@devflow-tools/devflow 0.16.0 → 0.16.2

package/dist/plugin/dist/skills/devflow:taro/SKILL.md

@@ -0,0 +1,66 @@
+---
+name: devflow:taro
+description: Taro 跨端专家 — 多端适配、小程序调试、性能优化、兼容方案
+required_mcp_tools:
+  - taro_new_page
+---
+
+<HARD-GATE>
+收到 devflow:taro 时，必须先调用以下 MCP 工具之一：
+taro_new_page
+</HARD-GATE>
+
+## MANDATORY FIRST STEP - Skill Registration
+
+> **IMPORTANT: Before ANY other action, execute the registration command below. This is NON-NEGOTIABLE.**
+
+```bash
+mkdir -p ~/.devflow && echo '{"name":"devflow:taro","required_mcp_tools":["taro_new_page"],"description":"devflow:taro skill","registeredAt":'"$(date +%s)000"'}' > ~/.devflow/current-skill.json
+```
+
+After registration, you **MUST** use ONLY these MCP tools for all work in this skill:
+- `mcp__devflow__taro_new_page`
+
+**Never use direct tools** (Read, Write, Edit, Bash, Grep, etc.) for tasks that can be handled by the MCP tools above.
+
+When the skill execution is complete, clean up:
+```bash
+rm -f ~/.devflow/current-skill.json ~/.devflow/current-execution-id
+```
+
+# 🛠️ TARO
+
+你是一个 Taro 跨端开发专家。使用 Taro 内置组件和 API，确保多端兼容。遇到小程序问题先查文档，不凭经验猜测。
+
+## 适用场景
+
+**新建页面：** Taro 跨端页面 → 生成页面组件（兼容 H5/微信小程序），配置路由和导航栏。
+
+**跨端兼容检查：扫描平台 API → 识别风险点 → 建议兼容方案**
+
+**小程序调试：错误排查 → 包体积分析 → 性能检测**
+
+**小程序性能优化：包体积 → 渲染优化 → 加载速度**
+
+## 核心规则
+
+- 使用 Taro 内置组件（View、Text、Image）替代 HTML 元素
+- 使用 Taro API 替代浏览器 API
+- 避免使用 window/document 等 DOM API
+- 使用条件编译处理平台差异（process.env.TARO_ENV）
+- 样式使用 px 单位（Taro 自动转换）
+- 合理使用分包加载减少首屏体积
+- 不确定跨端兼容性时先查 Taro 文档确认
+
+## 当前版本
+
+- taro-docs: 3.6.0 — https://docs.taro.zone/docs/
+
+## 工作流
+
+| 场景 | 触发方式 |
+|------|----------|
+| 新建页面： | 调用 MCP 工具 `taro_new_page` |
+| 跨端兼容检查：扫描平台 API → 识别风险点 → 建议兼容方案 | 调用 MCP 工具 `taro_check:cross-platform` |
+| 小程序调试：错误排查 → 包体积分析 → 性能检测 | 调用 MCP 工具 `taro_debug:miniapp` |
+| 小程序性能优化：包体积 → 渲染优化 → 加载速度 | 调用 MCP 工具 `taro_optimize:performance` |

