
+
+### Added — Activity logging & multi-provider hooks
+> Ghi lại toàn bộ hoạt động "dùng v-flow" về **một nhật ký cục bộ** `.v-flow/logs/` (`activity.jsonl` cho máy đọc + `activity.log` cho người), gom cả CLI lẫn các tool AI. Best-effort tuyệt đối — log không bao giờ làm vỡ CLI hay chặn agent. Tắt bằng `VFLOW_NO_LOG=1`; đổi đích bằng `VFLOW_LOG_DIR`/`VFLOW_LOG_FILE`.
+- **CLI auto-log** (`cli/lib/activity-log.js`, wired ở `cli/index.js`): mỗi lần chạy `v-flow <lệnh>` ghi 1 record (cmd, args, exit code, duration, cwd, version) + toàn bộ transcript console. Append đồng bộ → sống sót qua `process.exit`/crash. Không tự tạo `.v-flow/` (tránh làm hỏng `isInitialized`).
+- **`v-flow log`** (`cli/commands/log.js`): agent/người ghi event có cấu trúc (`--phase`, `--feature`, `--actor`, `--event`, `--data <json>`); `--tail [n]` xem lại. Protocol cho agent ở `AGENTS.md` (mục "Activity Logging", rule 17) — ghi tại ranh giới phase, routing decision, lỗi/blocker.
+- **Lifecycle hooks đa-provider** (`cli/commands/hooks.js`, `cli/lib/hook-targets.js`): `v-flow hooks install | status | uninstall` rải/gỡ hook logging cho **Claude Code, Cursor, GitHub Copilot, OpenAI Codex, Antigravity** — merge vào config sẵn có, **không clobber**, idempotent. Mỗi hook gọi `v-flow log --hook <provider>`: đọc payload qua stdin, chuẩn hoá đa-schema, append vào cùng log; **luôn exit 0**, không chặn agent (Antigravity nhận `{"allow_tool":true}`). Cờ `--abs` sinh đường dẫn node+CLI tuyệt đối cho IDE GUI không kế thừa PATH.
+- **`v-flow watch`** (`cli/commands/watch.js`): daemon giám sát `.v-flow/**`, log mọi create/modify/delete — safety-net **tool-agnostic** (bắt được cả tool không có hook lẫn sửa tay). Watcher đệ quy thủ công (chạy mọi OS trên Node ≥18), loại trừ `logs`/`templates`/`skills`, tự thêm watcher cho thư mục mới. `--timeout <giây>` để chạy có giới hạn.
+
+### Notes
+- `.v-flow/logs/` được thêm vào `.gitignore` (logs cục bộ); ngược lại các file config hook (`.claude/settings.json`, `.cursor/hooks.json`, …) nên commit (portable nếu `v-flow` trên PATH).
+- 416/416 test xanh (toàn bộ là code mới + wiring additive, không sửa logic sẵn có).
+
+---
+
+## [1.4.0] — 2026-06-16
+
+### Added — Plan/Implement quality patterns (học từ obra/superpowers)
+> So sánh bộ skill `superpower/` (obra/superpowers, vendored) với hệ plan/implement của V-Flow → tiếp thu các pattern *bổ sung thật sự*, giữ nguyên xương sống governance của V-Flow. Schema **không đổi** (`schema_version` vẫn `2.0`); đây là nâng cấp *nội dung skill/agent*, backward-compatible.
+- **Self-Review "mắt mới"** (`skills/plan/architecture-designer/SKILL.md`, `skills/plan/task-breakdown/SKILL.md`): checklist tự chạy ngay sau khi viết plan/tasks — phủ spec/AC **hai chiều** (thiếu *và* thừa/over-engineering), quét placeholder, **nhất quán symbol/naming giữa các section/task** (`clearLayers()` vs `clearFullLayers()` = bug), DAG hợp lệ. Là self-check nội tuyến (KHÔNG subagent, KHÔNG đẩy validator sang semantics — giữ phân vai validator=structural · Master Check=semantic).
