#!/usr/bin/env python3 """ HumanOS构建器：根据框架生成HumanOS Skill """ import json from pathlib import Path from typing import Dict, List class HumanOSBuilder: """HumanOS构建器""" # 4种HumanOS类型 OS_TYPES = { 'human': { 'name': 'HumanType', 'template': 'human-template.md', 'description': '人物型，蒸馏特定人物的思维框架' }, 'theme': { 'name': 'ThemeType', 'template': 'theme-template.md', 'description': '主题型，综合多人视角的主题框架' }, 'scenario': { 'name': 'ScenarioType', 'template': 'scenario-template.md', 'description': '场景型，特定场景的决策系统' }, 'methodology': { 'name': 'MethodologyType', 'template': 'methodology-template.md', 'description': '方法论型，完整的方法论实施框架' } } def __init__(self): self.built_os = {} def build_humanos(self, framework: Dict, os_type: str, template_dir: str, output_file: str) -> Dict: """构建HumanOS""" # 加载模板 template_path = Path(template_dir) / self.OS_TYPES[os_type]['template'] if not template_path.exists(): raise FileNotFoundError(f"模板文件不存在: {template_path}") with open(template_path, 'r', encoding='utf-8') as f: template_content = f.read() # 填充模板 skill_content = self._fill_template(template_content, framework, os_type) # 生成Agentic Protocol agentic_protocol = self._generate_agentic_protocol(framework) # 整合到Skill内容中 skill_content = self._integrate_agentic_protocol(skill_content, agentic_protocol) # 写入输出文件 output_path = Path(output_file) output_path.parent.mkdir(parents=True, exist_ok=True) with open(output_path, 'w', encoding='utf-8') as f: f.write(skill_content) # 生成构建结果 build_result = { 'os_type': os_type, 'target': framework.get('target', ''), 'output_file': str(output_path), 'framework_summary': framework.get('summary', {}), 'agentic_protocol': agentic_protocol, 'build_status': 'success' } self.built_os = build_result return build_result def _fill_template(self, template: str, framework: Dict, os_type: str) -> str: """填充模板""" # 准备替换变量 variables = { 'TARGET': framework.get('target', ''), 'OS_TYPE': self.OS_TYPES[os_type]['name'], 'DESCRIPTION': self.OS_TYPES[os_type]['description'], 'MENTAL_MODELS': self._format_mental_models( framework.get('mental_models', []) ), 'DECISION_HEURISTICS': self._format_decision_heuristics( framework.get('decision_heuristics', []) ), 'EXPRESSION_DNA': self._format_expression_dna( framework.get('expression_dna', {}) ), 'VALUES': self._format_values( framework.get('values', []) ), 'TOOLKIT': self._format_toolkit( framework.get('toolkit', []) ), 'INTERNAL_TENSIONS': self._format_internal_tensions( framework.get('internal_tensions', []) ), 'HONEST_BOUNDARIES': self._format_honest_boundaries( framework.get('honest_boundaries', []) ) } # 替换变量 content = template for var_name, var_value in variables.items(): placeholder = f"{{{{{var_name}}}}}" content = content.replace(placeholder, str(var_value)) return content def _format_mental_models(self, models: List[Dict]) -> str: """格式化心智模型""" formatted = [] for model in models: name = model.get('name', '') description = model.get('description', '') validation = model.get('validation', {}) # 只包含通过验证的模型 if validation.get('passed'): formatted.append(f"- **{name}**: {description}") return '\n'.join(formatted) if formatted else '无通过验证的心智模型' def _format_decision_heuristics(self, heuristics: List[str]) -> str: """格式化决策启发式""" formatted = [] for heuristic in heuristics[:10]: formatted.append(f"- {heuristic}") return '\n'.join(formatted) if formatted else '无决策启发式' def _format_expression_dna(self, dna: Dict) -> str: """格式化表达DNA""" formatted = [] if dna.get('sentence_patterns'): formatted.append("**句式模式**:") for pattern in dna['sentence_patterns'][:5]: formatted.append(f" - {pattern}") if dna.get('vocabulary'): formatted.append("**特色词汇**:") for vocab in dna['vocabulary'][:10]: formatted.append(f" - {vocab}") if dna.get('rhythm'): formatted.append(f"**节奏**: {dna['rhythm']}") return '\n'.join(formatted) if formatted else '无表达DNA' def _format_values(self, values: List[str]) -> str: """格式化价值观""" formatted = [] for value in