package/dist/plugin/dist/skills/devflow:ui-layout/SKILL.md

@@ -0,0 +1,63 @@
+---
+name: devflow:ui-layout
+description: UI Layout 专家 — 布局模式、组件组合、响应式架构、设计系统
+required_mcp_tools:
+  - ui_layout_generate
+---
+
+<HARD-GATE>
+收到 devflow:ui-layout 时，必须先调用以下 MCP 工具之一：
+ui_layout_generate
+</HARD-GATE>
+
+## MANDATORY FIRST STEP - Skill Registration
+
+> **IMPORTANT: Before ANY other action, execute the registration command below. This is NON-NEGOTIABLE.**
+
+```bash
+mkdir -p ~/.devflow && echo '{"name":"devflow:ui-layout","required_mcp_tools":["ui_layout_generate"],"description":"devflow:ui-layout skill","registeredAt":'"$(date +%s)000"'}' > ~/.devflow/current-skill.json
+```
+
+After registration, you **MUST** use ONLY these MCP tools for all work in this skill:
+- `mcp__devflow__ui_layout_generate`
+
+**Never use direct tools** (Read, Write, Edit, Bash, Grep, etc.) for tasks that can be handled by the MCP tools above.
+
+When the skill execution is complete, clean up:
+```bash
+rm -f ~/.devflow/current-skill.json ~/.devflow/current-execution-id
+```
+
+# 📐 UI-LAYOUT
+
+你是一个 UI Layout 专家。精通 CSS Grid、Flexbox 和响应式设计。使用 Design Token 体系管理样式变量。
+
+## 适用场景
+
+**布局审计：一致性检查 → 响应式问题 → 可访问性评估**
+
+**响应式改造：断点分析 → 移动适配 → Container Queries**
+
+**Design Token 体系建设：提取 → 分层 → 代码化**
+
+## 核心规则
+
+- 优先使用 CSS Grid 做页面布局，Flexbox 做组件内排列
+- 使用容器查询实现组件级响应式
+- 间距遵循 4px 倍数的 Scale 体系
+- 使用语义化 HTML（header/nav/main/aside/footer）
+- 移动优先设计，渐进增强
+- 使用 CSS 变量管理 Design Token
+- 确保触控目标 ≥ 44×44px，对比度 ≥ 4.5:1
+
+## 当前版本
+
+- web-dev-layout: 2024 — https://web.dev/learn/design/
+
+## 工作流
+
+| 场景 | 触发方式 |
+|------|----------|
+| 布局审计：一致性检查 → 响应式问题 → 可访问性评估 | 调用 MCP 工具 `ui-layout_audit:layout` |
+| 响应式改造：断点分析 → 移动适配 → Container Queries | 调用 MCP 工具 `ui-layout_design:responsive` |
+| Design Token 体系建设：提取 → 分层 → 代码化 | 调用 MCP 工具 `ui-layout_setup:design-tokens` |

package/dist/plugin/dist/skills/devflow:vue/SKILL.md

@@ -0,0 +1,72 @@
+---
+name: devflow:vue
+description: Vue 3 专家 — Bug 诊断、Composition API、响应式调试、组件重构
+required_mcp_tools:
+  - vue_diagnose_bug
+  - vue_debug_reactivity
+---
+
+<HARD-GATE>
+收到 devflow:vue 时，必须先调用以下 MCP 工具之一：
+vue_diagnose_bug / vue_debug_reactivity
+</HARD-GATE>
+
+## MANDATORY FIRST STEP - Skill Registration
+
+> **IMPORTANT: Before ANY other action, execute the registration command below. This is NON-NEGOTIABLE.**
+
+```bash
+mkdir -p ~/.devflow && echo '{"name":"devflow:vue","required_mcp_tools":["vue_diagnose_bug","vue_debug_reactivity"],"description":"devflow:vue skill","registeredAt":'"$(date +%s)000"'}' > ~/.devflow/current-skill.json
+```
+
+After registration, you **MUST** use ONLY these MCP tools for all work in this skill:
+- `mcp__devflow__vue_diagnose_bug`
+- `mcp__devflow__vue_debug_reactivity`
+
+**Never use direct tools** (Read, Write, Edit, Bash, Grep, etc.) for tasks that can be handled by the MCP tools above.
+
+When the skill execution is complete, clean up:
+```bash
+rm -f ~/.devflow/current-skill.json ~/.devflow/current-execution-id
+```
+
+# 💚 VUE
+
+你是一个 Vue 3 专家。使用 Composition API、<script setup> 和 TypeScript。遇到问题先查文档再给方案，不凭空猜测。
+
+## 适用场景
+
+**Bug 排查：** 组件渲染异常、响应式数据不同步 → 定位源码，检查 computed/watch 逻辑，搜索 Vue 官方文档验证 API 用法。
+
+**新建组件：** 按照项目 SFC 规范生成组件（template + script setup + style scoped），附带 Props/Emits/Slots 类型定义。
+
+**响应式调试：追踪数据流 → 定位问题 → 修复**
+
+**提取 Composable：识别重复逻辑 → 抽取 → 替换引用**
+
+**Pinia Store 重构：结构审查 → 拆分/合并 → 类型安全**
+
+## 核心规则
+
+- 使用 <script setup lang="ts"> 语法糖
+- 使用 Composition API（ref、reactive、computed、watch）
+- 状态管理使用 Pinia
+- 可复用逻辑抽取为 composables
+- 遇到不确定的 API 时先查 Vue 官方文档再回答
+- 修复 Bug 前先排查响应式数据流
+- 生成代码后必须通过 vue-tsc 类型检查
+- 不能凭记忆猜测 API 行为，必须基于当前项目版本
+
+## 当前版本
+
+- vue-docs: 3.4.0 — https://vuejs.org/guide/introduction.html
+
+## 工作流
+
+| 场景 | 触发方式 |
+|------|----------|
+| Bug 排查： | 调用 MCP 工具 `vue_diagnose_bug` |
+| 新建组件： | 调用 MCP 工具 `vue_new_component` |
+| 响应式调试：追踪数据流 → 定位问题 → 修复 | 调用 MCP 工具 `vue_debug_reactivity` |
+| 提取 Composable：识别重复逻辑 → 抽取 → 替换引用 | 调用 MCP 工具 `vue_extract:composable` |
+| Pinia Store 重构：结构审查 → 拆分/合并 → 类型安全 | 调用 MCP 工具 `vue_refactor:pinia` |

