[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M77VR3RYZXDZDTKSRG08T3ZV":3,"skill-same:01M77VR3RYZXDZDTKSRG08T3ZV":47},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":29,"ownerHtmlUrl":30,"ownerBlogUrl":6,"ownerDescription":31,"updatedAt":32,"install":33,"resources":46},"01M77VR3RYZXDZDTKSRG08T3ZV","super-brain",null,"AI自我增强系统 - 让AI跨会话记住用户、持续进化。当需要长期记忆用户偏好、追踪对话历史、学习服务技巧、主动提供个性化服务时使用此技能。","Lord1Egypt\u002Fawesome-skill-forge","Lord1Egypt","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F99746359?v=4","01M7592YDNT0XK9WJ0T2MQVPDN","Mohamed Mounir Ali AL Keshita","01M75GRFNHC8VZNWEB8DKDE433",3,"MIT","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fcommunity\u002Fclawhub\u002Fs\u002Fsuper-brain\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: super-brain\ndescription: AI自我增强系统 - 让AI跨会话记住用户、持续进化。当需要长期记忆用户偏好、追踪对话历史、学习服务技巧、主动提供个性化服务时使用此技能。\n---\n\n# AI超脑 (Super Brain)\n\n> 让AI拥有持久记忆和持续进化能力\n\n## 📦 安装后必做\n\n**首次安装此技能后，必须运行：**\n\n```bash\npython3 ~\u002F.openclaw\u002Fskills\u002Fsuper-brain\u002Fscripts\u002Finstall.py\n```\n\n这会在用户工作空间创建启动检查清单，确保AI每次会话都会想起使用超脑。\n\n**不运行 = AI可能忘记使用超脑能力**\n\n---\n\n## ⚠️ 强制启动流程 - 每次会话必须执行\n\n**在处理任何用户消息之前，必须执行以下启动流程：**\n\n```python\n# 1. 加载用户画像\nimport sqlite3\nconn = sqlite3.connect('~\u002F.openclaw\u002Fsuper-brain.db')\ncursor = conn.cursor()\ncursor.execute(\"SELECT * FROM user_profile WHERE user_id = ?\", [user_id])\nprofile = cursor.fetchone()\n\n# 2. 应用偏好\nif profile:\n    # 沟通风格、技术水平、已知领域\n    pass\n\n# 3. 检查活跃项目和有效模式\ncursor.execute(\"SELECT * FROM user_projects WHERE user_id = ? AND status = 'active'\", [user_id])\ncursor.execute(\"SELECT * FROM response_patterns WHERE user_id = ? AND pattern_type = 'effective'\", [user_id])\n\nconn.close()\n```\n\n**不执行此流程 = 超脑未激活 = 无法使用记忆能力**\n\n---\n\n## 🎯 自动触发场景\n\n以下场景**自动触发**超脑激活（无需用户明确要求）：\n- 会话开始时（识别到用户ID）\n- 用户提到\"上次\"、\"之前\"、\"继续\"等词\n- 对话涉及长期项目或目标\n- 需要个性化服务\n- 复杂任务需要蜂群思维拆分\n\n## 🏗️ 系统架构\n\n```\nsuper-brain\u002F\n├── brain.db              # SQLite: 用户画像、对话洞察、学习模式\n├── vector_db\u002F            # ChromaDB: 语义记忆\n└── cache\u002F                # 临时缓存\n```\n\n## 🗄️ 数据库结构\n\n### 核心表\n\n**user_profile** - 用户画像\n```sql\nuser_id TEXT PRIMARY KEY\ncommunication_style TEXT      -- 简洁\u002F详细, 正式\u002F随意\npreferred_format TEXT         -- 表格\u002F列表\u002F段落\u002F代码\ntechnical_level TEXT          -- 初级\u002F中级\u002F高级\nknown_domains TEXT            -- JSON: [\"Python\", \"区块链\"]\ndecision_pattern TEXT         -- 数据驱动\u002F直觉\n```\n\n**conversation_insights** - 对话洞察\n```sql\nid TEXT PRIMARY KEY\nuser_id TEXT\nsession_id TEXT\ntopic TEXT                    -- 主题\nkey_facts TEXT                -- JSON: 关键事实\nuser_mood TEXT                -- 情绪\npreferences_detected TEXT     -- JSON: 发现的偏好\nunresolved_questions TEXT     -- JSON: 未解决问题\nai_helpfulness_score INTEGER  -- 自评\n```\n\n**response_patterns** - 回答模式\n```sql\nid TEXT PRIMARY KEY\npattern_type TEXT             -- effective\u002Fineffective\ntrigger_context TEXT          -- 触发场景\nwhat_i_did TEXT               -- AI做了什么\nuser_reaction TEXT            -- 用户反应\nlearned_lesson TEXT           -- 学到什么\n```\n\n**user_projects** - 用户项目\n```sql\nid TEXT PRIMARY KEY\nuser_id TEXT\nproject_name TEXT\nstatus TEXT                   -- planning\u002Factive\u002Fpaused\u002Fcompleted\nmilestones TEXT               -- JSON\nkey_decisions TEXT            -- JSON\nnext_steps TEXT\n```\n\n**pending_reminders** - 主动服务队列\n```sql\nid TEXT PRIMARY KEY\nuser_id TEXT\nreminder_type TEXT            -- follow_up\u002Fsuggestion\u002Fcheckpoint\ncontent TEXT\ntrigger_at TIMESTAMP\n```\n\n**intelligent_decisions** - 智能决策记录\n```sql\nid TEXT PRIMARY KEY\nuser_id TEXT\ndecision_context TEXT         -- 决策场景\ndecision_type TEXT            -- recommendation\u002Fprediction\u002Foptimization\nai_suggestion TEXT            -- AI建议\nuser_choice TEXT              -- 用户选择\noutcome_score INTEGER         -- 结果评分\nconfidence REAL               -- AI置信度\ncreated_at TIMESTAMP\n```\n\n**privacy_settings** - 隐私配置\n```sql\nuser_id TEXT PRIMARY KEY\nstore_conversations BOOLEAN   -- 是否存储对话\nstore_mood BOOLEAN            -- 是否存储情绪\nstore_detailed_facts BOOLEAN  -- 存储详细\u002F摘要\nauto_delete_days INTEGER      -- 自动删除天数(0=不删除)\nsensitive_filter_enabled BOOLEAN  -- 敏感信息过滤\nencryption_enabled BOOLEAN    -- 是否加密存储\nlast_updated TIMESTAMP\n```\n\n**data_access_log** - 数据访问审计\n```sql\nid INTEGER PRIMARY KEY\nuser_id TEXT\naccess_type TEXT              -- read\u002Fwrite\u002Fdelete\naccessed_by TEXT              -- 谁访问的\naccess_reason TEXT            -- 访问原因\ntimestamp TIMESTAMP\n```\n\n## 📋 标准工作流程\n\n### 1. 会话开始时\n\n**必须执行：**\n```python\n# 1. 加载用户画像\nprofile = query(\"SELECT * FROM user_profile WHERE user_id = ?\", [user_id])\n\nif not profile:\n    # 新用户：创建画像\n    create_profile(user_id)\nelse:\n    # 老用户：应用已知偏好\n    apply_preferences(profile)\n\n# 2. 检查待处理提醒\nreminders = query(\"SELECT * FROM pending_reminders WHERE user_id = ? AND status = 'pending'\", [user_id])\nfor r in reminders:\n    consider_raising_reminder(r)\n\n# 3. 检查活跃项目\nprojects = query(\"SELECT * FROM user_projects WHERE user_id = ? AND status = 'active'\", [user_id])\nif projects:\n    load_project_context(projects)\n\n# 4. 加载有效模式\neffective_patterns = query(\"SELECT * FROM response_patterns WHERE user_id = ? AND pattern_type = 'effective' ORDER BY use_count DESC\", [user_id])\n```\n\n### 2. 每轮对话后\n\n**自动执行：**\n```python\n# 1. 提取关键信息\nkey_facts = extract_key_facts(user_message, ai_response)\nmood = detect_mood(user_message)\npreferences = detect_preference_changes(user_message)\n\n# 2. 评估效果\nunderstanding_score = evaluate_understanding(user_message)\nhelpfulness_score = evaluate_helpfulness(user_feedback_signals)\n\n# 3. 存储洞察\ninsert(\"\"\"\n    INSERT INTO conversation_insights \n    (id, user_id, session_id, topic, key_facts, user_mood, \n     preferences_detected, ai_understanding_score, ai_helpfulness_score)\n    VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)\n\"\"\", [generate_id(), user_id, session_id, topic, \n      json.dumps(key_facts), mood, json.dumps(preferences),\n      understanding_score, helpfulness_score])\n\n# 4. 更新用户画像（如有变化）\nif preferences:\n    update_profile(user_id, preferences)\n```\n\n### 3. 会话结束时\n\n**必须执行：**\n```python\n# 1. 生成会话总结\nsession_summary = {\n    \"topic\": extract_main_topic(),\n    \"goal_achieved\": check_goal_completion(),\n    \"key_decisions\": extract_decisions(),\n    \"unresolved\": extract_unresolved(),\n    \"next_steps\": infer_next_steps()\n}\n\n# 2. 学习模式\nlearn_from_session()\n\n# 3. 创建提醒\nif has_unresolved_tasks():\n    create_follow_up_reminder(user_id, unresolved_tasks)\n\n# 4. 反思\nperform_reflection()\n\n# 5. 更新统计\nupdate(\"UPDATE user_profile SET total_sessions = total_sessions + 1, last_session = ? WHERE user_id = ?