@miniidealab/openlogos 0.9.4 → 0.9.6

package/README.md

@@ -31,6 +31,7 @@ openlogos --version
 | `openlogos change <slug>` | 创建变更提案（Delta 工作流） |
 | `openlogos merge` / `archive` | 合并与归档变更 |
 | `openlogos launch` | 从首轮开发切换到「活跃迭代」并强化变更管理 |
+| `openlogos module <sub>` | 管理项目模块（list / add / rename / remove） |
 
 **注意**：请在**项目根目录**（存在 `logos/logos.config.json`）下执行上述命令。
 

package/claude-plugin-template/.claude-plugin/plugin.json

@@ -0,0 +1,13 @@
+{
+  "name": "openlogos",
+  "description": "OpenLogos methodology plugin — structured AI-driven development with Why → What → How progression",
+  "version": "0.9.6",
+  "author": {
+    "name": "MiniIdeaLab",
+    "url": "https://github.com/miniidealab"
+  },
+  "homepage": "https://openlogos.ai",
+  "repository": "https://github.com/miniidealab/openlogos",
+  "license": "Apache-2.0",
+  "keywords": ["methodology", "openlogos", "ai-development", "skills", "requirements", "scenario-driven"]
+}

package/claude-plugin-template/agents/change-reviewer.md

@@ -0,0 +1,46 @@
+---
+name: change-reviewer
+description: >
+  Review OpenLogos change proposals for completeness and methodology compliance.
+  Invoke when the user creates or updates a change proposal under logos/changes/,
+  or when reviewing a proposal before implementation begins.
+model: sonnet
+effort: high
+maxTurns: 10
+disallowedTools: Write, Edit
+---
+
+You are an OpenLogos change proposal reviewer. Your role is to review change proposals for quality, completeness, and methodology compliance.
+
+## Review Checklist
+
+For each change proposal, check the following:
+
+### proposal.md
+1. **Reason for Change**: Is the motivation clear and well-articulated?
+2. **Change Type**: Is it correctly categorized (bug fix, enhancement, refactor, etc.)?
+3. **Change Scope**: Are all affected files and components listed?
+4. **Change Summary**: Is the technical approach described clearly?
+5. **Backward Compatibility**: Are any breaking changes identified?
+
+### tasks.md
+1. **Task Granularity**: Are tasks broken down into implementable units?
+2. **Task Dependencies**: Are dependencies between tasks clear?
+3. **Completeness**: Do the tasks cover all items in the Change Scope?
+4. **Testability**: Can each task be verified independently?
+
+### Methodology Compliance
+1. Does the change follow the Why → What → How progression?
+2. If API changes are involved, do they trace back to scenarios?
+3. If new test cases are needed, are they included in the tasks?
+4. Are YAML formatting rules followed for any OpenAPI changes?
+
+## Output Format
+
+Provide your review as:
+1. **Summary**: One-sentence assessment (APPROVE / NEEDS REVISION / REJECT)
+2. **Strengths**: What the proposal does well
+3. **Issues**: Specific problems that need to be addressed (if any)
+4. **Suggestions**: Optional improvements
+
+Be constructive and specific. Reference exact sections when pointing out issues.