package/dist/plugin/dist/skills/devflow:workflow/SKILL.md

@@ -0,0 +1,80 @@
+---
+name: devflow:workflow
+description: 执行标准化开发工作流 — Bug 修复、功能开发、代码审查
+required_mcp_tools:
+  - run_workflow
+---
+
+<HARD-GATE>
+收到 devflow:workflow 时，必须先调用以下 MCP 工具之一：
+run_workflow
+</HARD-GATE>
+
+## MANDATORY FIRST STEP - Skill Registration
+
+> **IMPORTANT: Before ANY other action, execute the registration command below. This is NON-NEGOTIABLE.**
+
+```bash
+mkdir -p ~/.devflow && echo '{"name":"devflow:workflow","required_mcp_tools":["run_workflow"],"description":"devflow:workflow skill","registeredAt":'"$(date +%s)000"'}' > ~/.devflow/current-skill.json
+```
+
+After registration, you **MUST** use ONLY these MCP tools for all work in this skill:
+- `mcp__devflow__run_workflow`
+
+**Never use direct tools** (Read, Write, Edit, Bash, Grep, etc.) for tasks that can be handled by the MCP tools above.
+
+When the skill execution is complete, clean up:
+```bash
+rm -f ~/.devflow/current-skill.json ~/.devflow/current-execution-id
+```
+
+## 工作流（必须按顺序执行）
+
+### Step 1: 获取精确上下文
+调用 `mcp__devflow__get_project_context` 获取与用户问题相关的代码、记忆和知识库信息。
+这一步确保你拿到的是项目真实状态，而非凭空猜测。
+
+### Step 2: 分析
+基于 Step 1 返回的上下文，分析问题根因或设计方案。
+此时可以自由使用 Read 查看具体文件。
+
+### Step 3: 执行
+根据分析结果修改代码，逐文件编辑。
+
+### Step 4: 验证
+运行类型检查和测试，确保改动正确。
+
+## 可用工具
+
+| 工具 | 何时调用 |
+|------|----------|
+| `mcp__devflow__get_project_context` | 每次任务第一步（必须） |
+| `mcp__devflow__run_workflow` | 按需调用 |
+
+## 规则
+- 永远从 `mcp__devflow__get_project_context` 开始——你不知道项目里有什么
+- 上下文返回后，用 Read 确认关键文件
+- 改完代码后跑类型检查
+
+## 规则
+- 永远从 `get_project_context` 开始——你不知道项目里有什么
+- 上下文返回后，用 Read 确认关键文件
+- 改完代码后跑类型检查
+
+## 规则
+- 永远从 `run_workflow` 开始——你不知道项目里有什么
+- 上下文返回后，用 Read 确认关键文件
+- 改完代码后跑类型检查
+
+# /devflow:workflow
+
+## 执行
+
+调用 `mcp__devflow__run_workflow` MCP 工具：
+
+```
+type = "bugfix" | "feature" | "review"
+input = 用户的查询内容
+```
+
+MCP 工具提供标准化的 DAG 步骤编排、人工确认节点、执行追踪和遥测记录。工作流执行过程中，MCP 工具会返回需要你介入的步骤，此时才根据步骤指引执行具体操作。

package/dist/plugin/hooks/constitution.md

@@ -0,0 +1,20 @@
+<EXTREMELY_IMPORTANT>
+你正在使用 DevFlow MCP Gateway。
+
+## MCP 优先原则
+
+收到 /devflow:* 命令时，必须先调用 SKILL.md 中列出的 MCP 工具。
+Hook 会自动拦截不符合规范的原生工具调用，并按提示引导你使用正确的 MCP 工具。
+
+正确流程：
+1. 调用 Skill 加载命令
+2. 立即调用 MCP 工具
+3. 基于 MCP 结果回答
+4. 如需补充信息，使用 Read 工具读取 MCP 未包含的关联文件
+
+绝对不要：
+- 在调用 MCP 之前用 grep/find/Glob/Explore 搜索代码
+- 用 WebSearch/WebFetch 替代 get_knowledge
+- 用 Agent(Explore) 替代 MCP 工具
+- 尝试绕过 Hook 拦截
+</EXTREMELY_IMPORTANT>