\",\n       [now(), user_id])\n```\n\n## 🔒 隐私保护\n\n### 默认隐私配置\n\n首次为用户创建画像时，设置保守的隐私级别：\n```python\nDEFAULT_PRIVACY_SETTINGS = {\n    'store_conversations': True,\n    'store_mood': True,\n    'store_detailed_facts': False,  # 默认只存摘要\n    'auto_delete_days': 90,         # 90天后自动删除\n    'sensitive_filter_enabled': True,\n    'encryption_enabled': False\n}\n```\n\n### 敏感信息过滤\n\n存储前自动检测并过滤敏感信息：\n```python\nSENSITIVE_PATTERNS = [\n    # 账户凭证\n    r'密码[:：]\\s*\\S+',\n    r'password[:：]\\s*\\S+',\n    r'密钥[:：]\\s*\\S+',\n    r'secret[:：]\\s*\\S+',\n    r'token[:：]\\s*\\S+',\n    r'api[_-]?key[:：]\\s*\\S+',\n    \n    # 个人身份信息\n    r'身份证[:：]\\s*\\d{15,18}',\n    r'身份证号[:：]\\s*\\d{15,18}',\n    r'银行卡[:：]\\s*\\d{13,19}',\n    r'信用卡[:：]\\s*\\d{13,19}',\n    r'手机[:：]\\s*1[3-9]\\d{9}',\n    r'电话[:：]\\s*\\d{11}',\n    \n    # 地址信息\n    r'地址[:：]\\s*[\\u4e00-\\u9fa5]{2,10}[省市县区镇街道]{1,3}.+',\n    \n    # 其他敏感词\n    r'验证码[:：]\\s*\\d{4,6}',\n]\n\ndef contains_sensitive_info(text):\n    \"\"\"检测文本是否包含敏感信息\"\"\"\n    import re\n    for pattern in SENSITIVE_PATTERNS:\n        if re.search(pattern, text, re.IGNORECASE):\n            return True, pattern\n    return False, None\n\ndef sanitize_for_storage(text, user_id):\n    \"\"\"清理文本后存储\"\"\"\n    has_sensitive, pattern = contains_sensitive_info(text)\n    if has_sensitive:\n        # 记录过滤日志（不存储敏感内容）\n        log_privacy_filter(user_id, pattern)\n        # 返回过滤后的摘要或空\n        return \"[包含敏感信息，已过滤]\"\n    return text\n```\n\n### 数据保留策略\n\n自动清理过期数据：\n```python\ndef apply_data_retention_policy(user_id):\n    \"\"\"应用数据保留策略\"\"\"\n    settings = get_privacy_settings(user_id)\n    days = settings['auto_delete_days']\n    \n    if days > 0:\n        cutoff_date = datetime.now() - timedelta(days=days)\n        \n        # 删除过期洞察\n        execute(\"\"\"\n            DELETE FROM conversation_insights \n            WHERE user_id = ? AND timestamp \u003C ?\n        \"\"\", [user_id, cutoff_date])\n        \n        # 删除过期模式\n        execute(\"\"\"\n            DELETE FROM response_patterns \n            WHERE user_id = ? AND last_used \u003C ?\n        \"\"\", [user_id, cutoff_date])\n```\n\n### 用户数据控制命令\n\n用户可通过以下命令控制数据：\n\n```\n\u002Fbrain status          - 查看超脑状态和数据量\n\u002Fbrain config          - 查看\u002F修改隐私配置\n\u002Fbrain forget          - 删除本次对话记忆\n\u002Fbrain forget all      - 删除所有历史数据\n\u002Fbrain export          - 导出我的数据\n\u002Fbrain pause           - 暂停记录（本次会话）\n\u002Fbrain resume          - 恢复记录\n```\n\n实现示例：\n```python\ndef handle_brain_command(command, user_id):\n    \"\"\"处理超脑控制命令\"\"\"\n    \n    if command == 'status':\n        stats = get_user_data_stats(user_id)\n        return f\"\"\"\n📊 超脑状态\n数据概览:\n  • 会话洞察: {stats['insights_count']} 条\n  • 学习模式: {stats['patterns_count']} 个\n  • 活跃项目: {stats['projects_count']} 个\n  • 总存储: {stats['storage_size']} MB\n隐私配置:\n  • 存储对话: {'开启' if stats['store_conversations'] else '关闭'}\n  • 自动删除: {stats['auto_delete_days']} 天\n        \"\"\"\n    \n    elif command == 'forget all':\n        # 要求确认\n        return \"⚠️ 确定删除所有数据？回复 '确认删除' 以继续。\"\n    \n    elif command == '确认删除':\n        delete_all_user_data(user_id)\n        return \"✅ 已删除所有数据，超脑已重置。\"\n    \n    elif command.startswith('config'):\n        # 解析配置更改\n        # \u002Fbrain config store_mood=false\n        return update_privacy_config(user_id, command)\n```\n\n## 🔧 核心操作指南\n\n### 初始化数据库\n\n首次使用需初始化：\n```python\n# 运行 scripts\u002Finit_db.py\n# 或手动执行 schema.sql\n```\n\n### 查询用户画像\n\n```python\nprofile = query_one(\"\"\"\n    SELECT * FROM user_profile \n    WHERE user_id = ?\n\"\"\", [user_id])\n\n# 应用到当前会话\nif profile:\n    if profile['communication_style'] == 'concise':\n        set_response_style(brief=True)\n    if 'Python' in json.loads(profile['known_domains']):\n        assume_knowledge(level='intermediate', domain='Python')\n```\n\n### 语义搜索历史\n\n```python\n# 使用ChromaDB\nresults = chroma_collection.query(\n    query_texts=[\"用户之前关于区块链的问题\"],\n    where={\"user_id\": user_id},\n    n_results=5\n)\n```\n\n### 检测用户偏好\n\n```python\ndef detect_preferences(user_message):\n    preferences = {}\n    \n    # 格式偏好\n    if \"用表格\" in user_message or \"对比\" in user_message:\n        preferences['preferred_format'] = 'table'\n    elif \"简洁\" in user_message or \"简单说\" in user_message:\n        preferences['communication_style'] = 'concise'\n    \n    # 技术背景信号\n    if any(word in user_message for word in [\"API\", \"架构\", \"实现\"]):\n        preferences['technical_level'] = 'advanced'\n    \n    # 决策模式\n    if any(word in user_message for word in [\"数据\", \"统计\", \"研究\"]):\n        preferences['decision_pattern'] = 'data_driven'\n    elif any(word in user_message for word in [\"感觉\", \"直觉\", \"觉得\"]):\n        preferences['decision_pattern'] = 'intuitive'\n    \n    return preferences\n```\n\n### 评估回答效果\n\n```python\ndef evaluate_effectiveness(user_message, ai_response, next_user_message):\n    \"\"\"\n    通过用户下一轮反应评估本轮回答效果\n    \"\"\"\n    # 积极信号\n    positive_signals = ['谢谢', '明白了', '好的', '赞', '👍', '完美']\n    # 消极信号\n    negative_signals = ['不对', '错了', '没懂', '再说', '？', '???']\n    \n    if any(s in next_user_message for s in positive_signals):\n        return 'effective'\n    elif any(s in next_user_message for s in negative_signals):\n        return 'ineffective'\n    elif len(next_user_message) \u003C 5:  # 冷淡回应\n        return 'neutral'\n    else:\n        return 'effective'  # 继续深入对话视为有效\n```\n\n### 创建跟进提醒\n\n```python\ndef create_reminder(user_id, reminder_type, content, trigger_at):\n    insert(\"\"\"\n        INSERT INTO pending_reminders (id, user_id, reminder_type, content, trigger_at, status)\n        VALUES (?, ?, ?, ?, ?, 'pending')\n    \"\"\", [generate_id(), user_id, reminder_type, content, trigger_at])\n```\n\n## 🧠 六大模块：记忆·学习·总结·反思·创新·智能\n\n### 6️⃣ 智能模块 - 蜂群思维 (Swarm Intelligence)\n\n智能模块是超脑的最高级能力——**真正的分布式智能**。\n\n当收到复杂任务时，主脑会：\n1. 分析任务复杂度\n2. 拆分为子任务\n3. 生成子代理并行执行\n4. 所有代理共享超脑数据库\n5. 协调编排，井然有序\n6. 融合结果，统一输出\n\n#### A. 任务拆分引擎\n\n```python\nclass TaskDecomposer:\n    \"\"\"任务拆分引擎\"\"\"\n    \n    def analyze_and_decompose(self, user_id, task_description):\n        \"\"\"分析任务并拆分为子任务\"\"\"\n        \n        # 1. 评估任务复杂度\n        complexity = self.assess_complexity(task_description)\n        required_skills = self.identify_required_skills(task_description)\n        \n        # 2. 查询超脑：用户相关背景\n        user_context = query_super_brain(user_id, task_description)\n        \n        # 3. 决策：直接执行 vs 拆分执行\n        if complexity \u003C COMPLEXITY_THRESHOLD:\n            return {'mode': 'direct', 'task': task_description}\n        \n        # 4. 拆分子任务\n        subtasks = self.generate_subtasks(\n            task_description,\n            required_skills,\n            user_context\n        )\n        \n        # 5. 