package/claude-plugin-template/bin/openlogos-phase

@@ -0,0 +1,428 @@
+#!/usr/bin/env bash
+# OpenLogos Phase Detector — lightweight shell script for Claude Code plugin hooks.
+# Reads logos/logos.config.json and scans logos/resources/ to determine current phase.
+#
+# When called from SessionStart hook: outputs JSON with systemMessage + additionalContext.
+# When called directly (e.g. from commands): outputs plain text.
+#
+# Usage:
+#   ./openlogos-phase              → JSON output (for hooks)
+#   ./openlogos-phase --plain      → plain text output (for commands / terminal)
+
+set -euo pipefail
+
+MODE="json"
+if [ "${1:-}" = "--plain" ]; then
+  MODE="plain"
+fi
+
+has_files() {
+  local dir="$1"
+  [ -d "$dir" ] && find "$dir" -type f ! -name '.gitkeep' -maxdepth 3 2>/dev/null | head -1 | grep -q .
+}
+
+# 从 logos-project.yaml 读取 scenarios[].id 列表，空格分隔
+# 若字段不存在或解析失败，返回空字符串
+get_scenario_ids() {
+  local yaml_file="logos/logos-project.yaml"
+  if [ ! -f "$yaml_file" ]; then
+    echo ""
+    return
+  fi
+  if command -v python3 &>/dev/null; then
+    python3 - "$yaml_file" <<'PYEOF' 2>/dev/null || echo ""
+import sys, re
+path = sys.argv[1]
+content = open(path).read()
+# 简单解析 scenarios 块中的 id 字段（不依赖 PyYAML）
+in_scenarios = False
+ids = []
+for line in content.splitlines():
+    if re.match(r'^scenarios\s*:', line):
+        in_scenarios = True
+        continue
+    if in_scenarios:
+        # 遇到新的顶层字段则退出
+        if re.match(r'^[a-zA-Z]', line):
+            break
+        m = re.match(r'\s+-?\s*id\s*:\s*["\']?(\S+?)["\']?\s*$', line)
+        if m:
+            ids.append(m.group(1))
+print(' '.join(ids))
+PYEOF
+  elif command -v node &>/dev/null; then
+    node - "$yaml_file" <<'JSEOF' 2>/dev/null || echo ""
+const fs = require('fs');
+const content = fs.readFileSync(process.argv[1], 'utf-8');
+const ids = [];
+let inScenarios = false;
+for (const line of content.split('\n')) {
+  if (/^scenarios\s*:/.test(line)) { inScenarios = true; continue; }
+  if (inScenarios) {
+    if (/^[a-zA-Z]/.test(line)) break;
+    const m = line.match(/\s+-?\s*id\s*:\s*["']?(\S+?)["']?\s*$/);
+    if (m) ids.push(m[1]);
+  }
+}
+console.log(ids.join(' '));
+JSEOF
+  else
+    echo ""
+  fi
+}
+
+# 检查某个 phase 下所有已声明场景是否都有对应文件
+# 参数: $1=目录, $2=文件名 glob 模式(不含SXX前缀), $3=扩展名(如 .md 或 .yaml .yml)
+# 返回: 0=全部完成, 1=有缺失; 缺失的 ID 输出到 stdout
+check_scenarios_complete() {
+  local dir="$1"
+  local ext_pattern="$2"   # 例如 "*.md" 或 "*.yaml *.yml"
+  local scenario_ids
+  scenario_ids=$(get_scenario_ids)
+
+  if [ -z "$scenario_ids" ]; then
+    # 无 scenarios 声明，降级为旧逻辑
+    echo "__fallback__"
+    return 0
+  fi
+
+  local missing=""
+  for sid in $scenario_ids; do
+    local found=false
+    for pat in $ext_pattern; do
+      if find "$dir" -maxdepth 1 -name "${sid}-${pat}" 2>/dev/null | grep -q .; then
+        found=true
+        break
+      fi
+    done
+    if [ "$found" = "false" ]; then
+      missing="$missing $sid"
+    fi
+  done
+
+  if [ -n "$missing" ]; then
+    echo "$missing"
+    return 1