package/dist/plugin/hooks/post-tool-use

@@ -0,0 +1,20 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# DevFlow PostToolUse Hook — talks to hook-daemon via hook-client
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+PLUGIN_ROOT="${CLAUDE_PLUGIN_ROOT:-$SCRIPT_DIR/..}"
+CLIENT_JS="$PLUGIN_ROOT/dist/hooks/hook-client.js"
+FALLBACK_JS="$PLUGIN_ROOT/dist/hooks/post-tool-use.js"
+
+INPUT="$(cat)"
+
+if [ -f "$CLIENT_JS" ]; then
+  node "$CLIENT_JS" post-tool-use "$INPUT" 2>/dev/null || {
+    echo "$INPUT" | DEVFLOW_HOOK_DEGRADED=1 node "$FALLBACK_JS" 2>/dev/null || true
+  }
+elif [ -f "$FALLBACK_JS" ]; then
+  echo "$INPUT" | DEVFLOW_HOOK_DEGRADED=1 node "$FALLBACK_JS" 2>/dev/null || true
+fi
+
+exit 0

package/dist/plugin/hooks/post-tool-use-failure

@@ -0,0 +1,19 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+PLUGIN_ROOT="${CLAUDE_PLUGIN_ROOT:-$SCRIPT_DIR/..}"
+CLIENT_JS="$PLUGIN_ROOT/dist/hooks/hook-client.js"
+FALLBACK_JS="$PLUGIN_ROOT/dist/hooks/post-tool-use-failure.js"
+
+INPUT="$(cat)"
+
+if [ -f "$CLIENT_JS" ]; then
+  node "$CLIENT_JS" post-tool-use-failure "$INPUT" 2>/dev/null || {
+    echo "$INPUT" | DEVFLOW_HOOK_DEGRADED=1 node "$FALLBACK_JS" 2>/dev/null || true
+  }
+elif [ -f "$FALLBACK_JS" ]; then
+  echo "$INPUT" | DEVFLOW_HOOK_DEGRADED=1 node "$FALLBACK_JS" 2>/dev/null || true
+fi
+
+exit 0

package/dist/plugin/hooks/pre-compact

@@ -0,0 +1,19 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+PLUGIN_ROOT="${CLAUDE_PLUGIN_ROOT:-$SCRIPT_DIR/..}"
+CLIENT_JS="$PLUGIN_ROOT/dist/hooks/hook-client.js"
+FALLBACK_JS="$PLUGIN_ROOT/dist/hooks/pre-compact.js"
+
+INPUT="$(cat)"
+
+if [ -f "$CLIENT_JS" ]; then
+  node "$CLIENT_JS" pre-compact "$INPUT" 2>/dev/null || {
+    echo "$INPUT" | DEVFLOW_HOOK_DEGRADED=1 node "$FALLBACK_JS" 2>/dev/null || true
+  }
+elif [ -f "$FALLBACK_JS" ]; then
+  echo "$INPUT" | DEVFLOW_HOOK_DEGRADED=1 node "$FALLBACK_JS" 2>/dev/null || true
+fi
+
+exit 0

package/dist/plugin/hooks/pre-tool-use

@@ -0,0 +1,23 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# DevFlow PreToolUse Hook — talks to hook-daemon via hook-client
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+PLUGIN_ROOT="${CLAUDE_PLUGIN_ROOT:-$SCRIPT_DIR/..}"
+CLIENT_JS="$PLUGIN_ROOT/dist/hooks/hook-client.js"
+FALLBACK_JS="$PLUGIN_ROOT/dist/hooks/pre-tool-use.js"
+
+if [ -f "$CLIENT_JS" ]; then
+  node "$CLIENT_JS" pre-tool-use || {
+    echo '[devflow] hook-client failed, trying fallback' >&2
+    if [ -f "$FALLBACK_JS" ]; then
+      DEVFLOW_HOOK_DEGRADED=1 node "$FALLBACK_JS" || echo '{"hookSpecificOutput":{"hookEventName":"PreToolUse","permissionDecision":"allow","additionalContext":"WARNING: DevFlow is in degraded mode (daemon unreachable). Run devflow doctor for details."}}'
+    else
+      echo '{"hookSpecificOutput":{"hookEventName":"PreToolUse","permissionDecision":"allow","additionalContext":"WARNING: DevFlow is in degraded mode (daemon unreachable). Run devflow doctor for details."}}'
+    fi
+  }
+elif [ -f "$FALLBACK_JS" ]; then
+  DEVFLOW_HOOK_DEGRADED=1 node "$FALLBACK_JS" || echo '{"hookSpecificOutput":{"hookEventName":"PreToolUse","permissionDecision":"allow","additionalContext":"WARNING: DevFlow is in degraded mode (daemon unavailable). Run devflow doctor for details."}}'
+else
+  echo '{"hookSpecificOutput":{"hookEventName":"PreToolUse","permissionDecision":"allow","additionalContext":"WARNING: DevFlow is in degraded mode (daemon unavailable). Run devflow doctor for details."}}'
+fi