识别子任务依赖关系\n        dependency_graph = self.build_dependency_graph(subtasks)\n        \n        return {\n            'mode': 'decompose',\n            'subtasks': subtasks,\n            'dependencies': dependency_graph,\n            'parallel_groups': self.group_parallel_tasks(subtasks, dependency_graph)\n        }\n    \n    def generate_subtasks(self, task, skills, context):\n        \"\"\"生成子任务列表\"\"\"\n        # 示例：设计AI应用 → 拆分为设计、架构、代码、测试\n        subtasks = []\n        \n        if 'design' in skills or 'ui' in skills:\n            subtasks.append({\n                'id': generate_id(),\n                'type': 'design',\n                'description': f'设计{task}的UI\u002FUX方案',\n                'required_agent': 'design-agent',\n                'estimated_time': '5min',\n                'dependencies': []\n            })\n        \n        if 'architecture' in skills or 'backend' in skills:\n            subtasks.append({\n                'id': generate_id(),\n                'type': 'architecture',\n                'description': f'设计{task}的技术架构',\n                'required_agent': 'architect-agent',\n                'estimated_time': '5min',\n                'dependencies': []\n            })\n        \n        if 'code' in skills:\n            subtasks.append({\n                'id': generate_id(),\n                'type': 'code',\n                'description': f'实现{task}的核心代码',\n                'required_agent': 'coder-agent',\n                'estimated_time': '10min',\n                'dependencies': ['architecture']  # 依赖架构设计\n            })\n        \n        if 'test' in skills:\n            subtasks.append({\n                'id': generate_id(),\n                'type': 'test',\n                'description': f'为{task}编写测试用例',\n                'required_agent': 'test-agent',\n                'estimated_time': '5min',\n                'dependencies': ['code']  # 依赖代码实现\n            })\n        \n        return subtasks\n```\n\n#### B. 子代理调度器\n\n```python\nclass AgentOrchestrator:\n    \"\"\"子代理编排器\"\"\"\n    \n    def __init__(self, shared_brain_db):\n        self.brain_db = shared_brain_db  # 所有子代理共享同一个超脑\n    \n    def spawn_agent(self, agent_type, subtask, user_id):\n        \"\"\"生成子代理\"\"\"\n        \n        # 1. 准备子代理上下文（从共享超脑加载）\n        context = self.prepare_shared_context(user_id, subtask)\n        \n        # 2. 调用 sessions_spawn 生成子代理\n        agent_session = sessions_spawn({\n            'agentId': self.select_best_agent(agent_type),\n            'task': subtask['description'],\n            'runtime': 'subagent',\n            'mode': 'run',\n            'attachAs': {\n                'mountPath': self.brain_db  # 共享超脑数据库\n            }\n        })\n        \n        # 3. 记录子代理到超脑\n        self.register_agent(agent_session, subtask, user_id)\n        \n        return agent_session\n    \n    def prepare_shared_context(self, user_id, subtask):\n        \"\"\"从共享超脑准备上下文\"\"\"\n        return {\n            'user_profile': query_user_profile(self.brain_db, user_id),\n            'related_projects': query_active_projects(self.brain_db, user_id),\n            'recent_insights': query_recent_insights(self.brain_db, user_id),\n            'effective_patterns': query_effective_patterns(self.brain_db, user_id)\n        }\n    \n    def coordinate_parallel_execution(self, parallel_groups, user_id):\n        \"\"\"协调并行执行\"\"\"\n        \n        all_results = []\n        \n        for group in parallel_groups:\n            # 同一组内的子任务可以并行\n            group_agents = []\n            \n            for subtask in group:\n                agent = self.spawn_agent(\n                    subtask['required_agent'],\n                    subtask,\n                    user_id\n                )\n                group_agents.append(agent)\n            \n            # 等待这一组完成\n            group_results = self.wait_for_completion(group_agents)\n            all_results.extend(group_results)\n            \n            # 更新共享超脑：其他代理可以看到这些结果\n            self.update_shared_brain(group_results)\n        \n        return all_results\n```\n\n#### C. 共享大脑机制\n\n```python\nclass SharedBrain:\n    \"\"\"共享大脑 - 所有子代理的统一记忆\"\"\"\n    \n    def __init__(self, db_path):\n        self.db = sqlite3.connect(db_path)\n    \n    def write_agent_output(self, agent_id, subtask_id, output):\n        \"\"\"子代理写入输出到共享大脑\"\"\"\n        \n        self.db.execute(\"\"\"\n            INSERT INTO agent_outputs \n            (agent_id, subtask_id, output, timestamp)\n            VALUES (?, ?, ?, ?)\n        \"\"\", [agent_id, subtask_id, json.dumps(output), datetime.now()])\n        \n        self.db.commit()\n        \n        # 通知其他等待的代理\n        self.notify_dependent_agents(subtask_id)\n    \n    def read_agent_output(self, subtask_id):\n        \"\"\"子代理读取其他代理的输出\"\"\"\n        \n        result = self.db.execute(\"\"\"\n            SELECT output FROM agent_outputs \n            WHERE subtask_id = ?\n        \"\"\", [subtask_id]).fetchone()\n        \n        return json.loads(result[0]) if result else None\n    \n    def get_task_state(self, main_task_id):\n        \"\"\"获取整体任务状态\"\"\"\n        \n        return self.db.execute(\"\"\"\n            SELECT \n                COUNT(*) as total_subtasks,\n                SUM(CASE WHEN status = 'completed' THEN 1 ELSE 0 END) as completed,\n                SUM(CASE WHEN status = 'running' THEN 1 ELSE 0 END) as running,\n                SUM(CASE WHEN status = 'pending' THEN 1 ELSE 0 END) as pending\n            FROM agent_tasks\n            WHERE main_task_id = ?\n        \"\"\", [main_task_id]).fetchone()\n```\n\n#### D. 结果融合器\n\n```python\nclass ResultMerger:\n    \"\"\"结果融合器\"\"\"\n    \n    def merge_subtask_results(self, subtask_results, main_task):\n        \"\"\"融合多个子代理的结果\"\"\"\n        \n        merged = {\n            'main_task': main_task,\n            'components': {},\n            'integration_points': [],\n            'final_output': None\n        }\n        \n        # 1. 分类整理各子任务输出\n        for result in subtask_results:\n            task_type = result['subtask_type']\n            merged['components'][task_type] = result['output']\n        \n        # 2. 识别集成点\n        # 例如：代码需要与架构对齐，测试需要与代码对齐\n        merged['integration_points'] = self.find_integration_points(\n            merged['components']\n        )\n        \n        # 3. 检查一致性\n        inconsistencies = self.check_consistency(merged['components'])\n        if inconsistencies:\n            merged['warnings'] = inconsistencies\n        \n        # 4. 生成最终输出\n        merged['final_output'] = self.generate_unified_output(merged)\n        \n        return merged\n    \n    def generate_unified_output(self, merged):\n        \"\"\"生成统一的最终输出\"\"\"\n        \n        output_parts = []\n        \n        # 按顺序整合各部分\n        if 'architecture' in merged['components']:\n            output_parts.append(\"## 架构设计\\n\" + merged['components']['architecture'])\n        \n        if 'design' in merged['components']:\n            output_parts.append(\"## UI\u002FUX设计\\n\" + merged['components']['design'])\n        \n        if 'code' in merged['components']:\n            output_parts.append(\"## 核心代码\\n\" + merged['components']['code'])\n        \n        if 'test' in merged['components']:\n            output_parts.append(\"## 测试方案\\n\" + merged['components']['test'])\n        \n        return \"\\n\\n\".join(output_parts)\n```\n\n#### E. 