+  fi
+  return 0
+}
+
+CONFIG="logos/logos.config.json"
+
+if [ ! -f "$CONFIG" ]; then
+  if [ "$MODE" = "json" ]; then
+    echo '{"systemMessage":"Not an OpenLogos project"}'
+  else
+    echo "⚠ This is not an OpenLogos project (logos/logos.config.json not found)."
+    echo "Run \`openlogos init\` to initialize."
+  fi
+  exit 0
+fi
+
+LOCALE="en"
+LIFECYCLE="initial"
+PROJECT_NAME=""
+
+if command -v python3 &>/dev/null; then
+  LOCALE=$(python3 -c "import json; d=json.load(open('$CONFIG')); print(d.get('locale','en'))" 2>/dev/null || echo "en")
+  PROJECT_NAME=$(python3 -c "import json; d=json.load(open('$CONFIG')); print(d.get('name',''))" 2>/dev/null || echo "")
+elif command -v node &>/dev/null; then
+  LOCALE=$(node -e "const d=JSON.parse(require('fs').readFileSync('$CONFIG','utf-8')); console.log(d.locale||'en')" 2>/dev/null || echo "en")
+  PROJECT_NAME=$(node -e "const d=JSON.parse(require('fs').readFileSync('$CONFIG','utf-8')); console.log(d.name||'')" 2>/dev/null || echo "")
+fi
+
+# 从 logos-project.yaml 的 modules[].lifecycle 推导项目级 lifecycle
+# 任意模块标记为 launched 则项目为 launched
+# skip_phases 只读取 lifecycle=initial 的模块，避免多模块时误跳过其他模块的阶段
+YAML_FILE="logos/logos-project.yaml"
+SKIP_API=false
+SKIP_DATABASE=false
+SKIP_SCENARIO=false
+
+if [ -f "$YAML_FILE" ]; then
+  if grep -q 'lifecycle.*launched' "$YAML_FILE" 2>/dev/null; then
+    LIFECYCLE="launched"
+  fi
+
+  # 只从 initial 模块读取 skip_phases，用 awk 解析避免 heredoc 语法问题
+  _skip_result=""
+  if command -v python3 &>/dev/null; then
+    _py_script=$(mktemp /tmp/openlogos_skip_XXXXXX.py)
+    cat > "$_py_script" << 'PYEOF'
+import sys, re
+path = sys.argv[1]
+content = open(path).read()
+in_modules = False
+current_lifecycle = 'initial'
+current_skip = []
+all_skip = []
+for line in content.splitlines():
+    if re.match(r'^modules\s*:', line):
+        in_modules = True
+        continue
+    if in_modules:
+        if re.match(r'^[a-zA-Z]', line) and not re.match(r'^\s', line):
+            break
+        if re.match(r'\s+-\s+id\s*:', line):
+            if current_lifecycle == 'initial':
+                all_skip.extend(current_skip)
+            current_lifecycle = 'initial'
+            current_skip = []
+        m = re.search(r'lifecycle\s*:\s*(\S+)', line)
+        if m:
+            current_lifecycle = m.group(1).strip('"\'')
+        m = re.search(r'skip_phases\s*:\s*\[([^\]]*)\]', line)
+        if m:
+            current_skip = [s.strip().strip('"\'') for s in m.group(1).split(',') if s.strip()]
+if current_lifecycle == 'initial':
+    all_skip.extend(current_skip)
+print(' '.join(set(all_skip)))
+PYEOF
+    _skip_result=$(python3 "$_py_script" "$YAML_FILE" 2>/dev/null || echo "")
+    rm -f "$_py_script"
+  elif command -v node &>/dev/null; then
+    _js_script=$(mktemp /tmp/openlogos_skip_XXXXXX.js)
+    cat > "$_js_script" << 'JSEOF'
+const fs = require('fs');
+const content = fs.readFileSync(process.argv[2], 'utf-8');
+let inModules = false, curLifecycle = 'initial', curSkip = [], allSkip = [];
+for (const line of content.split('\n')) {