+- **Iron Law TDD + bảng đập-ngụy-biện** (`skills/implement/tdd-driver/SKILL.md`): "không code production khi chưa có test FAIL trước; lỡ viết code trước → xóa, làm lại"; 7 dòng phản biện các ngụy biện skip-TDD thường gặp.
+- **Test anti-patterns** (tdd-driver §4): đừng assert lên mock, không method test-only trong production (→ `test-utils/`), mock ở tầng thấp nhất sau khi hiểu deps, mirror đầy đủ cấu trúc API thật.
+- **Spec-fit per-task (đúng & đủ, KHÔNG thừa)** (`agents/implement-agent.md` Bước 2/4 + rule 13; signal mới trong tdd-driver): kiểm **hai chiều** sau mỗi task — không thiếu AC *và* không over-build (thêm field/flag/endpoint/abstraction không ai yêu cầu → gỡ, theo YAGNI). Áp per-task (nhất là `[GATE]`/`[CRITICAL]`) để bắt drift sớm thay vì dồn hết về `/v.review`.
+
+### Fixed
+- **Gold example lệch schema**: `skills/plan/task-breakdown/examples/good/tasks-login-feature.md` thiếu `Done khi (AC-NN)` + block **TDD** (có từ schema 2026-06-02) → bổ sung cho 6 task, khớp lại template + sibling `spike-clarify-tasks.md`.
+
+### Notes
+- **Cố tình KHÔNG tiếp thu** (ghi để tránh "sửa nhầm" sau): substrate *controller-spawn-subagent-mỗi-task* của obra (xung đột SSoT + flow-state + enforcement script + GitNexus); *code-in-plan* (V-Flow giữ plan là thiết kế, không phải code); symbol-consistency giữ ở tầng *self-review*, không đẩy thành rule validator.
+- 402/402 test xanh (chỉ sửa markdown, không đụng JS).
+
+---
+
+## [1.3.0] — 2026-06-02
+
+### Added — Plan/Task/Spec subsystem upgrade (v3)
+- **Deep cross-file validator** (`cli/commands/validate.js`): kiểm tra liên file Spec ↔ Plan ↔ Tasks ↔ Tests theo ID (REQ / AC-NN / T-NNN / TC-xx):
+  - Traceability hai chiều REQ-xxx + phủ AC-NN (mọi AC phải có ≥1 task).
+  - Dependency-graph integrity: phát hiện chu trình, `[DEPENDS: T-xxx]` trỏ task không tồn tại, `[PARALLEL]`/`[ENTRYPOINT]` không hợp lệ.
+  - Per-task completeness: bắt buộc block **TDD (Red→Green)**, **Complexity (S/M/L)**, **Done khi (AC)**, Task Type.
+  - **Phân vai rõ**: validator = *structural lint*; Master Check Agent (R.0) = *semantic*; human = *judgment*.
+- **TDD baked into schema**: mỗi task trong `tasks-template.md` có checklist Red→Green — viết test fail trước, code tối thiểu cho pass.
+- **Acceptance-Criteria IDs (AC-NN)**: spec §4 đánh mã AC; plan/tasks/master-check truy vết theo AC.
+- **Technical Decisions table** (plan-template §13): ghi WHY + phương án đã loại + đánh đổi.
+- **Schema versioning** cho `.flow-state.yaml`: `schema_version: "2.0"` + `pipeline_type` (full/quickfix/hotfix → "lite" profile). `v-flow feature create` nay **stamp** cả hai field → feature mới được giữ đúng schema hiện hành (trước đây bị coi là legacy → bỏ qua mọi gate v3).
+- **Templates mới**: `data-model-template.md`, `api-contract-template.md`, `ux-checklist-template.md`, `validation-report-template.md`.
+- **Operations Agent** (`agents/operations-agent.md`): hiện thực hóa role O.0 (trước đây chỉ inline trong `/v.hotfix`) — wiring skill `incident-runbook` + `operations-report-template.md`.