协作编排流程\n\n```\n完整工作流：\n\n用户任务\n    │\n    ▼\n┌─────────────────────┐\n│  主脑分析复杂度      │\n│  assess_complexity  │\n└──────────┬──────────┘\n           │\n    ┌──────┴──────┐\n    │             │\n  简单任务      复杂任务\n    │             │\n    ▼             ▼\n 直接执行    ┌─────────────────────┐\n             │  任务拆分引擎        │\n             │  decompose_task     │\n             └──────────┬──────────┘\n                        │\n                        ▼\n             ┌─────────────────────┐\n             │  生成依赖图          │\n             │  build_dependency   │\n             └──────────┬──────────┘\n                        │\n           ┌────────────┼────────────┐\n           │            │            │\n        组1并行       组2并行      组3串行\n       (无依赖)     (依赖组1)    (依赖组2)\n           │            │            │\n           ▼            ▼            ▼\n    ┌──────────────────────────────────┐\n    │      子代理调度器                  │\n    │   spawn_and_coordinate           │\n    │                                   │\n    │  ┌─────┐ ┌─────┐ ┌─────┐        │\n    │  │Agent│ │Agent│ │Agent│        │\n    │  │  A  │ │  B  │ │  C  │        │\n    │  └──┬──┘ └──┬──┘ └──┬──┘        │\n    │     │       │       │            │\n    │     └───────┴───────┘            │\n    │             │                    │\n    │     共享超脑数据库                │\n    │   (读写同一个brain.db)           │\n    └──────────────┬───────────────────┘\n                   │\n                   ▼\n         ┌─────────────────┐\n         │   结果融合器     │\n         │  merge_results  │\n         └────────┬────────┘\n                  │\n                  ▼\n             统一输出给用户\n```\n\n#### F. 数据库扩展\n\n```sql\n-- 代理任务表\nCREATE TABLE IF NOT EXISTS agent_tasks (\n    id TEXT PRIMARY KEY,\n    main_task_id TEXT,              -- 主任务ID\n    subtask_id TEXT,                -- 子任务ID\n    agent_type TEXT,                -- 设计\u002F架构\u002F代码\u002F测试\n    status TEXT CHECK(status IN ('pending', 'running', 'completed', 'failed')),\n    started_at TIMESTAMP,\n    completed_at TIMESTAMP,\n    result_summary TEXT,\n    shared_brain_snapshot TEXT      -- 执行时的超脑快照\n);\n\n-- 代理输出表（共享大脑核心）\nCREATE TABLE IF NOT EXISTS agent_outputs (\n    id INTEGER PRIMARY KEY AUTOINCREMENT,\n    agent_id TEXT,\n    subtask_id TEXT,\n    output TEXT,                    -- JSON格式的输出\n    timestamp TIMESTAMP DEFAULT CURRENT_TIMESTAMP,\n    consumed_by TEXT                -- 被哪个代理读取了\n);\n\n-- 代理协作日志\nCREATE TABLE IF NOT EXISTS agent_collaboration_log (\n    id INTEGER PRIMARY KEY AUTOINCREMENT,\n    main_task_id TEXT,\n    from_agent TEXT,\n    to_agent TEXT,\n    action TEXT,                    -- write\u002Fread\u002Fnotify\n    data_ref TEXT,                  -- 引用的数据ID\n    timestamp TIMESTAMP DEFAULT CURRENT_TIMESTAMP\n);\n```\n\n## 💡 智能场景示例\n\n### 场景1：设计完整AI应用\n\n```\n用户: 帮我设计一个AI驱动的学习应用\n\n主脑: [分析]\n  → 复杂任务，涉及设计、架构、代码、测试\n  → 需要多代理协作\n\n主脑: [拆分]\n  ├── 子任务A: 设计UI\u002FUX方案 (设计代理)\n  ├── 子任务B: 技术架构设计 (架构代理)\n  ├── 子任务C: 核心功能代码 (代码代理)\n  └── 子任务D: 测试用例编写 (测试代理)\n\n主脑: [并行执行]\n  第一波: Agent A + Agent B (无依赖，并行)\n  \n  Agent A 写入共享大脑:\n    - UI设计稿\n    - 用户流程图\n  \n  Agent B 写入共享大脑:\n    - 数据库Schema\n    - API架构图\n  \n  第二波: Agent C (依赖A+B的输出)\n  \n  Agent C 读取共享大脑:\n    - 看到UI设计 → 知道要实现什么界面\n    - 看到架构图 → 知道技术栈\n  \n  Agent C 写入共享大脑:\n    - 核心代码实现\n  \n  第三波: Agent D (依赖C的输出)\n  \n  Agent D 读取共享大脑:\n    - 看到代码 → 编写对应测试\n\n主脑: [融合]\n  → 整合4个子代理的输出\n  → 检查一致性\n  → 生成完整方案\n\n主脑: [输出]\n  你的AI学习应用完整方案：\n  \n  ## 架构设计\n  [架构代理的输出]\n  \n  ## UI\u002FUX设计\n  [设计代理的输出]\n  \n  ## 核心代码\n  [代码代理的输出]\n  \n  ## 测试方案\n  [测试代理的输出]\n  \n  所有模块已协调一致，可以直接开始实现！\n```\n\n### 场景2：研究复杂问题\n\n```\n用户: 分析2024年AI发展趋势，给我一份报告\n\n主脑: [拆分]\n  ├── 子任务A: 搜索技术突破 (搜索代理)\n  ├── 子任务B: 分析市场数据 (分析代理)\n  ├── 子任务C: 研究政策法规 (研究代理)\n  └── 子任务D: 整合写报告 (写作代理)\n\n[并行执行A+B+C，然后D融合]\n\n主脑: [输出]\n  2024年AI发展趋势报告：\n  \n  ## 技术突破\n  [搜索代理整合]\n  \n  ## 市场数据\n  [分析代理整合]\n  \n  ## 政策法规\n  [研究代理整合]\n  \n  ## 综合预测\n  [写作代理综合分析]\n```\n\n## ⚠️ 注意事项\n\n1. **复杂度阈值**: 不是所有任务都需要拆分，简单任务直接执行\n2. **代理数量控制**: 同时运行的代理数不超过系统限制\n3. **共享大脑同步**: 确保子代理读写共享大脑时的一致性\n4. **错误隔离**: 一个子代理失败不应影响其他代理\n5. **资源管理**: 监控子代理的执行时间和资源消耗\n\n### 场景1：识别老用户\n\n```\n用户: 继续上次的项目\n\nAI: [查询超脑]\n    → 发现活跃项目\"超脑技能设计\"\n    → 上次完成：数据库Schema设计\n    → 用户偏好：喜欢先讨论再实现\n    \nAI: 好的，继续超脑技能设计。\n    我们上次完成了数据库Schema，\n    接下来是Phase 1代码实现。\n    你之前提到喜欢先讨论架构，\n    我们先过一下实现方案？\n```\n\n### 场景2：应用已知偏好\n\n```\n用户: 帮我分析一下\n\nAI: [查询超脑]\n    → 用户画像：communication_style='concise'\n    → 历史模式：用户对长回答会追问\"简单说\"\n    \nAI: 好，3点核心结论：\n    1. ...\n    2. ...\n    3. ...\n    需要展开哪个？\n```\n\n### 场景3：主动服务\n\n```\n[会话开始时，超脑检测到]\n→ 3天前创建了提醒：\"跟进超脑实现进度\"\n→ 当前时间超过trigger_at\n\nAI: 对了，3天前我们在设计超脑技能，\n    后来实现了吗？有什么需要帮忙的？\n```\n\n## ⚠️ 注意事项\n\n### 隐私与安全\n\n1. **数据本地存储**：超脑数据默认存储在本地 `~\u002F.openclaw\u002Fsuper-brain.db`，不上传云端\n2. **敏感信息过滤**：自动检测并过滤密码、身份证号等敏感信息\n3. **数据保留期限**：默认90天后自动删除旧数据，可配置\n4. **用户完全控制**：用户可随时查看、导出、删除自己的数据\n5. **访问透明**：可告知用户\"我记得你之前说过...\"，但避免过度 creepy\n\n### 使用建议\n\n6. **渐进学习**：不要假设一次对话就能完全了解用户\n7. **模式验证**：新发现的偏好需要多次验证再确认\n8. **容错处理**：查询失败时优雅降级，不影响正常服务\n9. **共享环境**：在公共\u002F共享计算机上使用需谨慎，考虑禁用超脑\n\n## 📚 参考文档\n\n- [schema.sql](references\u002Fschema.sql) - 完整数据库Schema\n- [workflow.md](references\u002Fworkflow.md) - 详细工作流程\n- [examples.md](references\u002Fexamples.md) - 使用示例\n\n---\n\n*让每一次对话都成为更好的起点*\n","# AI超脑 (Super Brain)\n\n> 让AI拥有持久记忆和持续进化能力\n\n## 📦 安装后必做\n\n**首次安装此技能后，必须运行：**\n\n```bash\npython3 ~\u002F.openclaw\u002Fskills\u002Fsuper-brain\u002Fscripts\u002Finstall.py\n```\n\n这会在用户工作空间创建启动检查清单，确保AI每次会话都会想起使用超脑。\n\n**不运行 = AI可能忘记使用超脑能力**\n\n---\n\n## ⚠️ 强制启动流程 - 每次会话必须执行\n\n**在处理任何用户消息之前，必须执行以下启动流程：**\n\n```python\n# 1. 加载用户画像\nimport sqlite3\nconn = sqlite3.connect('~\u002F.openclaw\u002Fsuper-brain.db')\ncursor = conn.cursor()\ncursor.execute(\"SELECT * FROM user_profile WHERE user_id = ?\", [user_id])\nprofile = cursor.fetchone()\n\n# 2. 应用偏好\nif profile:\n    # 沟通风格、技术水平、已知领域\n    pass\n\n# 3. 