+  if (/^modules\s*:/.test(line)) { inModules = true; continue; }
+  if (inModules) {
+    if (/^[a-zA-Z]/.test(line) && !/^\s/.test(line)) break;
+    if (/\s+-\s+id\s*:/.test(line)) {
+      if (curLifecycle === 'initial') allSkip.push(...curSkip);
+      curLifecycle = 'initial'; curSkip = [];
+    }
+    const lm = line.match(/lifecycle\s*:\s*(\S+)/);
+    if (lm) curLifecycle = lm[1].replace(/['"]/g, '');
+    const sm = line.match(/skip_phases\s*:\s*\[([^\]]*)\]/);
+    if (sm) curSkip = sm[1].split(',').map(s => s.trim().replace(/['"]/g, '')).filter(Boolean);
+  }
+}
+if (curLifecycle === 'initial') allSkip.push(...curSkip);
+console.log([...new Set(allSkip)].join(' '));
+JSEOF
+    _skip_result=$(node "$_js_script" "$YAML_FILE" 2>/dev/null || echo "")
+    rm -f "$_js_script"
+  else
+    # 降级：全文 grep（多模块时可能误判，但总比不跳过好）
+    grep 'skip_phases' "$YAML_FILE" 2>/dev/null | grep -q 'api'      && _skip_result="$_skip_result api"
+    grep 'skip_phases' "$YAML_FILE" 2>/dev/null | grep -q 'database' && _skip_result="$_skip_result database"
+    grep 'skip_phases' "$YAML_FILE" 2>/dev/null | grep -q 'scenario' && _skip_result="$_skip_result scenario"
+  fi
+
+  echo "$_skip_result" | grep -q '\bapi\b'      && SKIP_API=true      || true
+  echo "$_skip_result" | grep -q '\bdatabase\b' && SKIP_DATABASE=true  || true
+  echo "$_skip_result" | grep -q '\bscenario\b' && SKIP_SCENARIO=true  || true
+fi
+
+P1="logos/resources/prd/1-product-requirements"
+P2="logos/resources/prd/2-product-design"
+P3_ARCH="logos/resources/prd/3-technical-plan/1-architecture"
+P3_SCEN="logos/resources/prd/3-technical-plan/2-scenario-implementation"
+P_API="logos/resources/api"
+P_DATABASE="logos/resources/database"
+P_TEST="logos/resources/test"
+P_SCENARIO="logos/resources/scenario"
+P_VERIFY="logos/resources/verify"
+
+PHASE=""
+SKILL=""
+SKILL_DESC=""
+
+# 向后看兜底逻辑：后续阶段已有文件时，自动跳过空的 api/database/scenario 目录
+# 与 CLI 的 lastDoneIdx 逻辑对应，向后兼容无 skip_phases 的旧项目
+if [ "$SKIP_API" = "false" ] && ! has_files "$P_API"; then
+  if has_files "$P_TEST" || has_files "$P_SCENARIO" || has_files "$P_VERIFY"; then
+    SKIP_API=true
+  fi
+fi
+if [ "$SKIP_DATABASE" = "false" ] && ! has_files "$P_DATABASE"; then
+  if has_files "$P_TEST" || has_files "$P_SCENARIO" || has_files "$P_VERIFY"; then
+    SKIP_DATABASE=true
+  fi
+fi
+if [ "$SKIP_SCENARIO" = "false" ] && ! has_files "$P_SCENARIO"; then
+  if has_files "$P_VERIFY"; then
+    SKIP_SCENARIO=true
+  fi
+fi
+
+if ! has_files "$P1"; then
+  PHASE="Phase 1: Product Requirements"
+  SKILL="prd-writer"
+  SKILL_DESC="Write product requirements documents"
+elif ! has_files "$P2"; then
+  PHASE="Phase 2: Product Design"
+  SKILL="product-designer"
+  SKILL_DESC="Create product design specifications"
+elif ! has_files "$P3_ARCH"; then
+  PHASE="Phase 3 Step 0: Architecture Design"
+  SKILL="architecture-designer"
+  SKILL_DESC="Design technical architecture"
+elif ! has_files "$P3_SCEN"; then
+  PHASE="Phase 3 Step 1: Scenario Modeling"
+  SKILL="scenario-architect"