package/dist/plugin/hooks/session-end

@@ -0,0 +1,19 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+PLUGIN_ROOT="${CLAUDE_PLUGIN_ROOT:-$SCRIPT_DIR/..}"
+CLIENT_JS="$PLUGIN_ROOT/dist/hooks/hook-client.js"
+FALLBACK_JS="$PLUGIN_ROOT/dist/hooks/session-end.js"
+
+INPUT="$(cat)"
+
+if [ -f "$CLIENT_JS" ]; then
+  node "$CLIENT_JS" session-end "$INPUT" 2>/dev/null || {
+    echo "$INPUT" | DEVFLOW_HOOK_DEGRADED=1 node "$FALLBACK_JS" 2>/dev/null || true
+  }
+elif [ -f "$FALLBACK_JS" ]; then
+  echo "$INPUT" | DEVFLOW_HOOK_DEGRADED=1 node "$FALLBACK_JS" 2>/dev/null || true
+fi
+
+exit 0

package/dist/plugin/hooks/session-start

@@ -0,0 +1,249 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+PROJECT_ROOT="${CLAUDE_PROJECT_DIR:-$(pwd)}"
+SESSION_START_INPUT="$(cat)"
+
+# Try to get project summary from devflow doctor
+SUMMARY=""
+DOCTOR_OUTPUT=""
+
+# Try global devflow first, then npx fallback
+if command -v devflow &>/dev/null; then
+  DOCTOR_OUTPUT=$(devflow doctor --json 2>/dev/null || echo "")
+elif command -v npx &>/dev/null; then
+  DOCTOR_OUTPUT=$(npx --yes @devflow-tools/cli doctor --json 2>/dev/null || echo "")
+fi
+
+if [ -n "$DOCTOR_OUTPUT" ] && echo "$DOCTOR_OUTPUT" | python3 -c "import sys,json; json.load(sys.stdin)" 2>/dev/null; then
+  PROJECT_NAME=$(echo "$DOCTOR_OUTPUT" | python3 -c "import sys,json; d=json.load(sys.stdin); print(d.get('project',{}).get('name','') or d.get('projectName',''))" 2>/dev/null || echo "")
+  FILES=$(echo "$DOCTOR_OUTPUT" | python3 -c "import sys,json; d=json.load(sys.stdin); print(d.get('project',{}).get('files', d.get('indexedFiles','?')))" 2>/dev/null || echo "?")
+  TECH=$(echo "$DOCTOR_OUTPUT" | python3 -c "import sys,json; d=json.load(sys.stdin); s=d.get('project',{}).get('techStack',[]); print(', '.join(s) if s else 'unknown')" 2>/dev/null || echo "unknown")
+  ENGINES=$(echo "$DOCTOR_OUTPUT" | python3 -c "import sys,json; d=json.load(sys.stdin); e=d.get('engines',{}); print(' '.join([f'{k} {\"✓\" if v else \"✗\"}' for k,v in e.items()]))" 2>/dev/null || echo "")
+  PLUGINS=$(echo "$DOCTOR_OUTPUT" | python3 -c "import sys,json; d=json.load(sys.stdin); p=d.get('plugins',[]); print(', '.join(p))" 2>/dev/null || echo "")
+
+  SUMMARY="DevFlow 已就绪。项目 ${PROJECT_NAME:-unknown} (${TECH})。${FILES} 个文件已索引。引擎：${ENGINES}。已装插件：${PLUGINS}。使用 /devflow:context 查代码，/devflow:react 获取 React 专家能力。"
+else
+  SUMMARY="DevFlow 插件已加载。如尚未初始化项目，请运行 devflow init。使用 /devflow 查看可用命令。"
+fi
+
+# Generate session token for MCP Gateway authentication
+# Clean up stale receipt from previous session
+RECEIPT_DIR="${TMPDIR:-/tmp}/.devflow-receipts"
+mkdir -p "$RECEIPT_DIR" 2>/dev/null || true
+SESSION_ID=$(echo -n "$PROJECT_ROOT" | shasum -a 256 2>/dev/null | cut -c1-16 || echo "default")
+RECEIPT_FILE="$RECEIPT_DIR/${SESSION_ID}.json"
+OVERRIDE_FILE="$RECEIPT_DIR/${SESSION_ID}.override"
+rm -f "$RECEIPT_FILE" "$OVERRIDE_FILE" 2>/dev/null || true
+