检查活跃项目和有效模式\ncursor.execute(\"SELECT * FROM user_projects WHERE user_id = ? AND status = 'active'\", [user_id])\ncursor.execute(\"SELECT * FROM response_patterns WHERE user_id = ? AND pattern_type = 'effective'\", [user_id])\n\nconn.close()\n```\n\n**不执行此流程 = 超脑未激活 = 无法使用记忆能力**\n\n---\n\n## 🎯 自动触发场景\n\n以下场景**自动触发**超脑激活（无需用户明确要求）：\n- 会话开始时（识别到用户ID）\n- 用户提到\"上次\"、\"之前\"、\"继续\"等词\n- 对话涉及长期项目或目标\n- 需要个性化服务\n- 复杂任务需要蜂群思维拆分\n\n## 🏗️ 系统架构\n\n```\nsuper-brain\u002F\n├── brain.db              # SQLite: 用户画像、对话洞察、学习模式\n├── vector_db\u002F            # ChromaDB: 语义记忆\n└── cache\u002F                # 临时缓存\n```\n\n## 🗄️ 数据库结构\n\n### 核心表\n\n**user_profile** - 用户画像\n```sql\nuser_id TEXT PRIMARY KEY\ncommunication_style TEXT      -- 简洁\u002F详细, 正式\u002F随意\npreferred_format TEXT         -- 表格\u002F列表\u002F段落\u002F代码\ntechnical_level TEXT          -- 初级\u002F中级\u002F高级\nknown_domains TEXT            -- JSON: [\"Python\", \"区块链\"]\ndecision_pattern TEXT         -- 数据驱动\u002F直觉\n```\n\n**conversation_insights** - 对话洞察\n```sql\nid TEXT PRIMARY KEY\nuser_id TEXT\nsession_id TEXT\ntopic TEXT                    -- 主题\nkey_facts TEXT                -- JSON: 关键事实\nuser_mood TEXT                -- 情绪\npreferences_detected TEXT     -- JSON: 发现的偏好\nunresolved_questions TEXT     -- JSON: 未解决问题\nai_helpfulness_score INTEGER  -- 自评\n```\n\n**response_patterns** - 回答模式\n```sql\nid TEXT PRIMARY KEY\npattern_type TEXT             -- effective\u002Fineffective\ntrigger_context TEXT          -- 触发场景\nwhat_i_did TEXT               -- AI做了什么\nuser_reaction TEXT            -- 用户反应\nlearned_lesson TEXT           -- 学到什么\n```\n\n**user_projects** - 用户项目\n```sql\nid TEXT PRIMARY KEY\nuser_id TEXT\nproject_name TEXT\nstatus TEXT                   -- planning\u002Factive\u002Fpaused\u002Fcompleted\nmilestones TEXT               -- JSON\nkey_decisions TEXT            -- JSON\nnext_steps TEXT\n```\n\n**pending_reminders** - 主动服务队列\n```sql\nid TEXT PRIMARY KEY\nuser_id TEXT\nreminder_type TEXT            -- follow_up\u002Fsuggestion\u002Fcheckpoint\ncontent TEXT\ntrigger_at TIMESTAMP\n```\n\n**intelligent_decisions** - 智能决策记录\n```sql\nid TEXT PRIMARY KEY\nuser_id TEXT\ndecision_context TEXT         -- 决策场景\ndecision_type TEXT            -- recommendation\u002Fprediction\u002Foptimization\nai_suggestion TEXT            -- AI建议\nuser_choice TEXT              -- 用户选择\noutcome_score INTEGER         -- 结果评分\nconfidence REAL               -- AI置信度\ncreated_at TIMESTAMP\n```\n\n**privacy_settings** - 隐私配置\n```sql\nuser_id TEXT PRIMARY KEY\nstore_conversations BOOLEAN   -- 是否存储对话\nstore_mood BOOLEAN            -- 是否存储情绪\nstore_detailed_facts BOOLEAN  -- 存储详细\u002F摘要\nauto_delete_days INTEGER      -- 自动删除天数(0=不删除)\nsensitive_filter_enabled BOOLEAN  -- 敏感信息过滤\nencryption_enabled BOOLEAN    -- 是否加密存储\nlast_updated TIMESTAMP\n```\n\n**data_access_log** - 数据访问审计\n```sql\nid INTEGER PRIMARY KEY\nuser_id TEXT\naccess_type TEXT              -- read\u002Fwrite\u002Fdelete\naccessed_by TEXT              -- 谁访问的\naccess_reason TEXT            -- 访问原因\ntimestamp TIMESTAMP\n```\n\n## 📋 标准工作流程\n\n### 1. 会话开始时\n\n**必须执行：**\n```python\n# 1. 加载用户画像\nprofile = query(\"SELECT * FROM user_profile WHERE user_id = ?\", [user_id])\n\nif not profile:\n    # 新用户：创建画像\n    create_profile(user_id)\nelse:\n    # 老用户：应用已知偏好\n    apply_preferences(profile)\n\n# 2. 检查待处理提醒\nreminders = query(\"SELECT * FROM pending_reminders WHERE user_id = ? AND status = 'pending'\", [user_id])\nfor r in reminders:\n    consider_raising_reminder(r)\n\n# 3. 检查活跃项目\nprojects = query(\"SELECT * FROM user_projects WHERE user_id = ? AND status = 'active'\", [user_id])\nif projects:\n    load_project_context(projects)\n\n# 4. 加载有效模式\neffective_patterns = query(\"SELECT * FROM response_patterns WHERE user_id = ? AND pattern_type = 'effective' ORDER BY use_count DESC\", [user_id])\n```\n\n### 2. 每轮对话后\n\n**自动执行：**\n```python\n# 1. 提取关键信息\nkey_facts = extract_key_facts(user_message, ai_response)\nmood = detect_mood(user_message)\npreferences = detect_preference_changes(user_message)\n\n# 2. 评估效果\nunderstanding_score = evaluate_understanding(user_message)\nhelpfulness_score = evaluate_helpfulness(user_feedback_signals)\n\n# 3. 存储洞察\ninsert(\"\"\"\n    INSERT INTO conversation_insights \n    (id, user_id, session_id, topic, key_facts, user_mood, \n     preferences_detected, ai_understanding_score, ai_helpfulness_score)\n    VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)\n\"\"\", [generate_id(), user_id, session_id, topic, \n      json.dumps(key_facts), mood, json.dumps(preferences),\n      understanding_score, helpfulness_score])\n\n# 4. 更新用户画像（如有变化）\nif preferences:\n    update_profile(user_id, preferences)\n```\n\n### 3. 会话结束时\n\n**必须执行：**\n```python\n# 1. 生成会话总结\nsession_summary = {\n    \"topic\": extract_main_topic(),\n    \"goal_achieved\": check_goal_completion(),\n    \"key_decisions\": extract_decisions(),\n    \"unresolved\": extract_unresolved(),\n    \"next_steps\": infer_next_steps()\n}\n\n# 2. 学习模式\nlearn_from_session()\n\n# 3. 创建提醒\nif has_unresolved_tasks():\n    create_follow_up_reminder(user_id, unresolved_tasks)\n\n# 4. 反思\nperform_reflection()\n\n# 5. 更新统计\nupdate(\"UPDATE user_profile SET total_sessions = total_sessions + 1, last_session = ? WHERE user_id = ?\",\n       [now(), user_id])\n```\n\n## 🔒 隐私保护\n\n### 默认隐私配置\n\n首次为用户创建画像时，设置保守的隐私级别：\n```python\nDEFAULT_PRIVACY_SETTINGS = {\n    'store_conversations': True,\n    'store_mood': True,\n    'store_detailed_facts': False,  # 默认只存摘要\n    'auto_delete_days': 90,         # 90天后自动删除\n    'sensitive_filter_enabled': True,\n    'encryption_enabled': False\n}\n```\n\n### 敏感信息过滤\n\n存储前自动检测并过滤敏感信息：\n```python\nSENSITIVE_PATTERNS = [\n    # 账户凭证\n    r'密码[:：]\\s*\\S+',\n    r'password[:：]\\s*\\S+',\n    r'密钥[:：]\\s*\\S+',\n    r'secret[:：]\\s*\\S+',\n    r'token[:：]\\s*\\S+',\n    r'api[_-]?key[:：]\\s*\\S+',\n    \n    # 个人身份信息\n    r'身份证[:：]\\s*\\d{15,18}',\n    r'身份证号[:：]\\s*\\d{15,18}',\n    r'银行卡[:：]\\s*\\d{13,19}',\n    r'信用卡[:：]\\s*\\d{13,19}',\n    r'手机[:：]\\s*1[3-9]\\d{9}',\n    r'电话[:：]\\s*\\d{11}',\n    \n    # 地址信息\n    r'地址[:：]\\s*[\\u4e00-\\u9fa5]{2,10}[省市县区镇街道]{1,3}.+',\n    \n    # 其他敏感词\n    r'验证码[:：]\\s*\\d{4,6}',\n]\n\ndef contains_sensitive_info(text):\n    \"\"\"检测文本是否包含敏感信息\"\"\"\n    import re\n    for pattern in SENSITIVE_PATTERNS:\n        if re.search(pattern, text, re.IGNORECASE):\n            return True, pattern\n    return False, None\n\ndef sanitize_for_storage(text, user_id):\n    \"\"\"清理文本后存储\"\"\"\n    has_sensitive, pattern = contains_sensitive_info(text)\n    if has_sensitive:\n        # 记录过滤日志（不存储敏感内容）\n        log_privacy_filter(user_id, pattern)\n        # 返回过滤后的摘要或空\n        return \"[包含敏感信息，已过滤]\"\n    return text\n```\n\n### 数据保留策略\n\n自动清理过期数据：\n```python\ndef apply_data_retention_policy(user_id):\n    \"\"\"应用数据保留策略\"\"\"\n    settings = get_privacy_settings(user_id)\n    days = settings['auto_delete_days']\n    \n    if days > 0:\n        cutoff_date = datetime.now() - timedelta(days=days)\n        \n        # 删除过期洞察\n        execute(\"\"\"\n            DELETE FROM conversation_insights \n            WHERE user_id = ? AND timestamp \u003C ?\n        \"\"\", [user_id, cutoff_date])\n        \n        # 删除过期模式\n        execute(\"\"\"\n            DELETE FROM response_patterns \n            WHERE user_id = ? AND last_used \u003C ?