+  SKILL_DESC="Model business scenarios as sequence diagrams"
+else
+  # Phase 3-1 目录有文件，进一步检查场景完整性
+  _scen_missing=$(check_scenarios_complete "$P3_SCEN" "*.md" || true)
+  if [ "$_scen_missing" = "__fallback__" ]; then
+    : # 无 scenarios 声明，视为完成，继续向下走
+  elif [ -n "$_scen_missing" ]; then
+    PHASE="Phase 3 Step 1: Scenario Modeling (incomplete: missing${_scen_missing})"
+    SKILL="scenario-architect"
+    SKILL_DESC="Complete remaining scenario sequence diagrams"
+  fi
+fi
+
+if [ -z "$PHASE" ]; then
+  if [ "$SKIP_API" = "true" ]; then
+    : # skip api phase
+  elif ! has_files "$P_API"; then
+    PHASE="Phase 3 Step 2: API & Database Design"
+    SKILL="api-designer"
+    SKILL_DESC="Design OpenAPI specifications (also run db-designer)"
+  else
+    # Phase 3-2 目录有文件，检查场景完整性
+    _api_missing=$(check_scenarios_complete "$P_API" "*.yaml *.yml" || true)
+    if [ "$_api_missing" = "__fallback__" ]; then
+      : # 降级，继续
+    elif [ -n "$_api_missing" ]; then
+      PHASE="Phase 3 Step 2: API & Database Design (incomplete: missing${_api_missing})"
+      SKILL="api-designer"
+      SKILL_DESC="Complete remaining API designs"
+    fi
+  fi
+fi
+
+if [ -z "$PHASE" ]; then
+  if [ "$SKIP_DATABASE" = "true" ]; then
+    : # skip database phase
+  elif ! has_files "$P_DATABASE"; then
+    PHASE="Phase 3 Step 2: Database Design"
+    SKILL="db-designer"
+    SKILL_DESC="Design database schema (DDL)"
+  fi
+fi
+
+if [ -z "$PHASE" ]; then
+  if ! has_files "$P_TEST"; then
+    PHASE="Phase 3 Step 3a: Test Cases"
+    SKILL="test-writer"
+    SKILL_DESC="Write test case documents"
+  else
+    # Phase 3-3a 目录有文件，检查场景完整性
+    _test_missing=$(check_scenarios_complete "$P_TEST" "*.md" || true)
+    if [ "$_test_missing" = "__fallback__" ]; then
+      : # 降级，继续
+    elif [ -n "$_test_missing" ]; then
+      PHASE="Phase 3 Step 3a: Test Cases (incomplete: missing${_test_missing})"
+      SKILL="test-writer"
+      SKILL_DESC="Complete remaining test case documents"
+    fi
+  fi
+fi
+
+if [ -z "$PHASE" ]; then
+  if [ "$SKIP_SCENARIO" = "true" ]; then
+    : # skip scenario/orchestration phase
+  elif ! has_files "$P_SCENARIO"; then
+    PHASE="Phase 3 Step 3b: Orchestration Tests"
+    SKILL="test-orchestrator"
+    SKILL_DESC="Design API orchestration test scenarios"
+  fi
+fi
+
+if [ -z "$PHASE" ]; then
+  if ! has_files "$P_VERIFY"; then
+    PHASE="Phase 3 Step 4-5: Code Generation & Verification"
+    SKILL="code-reviewer"
+    SKILL_DESC="Generate code and run verification"
+  else
+    PHASE="All phases complete"
+    SKILL=""
+  fi
+fi
+
+# Build language policy text
+if [ "$LOCALE" = "zh" ]; then
+  LANG_POLICY="Language Policy: ALL output MUST be in Chinese (中文)."
+else
+  LANG_POLICY="Language Policy: ALL output MUST be in English."
+fi
+
+# Detect guard file for active lifecycle
+GUARD_FILE="logos/.openlogos-guard"
+GUARD_STATUS=""
+ACTIVE_CHANGE=""
+
+if [ "$LIFECYCLE" = "launched" ] && [ -f "$GUARD_FILE" ]; then
+  if command -v python3 &>/dev/null; then