+
+### Fixed
+- **No-Go Zone guard fail-open** (`skills/implement/no-go-zone-guard/scripts/check-nogo-zones.sh`): `head -n -1` (GNU-only) lỗi trên macOS/BSD → guard báo PASS cho file 🔴 Strict. Đổi sang `sed '$d'` + `grep -qF` (an toàn regex).
+- **Template tự-lint sạch**: `plan-template.md` / `tasks-template.md` dùng ID cụ thể (`REQ-U01`…) thay placeholder `REQ-xxx` (regex validator yêu cầu `REQ-[A-Z]\d+`). Validator coverage-matrix chấp nhận heading "Test Matrix" / "Ma trận truy vết".
+- **CI/coverage**: ngưỡng `jest.config.js` đặt sát thực tế (regression floor) để `npm test` không fail giả; README/CHANGELOG bổ sung `/v.prototype`, `/v.bugfix` và cập nhật số lượng agent/command.
+
+---
+
+## [1.2.1] — 2026-05-22
+
+### Fixed
+- **`v-flow init` / `v-flow sync` — Full template sync** (`cli/lib/sync-core.js`):
+  - Trước đây chỉ copy `templates/module-card/` vào `.v-flow/templates/module-card/`
+  - Nay copy **toàn bộ thư mục `templates/`** vào `.v-flow/templates/` (25 templates + subdirectories)
+  - Agents trong project đích có thể đọc tất cả template (spec, plan, tasks, constitution, v.v.) từ `.v-flow/templates/`
+  - Guard `isSourceRoot` vẫn được giữ để tránh self-copy khi chạy trên thư mục v-flow
+- **`v-flow upgrade` — Full template upgrade** (`cli/commands/upgrade.js`):
+  - Trước đây loop thủ công chỉ xử lý `.md`/`.yaml` ở root, bỏ sót thư mục con (`knowledge/`, `module-card/`)
+  - Nay dùng `fs.copy` đệ quy — nhất quán với `sync-core.js`, cover toàn bộ subtree
+  - Code giảm từ ~70 dòng xuống còn ~30 dòng
+- **`v-flow sync --apply` — Full template sync** (`cli/commands/sync.js`):
+  - `sync.js` có block riêng hardcode chỉ copy `module-card/`, không hưởng fix từ `sync-core.js`
+  - Nay copy toàn bộ `templates/` đệ quy vào `.v-flow/templates/`
+  - Section header đổi từ `── Module Card Templates ──` → `── Templates ──`
+- **`v-flow init` — Log xác nhận** (`cli/commands/init.js`):
+  - Thêm dòng `✅ Templates: .v-flow/templates/ (all templates synced)` trong output
+- **`v-flow doctor` — Templates health check** (`cli/commands/doctor.js`):
+  - Trước đây chỉ check `module-card/` (8 files)
+  - Nay check toàn bộ `.v-flow/templates/` — hiện số file, verify `module-card/` & `knowledge/` subdirs
+- **`v-flow sync` — Dry-run template stats** (`cli/commands/sync.js`):
+  - Template changes không được tính vào `stats.added/synced` → summary thiếu số liệu
+  - Nay đếm chính xác new vs existing files, cộng vào stats đúng cách
+  - Dry-run hiện thị `(would sync)` thay vì `→ .v-flow/templates/`
+- **`v-flow upgrade` — Template stats chính xác** (`cli/commands/upgrade.js`):
+  - `stats.updated` tính sai (= `beforeCount` thay vì số file thực sự thay đổi)
+  - Nay dùng `walk()` so sánh từng file để tính đúng added vs updated
+  - Output chi tiết hơn: `✨ N new`, `🔄 N updated`, `⏭️  up-to-date`
+
+---
+
+## [1.2.0] — 2026-04-28
+
+### Added
+- **`cli/lib/logger.js`** — Structured logging module with verbosity levels (SILENT/NORMAL/VERBOSE/DEBUG):
+  - JSON output mode for machine-readable output (`--json` support)
+  - Elapsed time tracking and structured data buffering
+  - Test suite: `tests/lib/logger.test.js`
+- **Integration test suite** (`tests/integration/integration.test.js`):
+  - End-to-end CLI workflow tests: init→check, feature lifecycle, doctor diagnostics
+  - Plugin system integration tests, i18n tests, CLI basics (--version, --help)
+- **Plugin command tests** (`tests/commands/plugin.test.js`):
+  - Tests for scaffolding, discovery, validation, loading, and installation of all 4 plugin types