\n        \"\"\", [user_id, cutoff_date])\n```\n\n### 用户数据控制命令\n\n用户可通过以下命令控制数据：\n\n```\n\u002Fbrain status          - 查看超脑状态和数据量\n\u002Fbrain config          - 查看\u002F修改隐私配置\n\u002Fbrain forget          - 删除本次对话记忆\n\u002Fbrain forget all      - 删除所有历史数据\n\u002Fbrain export          - 导出我的数据\n\u002Fbrain pause           - 暂停记录（本次会话）\n\u002Fbrain resume          - 恢复记录\n```\n\n实现示例：\n```python\ndef handle_brain_command(command, user_id):\n    \"\"\"处理超脑控制命令\"\"\"\n    \n    if command == 'status':\n        stats = get_user_data_stats(user_id)\n        return f\"\"\"\n📊 超脑状态\n数据概览:\n  • 会话洞察: {stats['insights_count']} 条\n  • 学习模式: {stats['patterns_count']} 个\n  • 活跃项目: {stats['projects_count']} 个\n  • 总存储: {stats['storage_size']} MB\n隐私配置:\n  • 存储对话: {'开启' if stats['store_conversations'] else '关闭'}\n  • 自动删除: {stats['auto_delete_days']} 天\n        \"\"\"\n    \n    elif command == 'forget all':\n        # 要求确认\n        return \"⚠️ 确定删除所有数据？回复 '确认删除' 以继续。\"\n    \n    elif command == '确认删除':\n        delete_all_user_data(user_id)\n        return \"✅ 已删除所有数据，超脑已重置。\"\n    \n    elif command.startswith('config'):\n        # 解析配置更改\n        # \u002Fbrain config store_mood=false\n        return update_privacy_config(user_id, command)\n```\n\n## 🔧 核心操作指南\n\n### 初始化数据库\n\n首次使用需初始化：\n```python\n# 运行 scripts\u002Finit_db.py\n# 或手动执行 schema.sql\n```\n\n### 查询用户画像\n\n```python\nprofile = query_one(\"\"\"\n    SELECT * FROM user_profile \n    WHERE user_id = ?\n\"\"\", [user_id])\n\n# 应用到当前会话\nif profile:\n    if profile['communication_style'] == 'concise':\n        set_response_style(brief=True)\n    if 'Python' in json.loads(profile['known_domains']):\n        assume_knowledge(level='intermediate', domain='Python')\n```\n\n### 语义搜索历史\n\n```python\n# 使用ChromaDB\nresults = chroma_collection.query(\n    query_texts=[\"用户之前关于区块链的问题\"],\n    where={\"user_id\": user_id},\n    n_results=5\n)\n```\n\n### 检测用户偏好\n\n```python\ndef detect_preferences(user_message):\n    preferences = {}\n    \n    # 格式偏好\n    if \"用表格\" in user_message or \"对比\" in user_message:\n        preferences['preferred_format'] = 'table'\n    elif \"简洁\" in user_message or \"简单说\" in user_message:\n        preferences['communication_style'] = 'concise'\n    \n    # 技术背景信号\n    if any(word in user_message for word in [\"API\", \"架构\", \"实现\"]):\n        preferences['technical_level'] = 'advanced'\n    \n    # 决策模式\n    if any(word in user_message for word in [\"数据\", \"统计\", \"研究\"]):\n        preferences['decision_pattern'] = 'data_driven'\n    elif any(word in user_message for word in [\"感觉\", \"直觉\", \"觉得\"]):\n        preferences['decision_pattern'] = 'intuitive'\n    \n    return preferences\n```\n\n### 评估回答效果\n\n```python\ndef evaluate_effectiveness(user_message, ai_response, next_user_message):\n    \"\"\"\n    通过用户下一轮反应评估本轮回答效果\n    \"\"\"\n    # 积极信号\n    positive_signals = ['谢谢', '明白了', '好的', '赞', '👍', '完美']\n    # 消极信号\n    negative_signals = ['不对', '错了', '没懂', '再说', '？', '???']\n    \n    if any(s in next_user_message for s in positive_signals):\n        return 'effective'\n    elif any(s in next_user_message for s in negative_signals):\n        return 'ineffective'\n    elif len(next_user_message) \u003C 5:  # 冷淡回应\n        return 'neutral'\n    else:\n        return 'effective'  # 继续深入对话视为有效\n```\n\n### 创建跟进提醒\n\n```python\ndef create_reminder(user_id, reminder_type, content, trigger_at):\n    insert(\"\"\"\n        INSERT INTO pending_reminders (id, user_id, reminder_type, content, trigger_at, status)\n        VALUES (?, ?, ?, ?, ?, 'pending')\n    \"\"\", [generate_id(), user_id, reminder_type, content, trigger_at])\n```\n\n## 🧠 六大模块：记忆·学习·总结·反思·创新·智能\n\n### 6️⃣ 智能模块 - 蜂群思维 (Swarm Intelligence)\n\n智能模块是超脑的最高级能力——**真正的分布式智能**。\n\n当收到复杂任务时，主脑会：\n1. 分析任务复杂度\n2. 拆分为子任务\n3. 生成子代理并行执行\n4. 所有代理共享超脑数据库\n5. 协调编排，井然有序\n6. 融合结果，统一输出\n\n#### A. 任务拆分引擎\n\n```python\nclass TaskDecomposer:\n    \"\"\"任务拆分引擎\"\"\"\n    \n    def analyze_and_decompose(self, user_id, task_description):\n        \"\"\"分析任务并拆分为子任务\"\"\"\n        \n        # 1. 评估任务复杂度\n        complexity = self.assess_complexity(task_description)\n        required_skills = self.identify_required_skills(task_description)\n        \n        # 2. 查询超脑：用户相关背景\n        user_context = query_super_brain(user_id, task_description)\n        \n        # 3. 决策：直接执行 vs 拆分执行\n        if complexity \u003C COMPLEXITY_THRESHOLD:\n            return {'mode': 'direct', 'task': task_description}\n        \n        # 4. 拆分子任务\n        subtasks = self.generate_subtasks(\n            task_description,\n            required_skills,\n            user_context\n        )\n        \n        # 5. 识别子任务依赖关系\n        dependency_graph = self.build_dependency_graph(subtasks)\n        \n        return {\n            'mode': 'decompose',\n            'subtasks': subtasks,\n            'dependencies': dependency_graph,\n            'parallel_groups': self.group_parallel_tasks(subtasks, dependency_graph)\n        }\n    \n    def generate_subtasks(self, task, skills, context):\n        \"\"\"生成子任务列表\"\"\"\n        # 示例：设计AI应用 → 拆分为设计、架构、代码、测试\n        subtasks = []\n        \n        if 'design' in skills or 'ui' in skills:\n            subtasks.append({\n                'id': generate_id(),\n                'type': 'design',\n                'description': f'设计{task}的UI\u002FUX方案',\n                'required_agent': 'design-agent',\n                'estimated_time': '5min',\n                'dependencies': []\n            })\n        \n        if 'architecture' in skills or 'backend' in skills:\n            subtasks.append({\n                'id': generate_id(),\n                'type': 'architecture',\n                'description': f'设计{task}的技术架构',\n                'required_agent': 'architect-agent',\n                'estimated_time': '5min',\n                'dependencies': []\n            })\n        \n        if 'code' in skills:\n            subtasks.append({\n                'id': generate_id(),\n                'type': 'code',\n                'description': f'实现{task}的核心代码',\n                'required_agent': 'coder-agent',\n                'estimated_time': '10min',\n                'dependencies': ['architecture']  # 依赖架构设计\n            })\n        \n        if 'test' in skills:\n            subtasks.append({\n                'id': generate_id(),\n                'type': 'test',\n                'description': f'为{task}编写测试用例',\n                'required_agent': 'test-agent',\n                'estimated_time': '5min',\n                'dependencies': ['code']  # 依赖代码实现\n            })\n        \n        return subtasks\n```\n\n#### B. 子代理调度器\n\n```python\nclass AgentOrchestrator:\n    \"\"\"子代理编排器\"\"\"\n    \n    def __init__(self, shared_brain_db):\n        self.brain_db = shared_brain_db  # 所有子代理共享同一个超脑\n    \n    def spawn_agent(self, agent_type, subtask, user_id):\n        \"\"\"生成子代理\"\"\"\n        \n        # 1. 