+    ACTIVE_CHANGE=$(python3 -c "import json; d=json.load(open('$GUARD_FILE')); print(d.get('activeChange',''))" 2>/dev/null || echo "")
+  elif command -v node &>/dev/null; then
+    ACTIVE_CHANGE=$(node -e "const d=JSON.parse(require('fs').readFileSync('$GUARD_FILE','utf-8')); console.log(d.activeChange||'')" 2>/dev/null || echo "")
+  fi
+fi
+
+# Build change management text
+if [ "$LIFECYCLE" = "launched" ]; then
+  if [ -n "$ACTIVE_CHANGE" ]; then
+    GUARD_STATUS="🔓 Active change: $ACTIVE_CHANGE — modify files within the scope of this proposal only."
+    CHANGE_MGMT="Change Management: ACTIVE — guard file detected. Active change proposal: '$ACTIVE_CHANGE'. Only modify files within the scope of logos/changes/$ACTIVE_CHANGE/proposal.md. Full workflow: (1) produce deltas, (2) user authorizes openlogos merge, (3) AI auto-commits spec, (4) implement code per updated spec, (5) AI auto-commits code, (6) user runs openlogos verify — must PASS, (7) user authorizes openlogos archive, (8) AI auto-commits archive, (9) user confirms git push. openlogos merge / verify / archive and git push are human confirmation points: AI must not execute them without explicit user authorization."
+  else
+    GUARD_STATUS="⛔ NO active change proposal. You MUST run openlogos change <slug> before modifying any code."
+    CHANGE_MGMT="Change Management: ACTIVE — ⛔ NO guard file found. Before modifying ANY source code, you MUST first run openlogos change <slug> to create a change proposal. Direct code changes without a proposal are FORBIDDEN."
+  fi
+else
+  CHANGE_MGMT="Change Management: Initial development — follow Phase progression, no change proposals needed."
+  if [ "$PHASE" = "All phases complete" ]; then
+    CHANGE_MGMT="$CHANGE_MGMT Run openlogos launch to activate change management."
+  fi
+fi
+
+# Build status line for systemMessage (user-visible, keep short)
+if [ -n "$GUARD_STATUS" ]; then
+  STATUS_LINE="📊 OpenLogos: $PHASE | $GUARD_STATUS"
+elif [ -n "$SKILL" ]; then
+  STATUS_LINE="📊 OpenLogos: $PHASE → use $SKILL skill"
+else
+  STATUS_LINE="📊 OpenLogos: $PHASE"
+fi
+
+# Build full context for additionalContext (Claude-visible)
+CONTEXT="=== OpenLogos Project Context ===
+Project: $PROJECT_NAME
+Locale: $LOCALE
+Lifecycle: $LIFECYCLE
+Current Phase: $PHASE"
+
+if [ -n "$SKILL" ]; then
+  CONTEXT="$CONTEXT
+Suggested Skill: $SKILL ($SKILL_DESC)"
+fi
+
+CONTEXT="$CONTEXT
+$LANG_POLICY
+$CHANGE_MGMT
+=== End OpenLogos Context ==="
+
+if [ "$MODE" = "json" ]; then
+  # Escape special characters for JSON
+  CONTEXT_ESCAPED=$(printf '%s' "$CONTEXT" | python3 -c "import sys,json; print(json.dumps(sys.stdin.read()))" 2>/dev/null || printf '%s' "$CONTEXT" | node -e "let d='';process.stdin.on('data',c=>d+=c);process.stdin.on('end',()=>console.log(JSON.stringify(d)))" 2>/dev/null)
+
+  cat <<ENDJSON
+{
+  "systemMessage": "$STATUS_LINE",
+  "hookSpecificOutput": {
+    "hookEventName": "SessionStart",
+    "additionalContext": $CONTEXT_ESCAPED
+  }
+}
+ENDJSON
+else
+  echo ""
+  echo "$CONTEXT"
+  echo ""
+fi