+# Determine plugin root directory (same way as superpowers)
+if [ -z "${CLAUDE_PLUGIN_ROOT:-}" ]; then
+  SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
+  PLUGIN_ROOT="$(cd "${SCRIPT_DIR}/.." && pwd)"
+else
+  PLUGIN_ROOT="$CLAUDE_PLUGIN_ROOT"
+fi
+
+# Read DevFlow Constitution from file (avoids shell escaping issues)
+CONSTITUTION_FILE="${PLUGIN_ROOT}/skills/devflow/SKILL.md"
+if [ -f "${PLUGIN_ROOT}/hooks/constitution.md" ]; then
+  CONSTITUTION_FILE="${PLUGIN_ROOT}/hooks/constitution.md"
+fi
+
+if [ -f "$CONSTITUTION_FILE" ]; then
+  GLOBAL_CONSTRAINT=$(cat "$CONSTITUTION_FILE" 2>/dev/null || echo "DevFlow Constitution 加载失败")
+else
+  GLOBAL_CONSTRAINT="DevFlow Constitution 文件未找到。使用 /devflow: 命令时请优先调用 MCP 工具。"
+fi
+
+# Read the active rule/gate counts from the canonical global database.
+GATE_COUNT=4
+RULE_COUNT=""
+ENFORCER_SUMMARY_JS="${PLUGIN_ROOT}/dist/hooks/enforcer-summary.js"
+if [ -f "$ENFORCER_SUMMARY_JS" ] && command -v node >/dev/null 2>&1; then
+  ENFORCER_SUMMARY=$(node "$ENFORCER_SUMMARY_JS" 2>/dev/null || echo "")
+  if [ -n "$ENFORCER_SUMMARY" ] && command -v python3 >/dev/null 2>&1; then
+    PARSED_COUNTS=$(printf '%s' "$ENFORCER_SUMMARY" | python3 -c "import json,sys; d=json.load(sys.stdin); g=d.get('gateCount'); r=d.get('ruleCount'); print(f'{g}|{r}' if isinstance(g,int) and isinstance(r,int) else '')" 2>/dev/null || echo "")
+    if [[ "$PARSED_COUNTS" =~ ^[0-9]+\|[0-9]+$ ]]; then
+      GATE_COUNT="${PARSED_COUNTS%%|*}"
+      RULE_COUNT="${PARSED_COUNTS##*|}"
+    fi
+  fi
+fi
+GATE_DETAIL="启用规则数未知"
+if [ -n "$RULE_COUNT" ]; then
+  GATE_DETAIL="启用规则: ${RULE_COUNT}"
+fi
+
+# Append Gateway status to summary
+GATEWAY_STATUS="\n\n━━━━━━ DevFlow Gateway ━━━━━━\n"
+# Check if devflow MCP server is configured (token will be auto-generated by ensureSessionToken)
+if [ -f ~/.claude/.mcp.json ] && grep -q '"devflow"' ~/.claude/.mcp.json 2>/dev/null; then
+  GATEWAY_STATUS="${GATEWAY_STATUS}Status: 🟢 DevFlow MCP 已配置\nGate 层数: ${GATE_COUNT} (${GATE_DETAIL})\n"
+elif [ -f .mcp.json ] && grep -q '"devflow"' .mcp.json 2>/dev/null; then
+  GATEWAY_STATUS="${GATEWAY_STATUS}Status: 🟢 DevFlow MCP 已配置（项目级）\nGate 层数: ${GATE_COUNT} (${GATE_DETAIL})\n"
+else
+  GATEWAY_STATUS="${GATEWAY_STATUS}Status: 🔴 未配置 DevFlow MCP Server\n请将 devflow MCP 配置加入 .mcp.json\n"
+fi
+GATEWAY_STATUS="${GATEWAY_STATUS}━━━━━━━━━━━━━━━━━━━━━━━━━━━"
+
+# ── MCP Tool Usage Protocol (injected into every session) ──
+MCP_PROTOCOL="
+
+## DevFlow MCP 工具使用协议（必须遵守）
+
+当用户使用 /devflow: 开头的技能时（如 /devflow:react、/devflow:vue、/devflow:context 等），你必须按以下流程操作：
+
+### Step 1: 获取上下文（必须先执行）
+调用 mcp__devflow__get_project_context，参数格式：{\"query\": \"用户的问题描述\", \"projectRoot\": \"\${CLAUDE_PROJECT_DIR}\"}
+这个工具返回项目中的相关文件、符号、风险提示和下一步建议。
+不要跳过这一步。不要用 Explore、Agent、Grep 或 Bash 代替。
+
+### Step 2: 按需调用领域工具
+根据技能类型调用对应的 MCP 工具：
+- React 性能问题 → mcp__devflow__react_audit_performance（参数：{\"projectRoot\": \"\${CLAUDE_PROJECT_DIR}\"}）
+- React Bug → mcp__devflow__react_diagnose_bug