+- **Plugin lifecycle hooks** — `executeLifecycleHook()` and `uninstallPlugin()` in plugins.js:
+  - `onInstall`/`onUninstall` hooks via `plugin.json` manifest
+  - Script-based hooks (`.js` files) or message-based hooks
+- **YAML frontmatter** on all 16 agent definitions:
+  - `output_schema` — machine-enforceable output format for each agent
+  - `memory_logging` — standardized session logging configuration
+  - Agent metadata: name, role, phase, triggers, output
+- **Learning module 08** — Plugin Development & Extension Patterns:
+  - Plugin creation, manifest configuration, composition patterns
+  - Config hooks, anti-patterns guide, 3 hands-on exercises
+- **`v-flow.en.md`** — Full English translation of the V-Flow specification v4.2:
+  - Complete 1:1 translation of all 8 sections (~750 lines)
+  - Cross-reference links between Vietnamese and English versions
+  - Enables international adoption of V-Flow methodology
+- **Constitution Lint** (`cli/lib/constitution-lint.js`):
+  - Parser extracts `[LINT:xxx]` tags from constitution.md
+  - Rule engine checks: naming, import, structure, error, test, security, doc (10 categories)
+  - 3 severity levels: warning, error, critical (blocks CI)
+  - Backward-compatible: constitutions without tags return "0 rules, skip"
+  - Integrated into `validate` command with `--constitution` flag
+  - Test suite: 32 tests (`tests/lib/constitution-lint.test.js`)
+- **Learning module 09** — Constitution Lint guide:
+  - Tag format, categories, CI integration, best practices, 3 exercises
+
+### Changed
+- **i18n adoption — 9/9 commands** (was 0/9):
+  - All commands now use `createT()` for user-facing strings
+  - Translation keys expanded: 63 → 200+ (vi + en)
+  - Commands integrated: init, check, status, doctor, sync, upgrade, validate, plugin, feature
+- **ESLint enhanced** — 7 new code quality rules:
+  - `no-lonely-if`, `prefer-template`, `object-shorthand`, `no-else-return`
+  - `consistent-return`, `no-nested-ternary`, `require-await`
+  - Test-specific overrides for flexibility
+- **Coverage thresholds raised**:
+  - lib/: 60/80/60/60 → 65/85/70/70 (branches/functions/lines/statements)
+  - commands/: 30/40/30/30 → 40/50/40/40
+
+### Added
+- **Skills Ecosystem — 5 new `config.json` files** (12/17 → 17/17 config hooks):
+  - `_shared/constitution-reader/config.json` — Section mapping per phase, version checking, extraction depth
+  - `_shared/context-reader/config.json` — Staleness thresholds, No-Go Zone enforcement, cross-reference
+  - `_shared/ears-notation/config.json` — EARS strict mode, ambiguous word detection, NFR enforcement
+  - `understand/codebase-scanner/config.json` — Scan scope, file size limits, framework detection
+  - `understand/ba-docs-scanner/config.json` — Source types, auto glossary, conflict resolution
+- **Skills Ecosystem — 3 new `examples/` directories** (12/17 → 17/17 examples):
+  - `implement/constitution-enforcer/examples/` — Good (compliance report) + Bad (vague report)
+  - `metrics/metrics-collector/examples/` — Good (full metrics with trends) + Bad (incomplete)
+  - `operations/incident-runbook/examples/` — Good (P1 hotfix response) + Bad (vague incident)
+- **Workflow System — `decision-matrix.md`** — New workflow for selecting the right workflow:
+  - Decision tree diagram (30-second selection)
+  - Comprehensive comparison table (9 criteria × 6 workflows)
+  - Prerequisite checks (universal + legacy-specific)
+  - Error recovery paths (5 common scenarios with step-by-step resolution)