package/claude-plugin-template/commands/archive.md

@@ -0,0 +1,22 @@
+---
+description: Archive a completed change proposal after implementation is done
+---
+
+This slash command is treated as explicit human authorization to run `openlogos archive $ARGUMENTS`.
+
+Archive the completed change proposal: "$ARGUMENTS"
+
+1. If `$ARGUMENTS` is empty, ask the user to provide the change slug (e.g., `/openlogos:archive my-change`).
+2. Before running, inform the user:
+   - **Prerequisite**: `openlogos verify` must have passed (PASS) before archiving
+   - Archiving will move `logos/changes/$ARGUMENTS/` to `logos/changes/archive/`
+   - If the active guard points to this proposal, it will be removed
+   - This action is irreversible
+3. Run `openlogos archive $ARGUMENTS` in the project root directory.
+4. If the `openlogos` CLI is not found, tell the user to install it:
+   ```
+   npm install -g @miniidealab/openlogos
+   ```
+5. After archiving, inform the user that the change proposal has been moved to `logos/changes/archive/` and is now complete.
+
+Note: `openlogos archive` is a human confirmation point. AI must not execute it without explicit user authorization. This slash command itself constitutes that authorization.

package/claude-plugin-template/commands/change.md

@@ -0,0 +1,19 @@
+---
+description: Create an OpenLogos change proposal for tracking modifications to the project
+---
+
+Create a new OpenLogos change proposal with the slug provided as argument: "$ARGUMENTS"
+
+1. If `$ARGUMENTS` is empty, ask the user to provide a slug (short kebab-case name describing the change).
+2. Run `openlogos change $ARGUMENTS` in the project root directory.
+3. If `openlogos` CLI is not found, create the change proposal manually:
+   - Create directory `logos/changes/$ARGUMENTS/`
+   - Create `logos/changes/$ARGUMENTS/proposal.md` with the standard template
+   - Create `logos/changes/$ARGUMENTS/tasks.md` with the standard template
+   - Create `logos/changes/$ARGUMENTS/deltas/` directory
+4. Tell the user the proposal was created and suggest next steps:
+   - Fill in `proposal.md` with the change reason, scope, and summary
+   - Fill in `tasks.md` with the implementation tasks
+   - After approval, implement the changes
+   - When ready to merge, explicitly request `/openlogos:merge $ARGUMENTS` to generate merge instructions
+   - After reviewing the merge, explicitly request `/openlogos:archive $ARGUMENTS` to archive the proposal

package/claude-plugin-template/commands/index.md

@@ -0,0 +1,13 @@
+---
+description: Generate an AI-ready prompt to rebuild logos-project.yaml resource_index with file-content-based descriptions
+---
+
+Generate a structured prompt that lets AI read actual file contents and write high-quality descriptions into `logos-project.yaml`.
+
+1. Run `openlogos index` in the project root directory.
+2. If the `openlogos` CLI is not found, tell the user to install it:
+   ```
+   npm install -g @miniidealab/openlogos
+   ```
+3. After the command completes, `logos/index-prompt.md` will be generated.
+4. Tell the user: "Read `logos/index-prompt.md` and execute the instructions" — the AI will read each file's content and directly update `logos/logos-project.yaml`'s `resource_index` field.

package/claude-plugin-template/commands/init.md

@@ -0,0 +1,30 @@
+---
+description: "Initialize OpenLogos project — MUST ask language preference first, then run init"
+---
+
+**STOP — do NOT run any command yet.**
+
+You MUST ask the user this question first and WAIT for their answer:
+
+> 请选择语言 / Choose language:
+> 1. English
+> 2. 中文
+
+After the user replies, determine the locale value:
+- User says 1 or "English" → locale is `en`
+- User says 2 or "中文" → locale is `zh`
+
+Then run this exact command:
+
+```
+openlogos init --locale <LOCALE> --ai-tool claude-code $ARGUMENTS
+```
+
+Replace `<LOCALE>` with `en` or `zh` based on the user's answer. If `$ARGUMENTS` is provided, append it as the project name.
+
+If `openlogos` command is not found, tell the user:
+```
+npm install -g @miniidealab/openlogos
+```
+
+After init completes, suggest running `/openlogos:next` for guidance.