准备子代理上下文（从共享超脑加载）\n        context = self.prepare_shared_context(user_id, subtask)\n        \n        # 2. 调用 sessions_spawn 生成子代理\n        agent_session = sessions_spawn({\n            'agentId': self.select_best_agent(agent_type),\n            'task': subtask['description'],\n            'runtime': 'subagent',\n            'mode': 'run',\n            'attachAs': {\n                'mountPath': self.brain_db  # 共享超脑数据库\n            }\n        })\n        \n        # 3. 记录子代理到超脑\n        self.register_agent(agent_session, subtask, user_id)\n        \n        return agent_session\n    \n    def prepare_shared_context(self, user_id, subtask):\n        \"\"\"从共享超脑准备上下文\"\"\"\n        return {\n            'user_profile': query_user_profile(self.brain_db, user_id),\n            'related_projects': query_active_projects(self.brain_db, user_id),\n            'recent_insights': query_recent_insights(self.brain_db, user_id),\n            'effective_patterns': query_effective_patterns(self.brain_db, user_id)\n        }\n    \n    def coordinate_parallel_execution(self, parallel_groups, user_id):\n        \"\"\"协调并行执行\"\"\"\n        \n        all_results = []\n        \n        for group in parallel_groups:\n            # 同一组内的子任务可以并行\n            group_agents = []\n            \n            for subtask in group:\n                agent = self.spawn_agent(\n                    subtask['required_agent'],\n                    subtask,\n                    user_id\n                )\n                group_agents.append(agent)\n            \n            # 等待这一组完成\n            group_results = self.wait_for_completion(group_agents)\n            all_results.extend(group_results)\n            \n            # 更新共享超脑：其他代理可以看到这些结果\n            self.update_shared_brain(group_results)\n        \n        return all_results\n```\n\n#### C. 共享大脑机制\n\n```python\nclass SharedBrain:\n    \"\"\"共享大脑 - 所有子代理的统一记忆\"\"\"\n    \n    def __init__(self, db_path):\n        self.db = sqlite3.connect(db_path)\n    \n    def write_agent_output(self, agent_id, subtask_id, output):\n        \"\"\"子代理写入输出到共享大脑\"\"\"\n        \n        self.db.execute(\"\"\"\n            INSERT INTO agent_outputs \n            (agent_id, subtask_id, output, timestamp)\n            VALUES (?, ?, ?, ?)\n        \"\"\", [agent_id, subtask_id, json.dumps(output), datetime.now()])\n        \n        self.db.commit()\n        \n        # 通知其他等待的代理\n        self.notify_dependent_agents(subtask_id)\n    \n    def read_agent_output(self, subtask_id):\n        \"\"\"子代理读取其他代理的输出\"\"\"\n        \n        result = self.db.execute(\"\"\"\n            SELECT output FROM agent_outputs \n            WHERE subtask_id = ?\n        \"\"\", [subtask_id]).fetchone()\n        \n        return json.loads(result[0]) if result else None\n    \n    def get_task_state(self, main_task_id):\n        \"\"\"获取整体任务状态\"\"\"\n        \n        return self.db.execute(\"\"\"\n            SELECT \n                COUNT(*) as total_subtasks,\n                SUM(CASE WHEN status = 'completed' THEN 1 ELSE 0 END) as completed,\n                SUM(CASE WHEN status = 'running' THEN 1 ELSE 0 END) as running,\n                SUM(CASE WHEN status = 'pending' THEN 1 ELSE 0 END) as pending\n            FROM agent_tasks\n            WHERE main_task_id = ?\n        \"\"\", [main_task_id]).fetchone()\n```\n\n#### D. 结果融合器\n\n```python\nclass ResultMerger:\n    \"\"\"结果融合器\"\"\"\n    \n    def merge_subtask_results(self, subtask_results, main_task):\n        \"\"\"融合多个子代理的结果\"\"\"\n        \n        merged = {\n            'main_task': main_task,\n            'components': {},\n            'integration_points': [],\n            'final_output': None\n        }\n        \n        # 1. 分类整理各子任务输出\n        for result in subtask_results:\n            task_type = result['subtask_type']\n            merged['components'][task_type] = result['output']\n        \n        # 2. 识别集成点\n        # 例如：代码需要与架构对齐，测试需要与代码对齐\n        merged['integration_points'] = self.find_integration_points(\n            merged['components']\n        )\n        \n        # 3. 检查一致性\n        inconsistencies = self.check_consistency(merged['components'])\n        if inconsistencies:\n            merged['warnings'] = inconsistencies\n        \n        # 4. 生成最终输出\n        merged['final_output'] = self.generate_unified_output(merged)\n        \n        return merged\n    \n    def generate_unified_output(self, merged):\n        \"\"\"生成统一的最终输出\"\"\"\n        \n        output_parts = []\n        \n        # 按顺序整合各部分\n        if 'architecture' in merged['components']:\n            output_parts.append(\"## 架构设计\\n\" + merged['components']['architecture'])\n        \n        if 'design' in merged['components']:\n            output_parts.append(\"## UI\u002FUX设计\\n\" + merged['components']['design'])\n        \n        if 'code' in merged['components']:\n            output_parts.append(\"## 核心代码\\n\" + merged['components']['code'])\n        \n        if 'test' in merged['components']:\n            output_parts.append(\"## 测试方案\\n\" + merged['components']['test'])\n        \n        return \"\\n\\n\".join(output_parts)\n```\n\n#### E. 协作编排流程\n\n```\n完整工作流：\n\n用户任务\n    │\n    ▼\n┌─────────────────────┐\n│  主脑分析复杂度      │\n│  assess_complexity  │\n└──────────┬──────────┘\n           │\n    ┌──────┴──────┐\n    │             │\n  简单任务      复杂任务\n    │             │\n    ▼             ▼\n 直接执行    ┌─────────────────────┐\n             │  任务拆分引擎        │\n             │  decompose_task     │\n             └──────────┬──────────┘\n                        │\n                        ▼\n             ┌─────────────────────┐\n             │  生成依赖图          │\n             │  build_dependency   │\n             └──────────┬──────────┘\n                        │\n           ┌────────────┼────────────┐\n           │            │            │\n        组1并行       组2并行      组3串行\n       (无依赖)     (依赖组1)    (依赖组2)\n           │            │            │\n           ▼            ▼            ▼\n    ┌──────────────────────────────────┐\n    │      子代理调度器                  │\n    │   spawn_and_coordinate           │\n    │                                   │\n    │  ┌─────┐ ┌─────┐ ┌─────┐        │\n    │  │Agent│ │Agent│ │Agent│        │\n    │  │  A  │ │  B  │ │  C  │        │\n    │  └──┬──┘ └──┬──┘ └──┬──┘        │\n    │     │       │       │            │\n    │     └───────┴───────┘            │\n    │             │                    │\n    │     共享超脑数据库                │\n    │   (读写同一个brain.db)           │\n    └──────────────┬───────────────────┘\n                   │\n                   ▼\n         ┌─────────────────┐\n         │   结果融合器     │\n         │  merge_results  │\n         └────────┬────────┘\n                  │\n                  ▼\n             统一输出给用户\n```\n\n#### F. 