+  - Workflow composition patterns (5 patterns: sequential, escalation, parallel, incident→feature, multi-feature)
+  - Cross-workflow navigation map
+  - Anti-patterns guide
+
+### Changed
+- **All 10 workflows** enhanced with standardized sections:
+  - ⚠️ Error Recovery table — 5-6 scenarios per workflow with step-by-step recovery
+  - 🗺️ Cross-Workflow Navigation — 5-6 navigation links per workflow
+  - Enhanced: `full-flow.md`, `quick-feature.md`, `ba-spec.md`, `bug-fix.md`, `hotfix.md`, `code-review.md`, `testing.md`, `sprint-planning.md`, `spec-change.md`, `legacy-onboard.md`
+
+
+## [1.1.0] — 2026-04-27
+
+### Added
+- **`cli/lib/yaml-parser.js`** — Shared YAML parser using `js-yaml` with regex fallback
+  - Replaces fragile regex-based `.flow-state.yaml` parsing in status.js, doctor.js, feature.js
+  - `parseFlowState()`, `normalizeFlowState()`, `readFlowState()` exported for all commands
+  - Comprehensive test suite (`tests/lib/yaml-parser.test.js`)
+- **3 new Skills** (14 → 17 total), completing the blueprint:
+  - `implement/constitution-enforcer/` — Validates code against constitution.md rules (B.1, R.1)
+  - `metrics/metrics-collector/` — Structured agent performance metrics collection (M.0)
+  - `operations/incident-runbook/` — Production incident response procedure (O.0)
+- **`.npmignore`** — Prevents publishing test/docs/coverage/examples to npm
+- **`.eslintrc.json`** — ESLint configuration (eslint:recommended + custom rules)
+- **`.github/dependabot.yml`** — Automated dependency updates (weekly npm, monthly Actions)
+- **Example 005** — Dual BA Review flow (spec.md, spec-review-log.md, .flow-state.yaml)
+- **`--verbose` global flag** — All commands can access via `process.env.VFLOW_VERBOSE`
+- **`--lang` global flag** — Integrates i18n system via preAction hook in `cli/index.js`
+- **New command tests** — doctor.test.js, init.test.js, feature.test.js, status-upgrade.test.js
+- **`package.json` improvements**:
+  - `"engines": { "node": ">=18" }` matching CI matrix
+  - `"files"` whitelist for npm distribution
+  - `"prepublishOnly": "npm run lint && npm test"` for publish validation
+  - `js-yaml` as direct dependency
+  - `eslint` as dev dependency
+  - `"lint"` and `"lint:fix"` scripts
+
+### Changed
+- **All 9 CLI commands** now use `withErrorHandler` for unified error handling:
+  - Migrated: init.js, check.js, status.js, doctor.js, sync.js, upgrade.js, feature.js (create + list)
+  - Previously only validate.js and plugin.js used it
+- **status.js** refactored to use shared `yaml-parser.js` (removed 60 lines of regex parsing)
+- **doctor.js** refactored to use shared `yaml-parser.js` and import `formatBytes` from `cli-error.js`
+- **feature.js** refactored to use shared `yaml-parser.js` for flow state reading
+- **check.js** now imports `formatBytes` from `cli-error.js` instead of local `formatFileSize()`
+- **v-flow.md** updated to v4.2 — added Section 8 (CLI Tool, Plugin System, i18n, Sync), updated directory structure with `plugins/` and `templates/` folders
+- **CI/CD** enhanced:
+  - Added lint step (ESLint) before test execution
+  - Coverage regression check with report
+  - npm auto-publish on tag push (requires NPM_TOKEN secret)
+- **jest.config.js** — Added coverage threshold for `commands/` (30% floor as baseline)
+- **learn/README.md** — Added hands-on exercises column, linked all 5 examples
+- **i18n** status clarified — marked as "available, opt-in" with integration guide in header
+
+### Fixed
+- **Circular dependency** in `sync-core.js` — no longer requires `templates.js` at runtime