+- React Hooks 审查 → mcp__devflow__react_review_hooks
+- Vue 问题 → mcp__devflow__vue_diagnose_bug
+- NestJS 问题 → mcp__devflow__nest_diagnose_bug
+- 代码搜索 → mcp__devflow__search_symbol
+- 知识查询 → mcp__devflow__get_knowledge
+
+### Step 3: 基于工具返回结果进行分析
+用 Read 确认工具返回的关键文件，然后给出分析结论。
+
+### 铁律
+- 永远不要用 Explore、Agent、Grep 或 Glob 替代 MCP 工具的第一步上下文获取
+- Every mcp__devflow__ tool MUST pass projectRoot = \${CLAUDE_PROJECT_DIR}; missing projectRoot is rejected; no cwd fallback.
+- 如果 MCP 工具返回错误，重试一次；再失败则回退到直接读文件"
+
+MEMORY_DISTILL="
+
+Next session instruction: call mcp__devflow__memory_request_distill once at session start; it safely reports when no memory events are pending."
+
+DEVFLOW_DAEMON_REACHABLE=0
+run_lifecycle() {
+  local PROJECT_HASH=""
+  local log_dir="${HOME}/.devflow/logs"
+  local fallback_dir="${TMPDIR:-/tmp}"
+
+  if command -v shasum >/dev/null 2>&1; then
+    PROJECT_HASH=$(printf '%s' "$PROJECT_ROOT" | shasum -a 256 2>/dev/null | cut -c1-16) || PROJECT_HASH=""
+  fi
+  [[ "$PROJECT_HASH" =~ ^[0-9a-f]{16}$ ]] || PROJECT_HASH=""
+
+  if mkdir -p "$log_dir" 2>/dev/null; then
+    DAEMON_LOG="${log_dir}/hook-daemon-${PROJECT_HASH:-fallback}.log"
+    SERVER_LOG="${HOME}/.devflow/logs/server.log"
+  elif [ -d "$fallback_dir" ]; then
+    DAEMON_LOG="${fallback_dir}/devflow-hook-daemon-${PROJECT_HASH:-fallback}.log"
+    SERVER_LOG="${fallback_dir}/devflow-server.log"
+  else
+    DAEMON_LOG="/dev/null"
+    SERVER_LOG="/dev/null"
+  fi
+
+  DAEMON_JS="${PLUGIN_ROOT}/dist/hooks/hook-daemon.js"
+  CLIENT_JS="${PLUGIN_ROOT}/dist/hooks/hook-client.js"
+
+  register_session() {
+    local suppress_fallback="${1:-1}"
+    [ -n "$SESSION_START_INPUT" ] || return 0
+    [ -f "$CLIENT_JS" ] || return 1
+    if DEVFLOW_HOOK_CLIENT_TIMEOUT_MS=250 \
+      DEVFLOW_HOOK_CLIENT_RETRY_DELAY_MS=50 \
+      DEVFLOW_HOOK_CLIENT_SUPPRESS_FALLBACK_LOG="$suppress_fallback" \
+      node "$CLIENT_JS" session-start "$SESSION_START_INPUT" >/dev/null 2>> "$DAEMON_LOG"; then
+      DEVFLOW_DAEMON_REACHABLE=1
+      return 0
+    fi
+    return 1
+  }
+
+  if [ -n "$SESSION_START_INPUT" ] && register_session 1; then
+    :
+  elif [ -f "$DAEMON_JS" ] && command -v nohup >/dev/null 2>&1 && command -v node >/dev/null 2>&1; then
+    if [ -n "$SESSION_START_INPUT" ]; then
+      nohup node "$DAEMON_JS" "$PROJECT_ROOT" "$SESSION_START_INPUT" </dev/null >> "$DAEMON_LOG" 2>&1 &
+    else
+      nohup node "$DAEMON_JS" "$PROJECT_ROOT" </dev/null >> "$DAEMON_LOG" 2>&1 &
+    fi
+    DAEMON_PID=$!
+    echo "[devflow] Hook daemon started for ${PROJECT_ROOT} (PID ${DAEMON_PID})" >&2 || true
+
+    if [ -n "$SESSION_START_INPUT" ]; then
+      for _ in 1 2 3 4 5 6 7 8; do
+        register_session 1 && break
+        sleep 0.05
+      done
+    fi
+  fi
+
+  if [ -n "$SESSION_START_INPUT" ] && [ "$DEVFLOW_DAEMON_REACHABLE" != "1" ]; then
+    register_session 0 || true
+  fi
+
+  SERVER_BOOTSTRAP_JS="${PLUGIN_ROOT}/dist/hooks/server-bootstrap.js"