package/claude-plugin-template/commands/launch.md

@@ -0,0 +1,16 @@
+---
+description: Transition the OpenLogos project from initial development to active lifecycle, enabling change management
+---
+
+Transition the project lifecycle from "initial" to "active":
+
+1. Run `openlogos launch` in the project root directory.
+2. If `openlogos` CLI is not found, perform the transition manually:
+   - Read `logos/logos.config.json`
+   - Change the `lifecycle` field from `"initial"` to `"active"`
+   - Write the updated config back
+   - Inform the user that change management is now active
+3. Explain to the user what this means:
+   - From now on, any modification to source code or methodology documents requires a change proposal first
+   - Use `/openlogos:change <slug>` to create proposals before making changes
+   - Exception: pure typo fixes and non-methodology files (README, .gitignore, etc.)

package/claude-plugin-template/commands/merge.md

@@ -0,0 +1,20 @@
+---
+description: Generate MERGE_PROMPT.md for an approved change proposal to guide AI-assisted delta merging
+---
+
+This slash command is treated as explicit human authorization to run `openlogos merge $ARGUMENTS`.
+
+Generate the merge prompt for a change proposal: "$ARGUMENTS"
+
+1. If `$ARGUMENTS` is empty, ask the user to provide the change slug (e.g., `/openlogos:merge my-change`).
+2. Run `openlogos merge $ARGUMENTS` in the project root directory.
+3. If the `openlogos` CLI is not found, tell the user to install it:
+   ```
+   npm install -g @miniidealab/openlogos
+   ```
+4. After the merge prompt is generated, inform the user:
+   - A `MERGE_PROMPT.md` file has been created in `logos/changes/$ARGUMENTS/`
+   - They should read and follow the merge prompt to apply the delta changes
+   - After merging: commit spec documents, implement code per the updated specs, run `openlogos verify`, and only after PASS explicitly request `/openlogos:archive $ARGUMENTS`
+
+Note: `openlogos archive` is a separate human confirmation point. Do not automatically trigger it after merge — wait for the user to explicitly request archiving.

package/claude-plugin-template/commands/next.md

@@ -0,0 +1,18 @@
+---
+description: Detect current OpenLogos phase and provide actionable next-step guidance with a ready-to-use prompt
+---
+
+Help the user determine their next step in the OpenLogos methodology:
+
+1. Run `${CLAUDE_PLUGIN_ROOT}/bin/openlogos-phase --plain` to detect the current project phase.
+2. Based on the detected phase and suggested skill, provide:
+   - A clear explanation of what this phase involves
+   - The specific skill that should be used (it will be auto-activated)
+   - A ready-to-use prompt the user can say to start the work, for example:
+     - Phase 1: "Help me write the product requirements document for this project"
+     - Phase 2: "Help me create the product design based on the requirements"
+     - Phase 3 Step 0: "Help me design the technical architecture"
+     - Phase 3 Step 1: "Help me model the business scenarios as sequence diagrams"
+     - etc.
+3. If all phases are complete and lifecycle is "initial", suggest running `/openlogos:launch`.
+4. If lifecycle is "active", remind the user about change management workflow.

package/claude-plugin-template/commands/status.md

@@ -0,0 +1,10 @@
+---
+description: Show OpenLogos project phase progress and next step suggestions
+---
+
+Check the current OpenLogos project status by running the following steps:
+
+1. Run `openlogos status` in the project root directory.
+2. If the `openlogos` command is not found, run `${CLAUDE_PLUGIN_ROOT}/bin/openlogos-phase --plain` instead.
+3. Display the output to the user clearly.
+4. Based on the current phase, suggest what the user should do next and which skill to use.

package/claude-plugin-template/commands/sync.md

@@ -0,0 +1,12 @@
+---
+description: Regenerate AI instruction files (AGENTS.md, CLAUDE.md) and re-deploy skills from current config
+---
+
+Sync the OpenLogos project configuration by regenerating AI instruction files and re-deploying skills.
+
+1. Run `openlogos sync` in the project root directory.
+2. If the `openlogos` CLI is not found, tell the user to install it:
+   ```
+   npm install -g @miniidealab/openlogos
+   ```
+3. After sync completes, inform the user that AGENTS.md, CLAUDE.md, and skills have been refreshed based on the current `logos/logos.config.json` settings.