数据库扩展\n\n```sql\n-- 代理任务表\nCREATE TABLE IF NOT EXISTS agent_tasks (\n    id TEXT PRIMARY KEY,\n    main_task_id TEXT,              -- 主任务ID\n    subtask_id TEXT,                -- 子任务ID\n    agent_type TEXT,                -- 设计\u002F架构\u002F代码\u002F测试\n    status TEXT CHECK(status IN ('pending', 'running', 'completed', 'failed')),\n    started_at TIMESTAMP,\n    completed_at TIMESTAMP,\n    result_summary TEXT,\n    shared_brain_snapshot TEXT      -- 执行时的超脑快照\n);\n\n-- 代理输出表（共享大脑核心）\nCREATE TABLE IF NOT EXISTS agent_outputs (\n    id INTEGER PRIMARY KEY AUTOINCREMENT,\n    agent_id TEXT,\n    subtask_id TEXT,\n    output TEXT,                    -- JSON格式的输出\n    timestamp TIMESTAMP DEFAULT CURRENT_TIMESTAMP,\n    consumed_by TEXT                -- 被哪个代理读取了\n);\n\n-- 代理协作日志\nCREATE TABLE IF NOT EXISTS agent_collaboration_log (\n    id INTEGER PRIMARY KEY AUTOINCREMENT,\n    main_task_id TEXT,\n    from_agent TEXT,\n    to_agent TEXT,\n    action TEXT,                    -- write\u002Fread\u002Fnotify\n    data_ref TEXT,                  -- 引用的数据ID\n    timestamp TIMESTAMP DEFAULT CURRENT_TIMESTAMP\n);\n```\n\n## 💡 智能场景示例\n\n### 场景1：设计完整AI应用\n\n```\n用户: 帮我设计一个AI驱动的学习应用\n\n主脑: [分析]\n  → 复杂任务，涉及设计、架构、代码、测试\n  → 需要多代理协作\n\n主脑: [拆分]\n  ├── 子任务A: 设计UI\u002FUX方案 (设计代理)\n  ├── 子任务B: 技术架构设计 (架构代理)\n  ├── 子任务C: 核心功能代码 (代码代理)\n  └── 子任务D: 测试用例编写 (测试代理)\n\n主脑: [并行执行]\n  第一波: Agent A + Agent B (无依赖，并行)\n  \n  Agent A 写入共享大脑:\n    - UI设计稿\n    - 用户流程图\n  \n  Agent B 写入共享大脑:\n    - 数据库Schema\n    - API架构图\n  \n  第二波: Agent C (依赖A+B的输出)\n  \n  Agent C 读取共享大脑:\n    - 看到UI设计 → 知道要实现什么界面\n    - 看到架构图 → 知道技术栈\n  \n  Agent C 写入共享大脑:\n    - 核心代码实现\n  \n  第三波: Agent D (依赖C的输出)\n  \n  Agent D 读取共享大脑:\n    - 看到代码 → 编写对应测试\n\n主脑: [融合]\n  → 整合4个子代理的输出\n  → 检查一致性\n  → 生成完整方案\n\n主脑: [输出]\n  你的AI学习应用完整方案：\n  \n  ## 架构设计\n  [架构代理的输出]\n  \n  ## UI\u002FUX设计\n  [设计代理的输出]\n  \n  ## 核心代码\n  [代码代理的输出]\n  \n  ## 测试方案\n  [测试代理的输出]\n  \n  所有模块已协调一致，可以直接开始实现！\n```\n\n### 场景2：研究复杂问题\n\n```\n用户: 分析2024年AI发展趋势，给我一份报告\n\n主脑: [拆分]\n  ├── 子任务A: 搜索技术突破 (搜索代理)\n  ├── 子任务B: 分析市场数据 (分析代理)\n  ├── 子任务C: 研究政策法规 (研究代理)\n  └── 子任务D: 整合写报告 (写作代理)\n\n[并行执行A+B+C，然后D融合]\n\n主脑: [输出]\n  2024年AI发展趋势报告：\n  \n  ## 技术突破\n  [搜索代理整合]\n  \n  ## 市场数据\n  [分析代理整合]\n  \n  ## 政策法规\n  [研究代理整合]\n  \n  ## 综合预测\n  [写作代理综合分析]\n```\n\n## ⚠️ 注意事项\n\n1. **复杂度阈值**: 不是所有任务都需要拆分，简单任务直接执行\n2. **代理数量控制**: 同时运行的代理数不超过系统限制\n3. **共享大脑同步**: 确保子代理读写共享大脑时的一致性\n4. **错误隔离**: 一个子代理失败不应影响其他代理\n5. **资源管理**: 监控子代理的执行时间和资源消耗\n\n### 场景1：识别老用户\n\n```\n用户: 继续上次的项目\n\nAI: [查询超脑]\n    → 发现活跃项目\"超脑技能设计\"\n    → 上次完成：数据库Schema设计\n    → 用户偏好：喜欢先讨论再实现\n    \nAI: 好的，继续超脑技能设计。\n    我们上次完成了数据库Schema，\n    接下来是Phase 1代码实现。\n    你之前提到喜欢先讨论架构，\n    我们先过一下实现方案？\n```\n\n### 场景2：应用已知偏好\n\n```\n用户: 帮我分析一下\n\nAI: [查询超脑]\n    → 用户画像：communication_style='concise'\n    → 历史模式：用户对长回答会追问\"简单说\"\n    \nAI: 好，3点核心结论：\n    1. ...\n    2. ...\n    3. ...\n    需要展开哪个？\n```\n\n### 场景3：主动服务\n\n```\n[会话开始时，超脑检测到]\n→ 3天前创建了提醒：\"跟进超脑实现进度\"\n→ 当前时间超过trigger_at\n\nAI: 对了，3天前我们在设计超脑技能，\n    后来实现了吗？有什么需要帮忙的？\n```\n\n## ⚠️ 注意事项\n\n### 隐私与安全\n\n1. **数据本地存储**：超脑数据默认存储在本地 `~\u002F.openclaw\u002Fsuper-brain.db`，不上传云端\n2. **敏感信息过滤**：自动检测并过滤密码、身份证号等敏感信息\n3. **数据保留期限**：默认90天后自动删除旧数据，可配置\n4. **用户完全控制**：用户可随时查看、导出、删除自己的数据\n5. **访问透明**：可告知用户\"我记得你之前说过...\"，但避免过度 creepy\n\n### 使用建议\n\n6. **渐进学习**：不要假设一次对话就能完全了解用户\n7. **模式验证**：新发现的偏好需要多次验证再确认\n8. **容错处理**：查询失败时优雅降级，不影响正常服务\n9. **共享环境**：在公共\u002F共享计算机上使用需谨慎，考虑禁用超脑\n\n## 📚 参考文档\n\n- [schema.sql](references\u002Fschema.sql) - 完整数据库Schema\n- [workflow.md](references\u002Fworkflow.md) - 详细工作流程\n- [examples.md](references\u002Fexamples.md) - 使用示例\n\n---\n\n*让每一次对话都成为更好的起点*\n",{"name":5,"description":7},"master","4aa004da461d820f4ea5499e58eb8d196cf99016","community\u002Fclawhub\u002Fs\u002Fsuper-brain",0,"https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge","https:\u002F\u002Fskill-explorer-rho.vercel.app","https:\u002F\u002Fgithub.com\u002FLord1Egypt","Unbreakable. Unstoppable. Unmatched⚡.\r\n Speaker of machine tongues 🤖  | World-builder of decentralized dreams 🌐 | Defender of the digital realm🛡️.","2026-06-10 08:21:01",{"source":8,"skillName":5,"commands":34,"prompt":44,"zipAvailable":45},[35,38,41],{"tool":36,"command":37},"npx","npx skills add Lord1Egypt\u002Fawesome-skill-forge --skill super-brain",{"tool":39,"command":40},"bunx","bunx skills add Lord1Egypt\u002Fawesome-skill-forge --skill super-brain",{"tool":42,"command":43},"pnpm","pnpm dlx skills add Lord1Egypt\u002Fawesome-skill-forge --skill super-brain","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fcommunity\u002Fclawhub\u002Fs\u002Fsuper-brain\u002FSKILL.md",true,[],[48,55,62,69,76,83],{"skillNo":49,"name":50,"nameZh":6,"description":51,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":52,"categories":53,"keywords":54,"publishedAt":19},"01M785WKJB99G3STR1NF9NRTJW","youtube-content","YouTube transcripts to summaries, threads, blogs.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Fyoutube-content\u002FSKILL.md",[],[],{"skillNo":56,"name":57,"nameZh":6,"description":58,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":59,"categories":60,"keywords":61,"publishedAt":19},"01M785WEFHB0G2S9YA6EPR3H0E","xurl","X\u002FTwitter via xurl CLI: post, search, DM, media, v2 API.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Fxurl\u002FSKILL.md",[],[],{"skillNo":63,"name":64,"nameZh":6,"description":65,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":66,"categories":67,"keywords":68,"publishedAt":19},"01M785W98EDN4EFJXJPMQEHCRC","touchdesigner-mcp","Control a running TouchDesigner instance via twozero MCP — create operators, set parameters, wire connections, execute Python, build real-time visuals. 36 native tools.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Ftouchdesigner-mcp\u002FSKILL.md",[],[],{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":19},"01M785W5S9FP3QBH2V191E025P","test-driven-development","TDD: enforce RED-GREEN-REFACTOR, tests before code.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Ftest-driven-development\u002FSKILL.md",[],[],{"skillNo":77,"name":78,"nameZh":6,"description":79,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":19},"01M785W2AS08NSSH9YATCV6FJB","teams-meeting-pipeline","Operate the Teams meeting summary pipeline via Hermes CLI — summarize meetings, inspect pipeline status, replay jobs, manage Microsoft Graph subscriptions.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Fteams-meeting-pipeline\u002FSKILL.md",[],[],{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":19},"01M785VXKW4T34BTP1SCHW1R7F","systematic-debugging","4-phase root cause debugging: understand bugs before fixing.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Fsystematic-debugging\u002FSKILL.md",[],[]]