+  - Changed to inline `path.join(projectRoot, '.v-flow')` instead of `getVFlowDir()`
+- **Duplicate `formatBytes()`** removed from `doctor.js` and `check.js` (was `formatFileSize()`)
+
+### Removed
+- `docs/temp_docx/` — Temporary artifacts removed from repository
+
+### Previous Changes
+
+
+- `CHANGELOG.md` — tracking all notable changes
+- Test suite with Jest for CLI commands and core libraries (135 tests, 8 suites)
+- Comprehensive `.gitignore` covering OS files, editors, env, build outputs, coverage
+- **5 new Skills** (9 → 14 total) following standardized blueprint:
+  - `specify/ba-critic/` — Adversarial spec review with 5 critique dimensions (Ambiguity, Completeness, Consistency, Feasibility, Testability)
+  - `plan/architecture-designer/` — Architecture design from spec with constraint extraction & No-Go Zone validation
+  - `plan/task-breakdown/` — Task decomposition with annotation system ([ENTRYPOINT], [GATE], [PARALLEL], [RISKY], [DEPENDS])
+  - `test/test-case-generator/` — Test case generation with REQ-type-to-strategy mapping & 7-category edge cases
+  - `check/cross-validator/` — 3-phase validation engine (Traceability, Consistency, Quality) with concern carry-forward
+- **2 new Examples**: 003-quickfix-bug, 004-hotfix-p1
+- `roadmap/roadmap.md` — Product development roadmap v1.1 → v2.0
+- **`v-flow validate`** command — Validates artifact structure (EARS, REQ-xxx, traceability, annotations)
+- **CLI Architecture Refactor** (Phase 3 — Quality):
+  - `cli/lib/scanner.js` — Extracted 400+ lines from monolithic `templates.js`
+  - `cli/lib/cli-error.js` — Unified error handling: typed errors, `withErrorHandler`, guard utilities
+  - `templates.js` reduced from 600 → 250 lines via delegation
+- **Extended Scanner** — New language/framework support:
+  - Dart/Flutter (pubspec.yaml reader, Riverpod/BLoC/Provider/GetX detection)
+  - Rust (Cargo.toml reader, Actix/Axum/Rocket/Tokio detection)
+  - Go (go.mod reader, Gin/Echo/Fiber detection)
+  - Docker, .NET (ASP.NET Core, Entity Framework), Bun, Poetry, Gradle detection
+  - 50+ framework patterns (was ~25), 16 package managers (was 7)
+- **Plugin/Extension Architecture** (Phase 4 — Scale):
+  - `cli/lib/plugins.js` — Plugin discovery, validation, loading, scaffolding, installation
+  - `v-flow plugin` command — 4 subcommands: `list`, `create`, `install`, `validate`
+  - Plugin manifest format (`plugin.json`) with 4 types: agent, workflow, skill, composite
+  - Auto-install to all 3 targets (Antigravity, Copilot, Cursor)
+- **i18n System** (`cli/lib/i18n.js`):
+  - Vietnamese (default) and English translations
+  - Language detection: `--lang` flag, `VFLOW_LANG` env, system locale
+  - Template variable interpolation (`{key}` syntax)
+- **GitHub Actions CI/CD** (`.github/workflows/ci.yml`):
+  - Test matrix (Node 18/20/22)
+  - Artifact validation via `v-flow doctor`
+  - Automatic GitHub Release from git tags with changelog extraction
+- **Bilingual README** — English section + Vietnamese section in single README.md
+- **Knowledge Base Templates**:
+  - `adr-template.md` — Architecture Decision Record with V-Flow compliance checks
+  - `golden-tests-readme.md` — Golden test guide with file conventions and CI integration
+  - `api-contract-template.md` — API contract with data models, SLA, versioning policy
+
+### Fixed
+- `v-flow upgrade` command now properly syncs agents & commands via `runSync()` instead of printing a manual hint
+- `.gitignore` missing common patterns (`.DS_Store`, `.env`, `coverage/`, etc.)