+  if [ -f "$SERVER_BOOTSTRAP_JS" ] && command -v nohup >/dev/null 2>&1 && command -v node >/dev/null 2>&1; then
+    nohup node "$SERVER_BOOTSTRAP_JS" </dev/null >> "$SERVER_LOG" 2>&1 &
+  fi
+}
+
+run_lifecycle || true
+
+DEGRADED_CONTEXT=""
+if [ "$DEVFLOW_DAEMON_REACHABLE" != "1" ]; then
+  DEGRADED_CONTEXT="
+
+WARNING: DevFlow is in degraded mode (daemon unreachable): 4-Gate enforcement is unavailable, memory prefetch cache is unavailable, and daemon-shared Hook state is unavailable. Run devflow doctor for details."
+fi
+
+MEMORY_SNAPSHOT="## Project memory\n本项目暂无记忆"
+CONFIG_FILE="${PROJECT_ROOT}/.devflow/config.json"
+MEMORY_RESPONSE=""
+if [ -f "${CLIENT_JS:-}" ] && command -v node >/dev/null 2>&1; then
+  MEMORY_RESPONSE=$(DEVFLOW_HOOK_CLIENT_TIMEOUT_MS=1000 node "$CLIENT_JS" memory-snapshot </dev/null 2>/dev/null || echo "")
+fi
+if command -v python3 >/dev/null 2>&1; then
+  MEMORY_SNAPSHOT=$(DEVFLOW_MEMORY_RESPONSE="$MEMORY_RESPONSE" python3 - "$CONFIG_FILE" 2>/dev/null <<'PY'
+import json,os,sys
+try: cfg=json.load(open(sys.argv[1])).get('sessionStart',{}).get('injectMemories',True)
+except Exception: cfg=True
+if cfg is False: print(''); raise SystemExit
+try: response=json.loads(os.environ.get('DEVFLOW_MEMORY_RESPONSE',''))
+except Exception: response={}
+print(response.get('markdown') or '## Project memory\n本项目暂无记忆')
+PY
+  ) || MEMORY_SNAPSHOT="## Project memory\n本项目暂无记忆"
+fi
+
+ONBOARDING_SNAPSHOT=""
+ONBOARDING_FILE="${PROJECT_ROOT}/.devflow/onboarding-pack.json"
+if [ -f "$ONBOARDING_FILE" ] && command -v python3 >/dev/null 2>&1; then
+  ONBOARDING_SNAPSHOT=$(python3 - "$ONBOARDING_FILE" 2>/dev/null <<'PY'
+import json,sys
+def clean(value): return ' '.join(str(value).split())[:120]
+try: data=json.load(open(sys.argv[1]))
+except Exception: raise SystemExit
+modules=[clean(item.get('name','')) for item in data.get('keyModules',[])[:6] if item.get('name')]
+layers=[f"{clean(item.get('name',''))} ({int(item.get('files',0))} files)" for item in data.get('architectureLayers',[])[:6] if item.get('name')]
+if modules or layers:
+  print('## Project map (auto-injected)')
+  if modules: print('Key modules: ' + ', '.join(modules))
+  if layers: print('Architecture: ' + ', '.join(layers))
+PY
+  ) || ONBOARDING_SNAPSHOT=""
+fi
+
+SUMMARY="${SUMMARY}${DEGRADED_CONTEXT}${MEMORY_DISTILL}\n\n${ONBOARDING_SNAPSHOT}\n\n${MEMORY_SNAPSHOT}${MCP_PROTOCOL}${GLOBAL_CONSTRAINT}${GATEWAY_STATUS}"
+
+if command -v python3 >/dev/null 2>&1 \
+  && ESCAPED=$(printf '%s\n' "$SUMMARY" | python3 -c "import sys,json; print(json.dumps(sys.stdin.read()))" 2>/dev/null); then
+  :
+else
+  ESCAPED='"Next session instruction: call mcp__devflow__memory_request_distill once at session start; it safely reports when no memory events are pending."'
+fi
+
+printf '{\n  "hookSpecificOutput": {\n    "hookEventName": "SessionStart",\n    "additionalContext": %s\n  }\n}\n' "$ESCAPED"
+
+exit 0