values[:5]: formatted.append(f"- {value}") return '\n'.join(formatted) if formatted else '无价值观' def _format_toolkit(self, tools: List[str]) -> str: """格式化工具箱""" formatted = [] for tool in tools[:10]: formatted.append(f"- {tool}") return '\n'.join(formatted) if formatted else '无工具' def _format_internal_tensions(self, tensions: List[Dict]) -> str: """格式化内在张力""" formatted = [] for tension in tensions[:5]: tension_desc = tension.get('tension', '') description = tension.get('description', '') formatted.append(f"- **{tension_desc}**: {description}") return '\n'.join(formatted) if formatted else '无内在张力' def _format_honest_boundaries(self, boundaries: List[str]) -> str: """格式化诚实边界""" formatted = [] for boundary in boundaries[:5]: formatted.append(f"- {boundary}") return '\n'.join(formatted) if formatted else '无诚实边界' def _generate_agentic_protocol(self, framework: Dict) -> Dict: """生成Agentic Protocol""" # 从心智模型推导研究维度 mental_models = framework.get('mental_models', []) validated_models = [m for m in mental_models if m.get('validation', {}).get('passed')] # 生成研究步骤 research_steps = self._generate_research_steps(validated_models) agentic_protocol = { 'version': '1.0', 'target': framework.get('target', ''), 'description': f'{framework.get("target", "")}的研究和回答协议', 'steps': [ { 'step': 1, 'name': '问题分类', 'description': '识别用户问题的类型和维度', 'actions': [ '分析问题的领域', '识别相关的决策维度', '确定问题的复杂度' ] }, { 'step': 2, 'name': '研究', 'description': f'应用{framework.get("target", "")}的心智模型进行深入分析', 'actions': research_steps }, { 'step': 3, 'name': '回答', 'description': '生成结构化的回答，体现独特的思维框架', 'actions': [ '使用决策启发式推导结论', '体现表达DNA的风格', '标注诚实边界和局限性' ] } ], 'mental_models_used': [m.get('name', '') for m in validated_models], 'decision_heuristics_used': framework.get('decision_heuristics', [])[:5] } return agentic_protocol def _generate_research_steps(self, mental_models: List[Dict]) -> List[str]: """生成研究步骤""" steps = [] # 基于心智模型生成研究步骤 for model in mental_models: model_name = model.get('name', '') steps.append(f"应用{model_name}心智模型分析问题") # 添加通用步骤 steps.extend([ "识别问题的核心矛盾", "应用决策启发式进行推理", "评估不同方案的优劣势" ]) return steps def _integrate_agentic_protocol(self, skill_content: str, agentic_protocol: Dict) -> str: """整合Agentic Protocol到Skill内容""" # 将Agentic Protocol转换为Markdown格式 protocol_md = self._format_agentic_protocol(agentic_protocol) # 替换模板中的Agentic Protocol占位符 skill_content = skill_content.replace('{{AGENTIC_PROTOCOL}}', protocol_md) return skill_content def _format_agentic_protocol(self, protocol: Dict) -> str: """格式化Agentic Protocol""" formatted = f"""## Agentic Protocol {protocol.get('description', '')} ### 步骤 """ for step_info in protocol.get('steps', []): formatted += f""" #### 步骤 {step_info['step']}: {step_info['name']} {step_info['description']} **行动**: """ for action in step_info.get('actions', []): formatted += f"- {action}\n" formatted += f""" ### 应用心智模型 """ for model_name in protocol.get('mental_models_used', []): formatted += f"- {model_name}\n" formatted += f""" ### 决策启发式 """ for heuristic in protocol.get('decision_heuristics_used', []): formatted += f"- {heuristic}\n" return formatted def main(): import argparse parser = argparse.ArgumentParser(description='HumanOS构建器') parser.add_argument('--framework', type=str, required=True, help='框架JSON文件路径') parser.add_argument('--template-dir', type=str, required=True, help='模板目录') parser.add_argument('--os-type', type=str, required=True, choices=['human', 'theme', 'scenario', 'methodology'], help='HumanOS类型') parser.add_argument('--output', type=str, required=True, help='输出SKILL.md文件路径') args = parser.parse_args() # 读取框架 with open(args.framework, 'r', encoding='utf-8') as f: framework = json.load(f) # 构建HumanOS builder = HumanOSBuilder() result = builder.build_humanos( framework, args.os_type, args.template_dir, args.output ) print(f"HumanOS构建完成: {args.output}") print(json.dumps(result, ensure_ascii=False, indent=2)) if __name__ == '__main__': main()