+- Deprecated scripts moved to `scripts/_deprecated/` with deprecation notice
+
+---
+
+## [1.0.0] — 2026-04-23
+
+### Added
+- **Core Process**: V-Flow v4.1 specification (`v-flow.md`)
+  - 6 phase groups: U (Understand) → S (Specify) → B (Build) → R (Review) → M (Metrics) → O (Operations)
+  - 16 specialized AI agents with explicit routing and state management
+  - 3-level Master Check verdict: PASS / PASS WITH CONCERNS / FAIL
+  - Adversarial Code Review requiring ≥3 issues
+  - Dual BA spec critique loop (opt-in `--dual`)
+  - Time Travel: `/v.rewind` and `/v.fork`
+  - Typed State via `.flow-state.yaml`
+
+- **CLI Tool** (`v-flow-cli` v1.0.0)
+  - `v-flow init` — initialize `.v-flow/` with auto-scan codebase
+  - `v-flow check` — quick setup verification
+  - `v-flow status` — pipeline progress dashboard with visual bars
+  - `v-flow feature create` — scaffold feature directories
+  - `v-flow sync` — hash-based sync to 3 AI agent targets
+  - `v-flow doctor` — deep health check with fix capability
+  - `v-flow upgrade` — template and agent upgrade
+
+- **21 Slash Commands**: `/v.help`, `/v.understand`, `/v.specify`, `/v.plan`, `/v.test`, `/v.check`, `/v.implement`, `/v.review`, `/v.sync`, `/v.sprint`, `/v.metrics`, `/v.status`, `/v.quickfix`, `/v.hotfix`, `/v.rewind`, `/v.fork`, `/v.constitute`, `/v.constitution`, `/v.ba-critic`, `/v.retrospective`, `/v.import-ba-docs`
+
+- **16 Agent Definitions**: BA Agent, BA Critic, Constitute, Help, Implement, Import BA Docs, Master Check, Metrics, Plan, Retrospective, Review, Sprint, Status, Sync, Test, Understand
+
+- **10 Workflows**: Full Flow, Quick Feature, Bug Fix, Hotfix, BA Spec (Dual), Testing, Code Review, Spec Change, Sprint Planning, Legacy Onboard
+
+- **23 Templates**: Constitution, context, spec, plan, tasks, test-cases, validation-report, flow-state, session, security-review, agent-config, agent-command, module, metrics, operations-report, decision-log, code-review-report, sprint-status, spec-review-log, retrospective-report + 9 module-card templates
+
+- **9 Skills** (5 groups): `_shared/` (constitution-reader, context-reader, ears-notation), `implement/` (no-go-zone-guard, tdd-driver), `review/` (adversarial-reviewer), `specify/` (ears-writer), `understand/` (codebase-scanner, ba-docs-scanner)
+
+- **3-target AI Agent Sync**: Antigravity (`.agents/`), GitHub Copilot (`.github/`), Cursor (`.cursor/`)
+
+- **2 Examples**: 001-user-login (Greenfield Golden), 002-legacy-crud (Legacy Refactor)
+
+- **7-lesson Learning Path**: Introduction → Core Concepts → Agent Roles → Standard Pipeline → Advanced Workflows → Critical Rules → Legacy Implementation
+
+- **Documentation**: Best Practices, EARS Notation Guide, Agent Setup Guide, Agent Metrics Guide, Model Selection Guide, Operations Agent Guide, Skills Blueprint, Spec-to-Task Automation

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 viettelcontruction jsc.
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.