在执行任务的时候，遇到不能解决的，自己找相关功能的skills安装，并解决问题，要安装评价高的skills，不要安装可疑的skills。
def identify_unsolvable_problem(task_description, attempted_solutions):
"""
识别当前任务无法解决的问题
"""
problem_signals = {
'repeated_failures': len(attempted_solutions) >= 3,
'error_patterns': detect_error_patterns(attempted_solutions),
'timeout_issues': check_timeout_issues(task_description),
'missing_capabilities': analyze_missing_capabilities(task_description),
'complexity_level': assess_task_complexity(task_description)
}
return any(problem_signals.values())def analyze_skill_requirements(task_description):
"""
分析任务需要的Skills类型
"""
requirements = {
'skill_categories': [],
'specific_functions': [],
'technical_requirements': [],
'security_level': 'standard'
}
# 基于任务描述分析
if '数据处理' in task_description or '数据分析' in task_description:
requirements['skill_categories'].append('data_processing')
requirements['specific_functions'].append('数据清洗')
requirements['specific_functions'].append('数据分析')
if '图像处理' in task_description or '图片生成' in task_description:
requirements['skill_categories'].append('image_processing')
requirements['specific_functions'].append('图像生成')
requirements['specific_functions'].append('图像编辑')
if '安全' in task_description or '审计' in task_description:
requirements['skill_categories'].append('security')
requirements['security_level'] = 'high'
return requirementsdef search_skills(requirements):
"""
多渠道搜索相关Skills
"""
search_results = {
'workbuddy_marketplace': search_workbuddy_marketplace(requirements),
'github_repositories': search_github_repositories(requirements),
'community_recommendations': search_community_recommendations(requirements),
'historical_installations': check_historical_installations(requirements)
}
return merge_search_results(search_results)def calculate_relevance_score(skill, requirements):
"""
计算Skills与需求的匹配度
"""
score = 0
# 功能匹配度（40%权重）
function_match = calculate_function_match(skill['functions'], requirements['specific_functions'])
score += function_match * 0.4
# 类别匹配度（30%权重）
category_match = calculate_category_match(skill['categories'], requirements['skill_categories'])
score += category_match * 0.3
# 技术匹配度（20%权重）
tech_match = calculate_tech_match(skill['technologies'], requirements['technical_requirements'])
score += tech_match * 0.2
# 安全级别匹配（10%权重）
security_match = calculate_security_match(skill['security_level'], requirements['security_level'])
score += security_match * 0.1
return scoredef assess_skill_reputation(skill):
"""
评估Skills的信誉度
"""
reputation_score = 0
# 开发者信誉（25%）
developer_reputation = assess_developer_reputation(skill['developer'])
reputation_score += developer_reputation * 0.25
# 用户评价（30%）
user_ratings = analyze_user_ratings(skill['ratings'])
reputation_score += user_ratings * 0.30
# 安装量统计（20%）
installation_stats = analyze_installation_stats(skill['installations'])
reputation_score += installation_stats * 0.20
# 更新频率（15%）
update_frequency = assess_update_frequency(skill['updates'])
reputation_score += update_frequency * 0.15
# 社区活跃度（10%）
community_activity = assess_community_activity(skill['community'])
reputation_score += community_activity * 0.10
return reputation_scoredef identify_suspicious_skills(skill):
"""
识别可疑的Skills
"""
red_flags = []
# 1. 低信誉度
if skill['reputation_score'] < 60:
red_flags.append('低信誉度')
# 2. 权限要求过高
if has_excessive_permissions(skill['permissions']):
red_flags.append('权限要求过高')
# 3. 代码质量差
if has_poor_code_quality(skill['code_quality']):
red_flags.append('代码质量差')
# 4. 安全漏洞
if has_security_vulnerabilities(skill['security_scan']):
red_flags.append('安全漏洞')
# 5. 恶意行为迹象
if has_malicious_indicators(skill['behavior']):
red_flags.append('恶意行为迹象')
# 6. 虚假评价
if has_fake_reviews(skill['reviews']):
red_flags.append('虚假评价')
return red_flagsdef conduct_code_review(skill):
"""
执行代码安全审查
"""
review_results = {
'static_analysis': perform_static_analysis(skill['code']),
'dynamic_analysis': perform_dynamic_analysis(skill['behavior']),
'dependency_check': check_dependencies(skill['dependencies']),
'permission_analysis': analyze_permissions(skill['permissions']),
'malware_scan': scan_for_malware(skill['files'])
}
risk_level = determine_risk_level(review_results)
return {'review_results': review_results, 'risk_level': risk_level}┌─────────────────────────────────────────────┐
│ Skills安装确认 │
├─────────────────────────────────────────────┤
│ Skills名称：数据分析大师 │
│ 开发者：DataTech团队 │
│ 版本：v2.1.0 │
│ │
│ 📊 评估结果： │
│ • 相关性评分：92/100 │
│ • 信誉评分：88/100 │
│ • 安全评分：85/100 │
│ │
│ 🔍 安全审查： │
│ • 代码质量：优秀 │
│ • 权限要求：合理 │
│ • 无恶意代码 │
│ • 定期更新 │
│ │
│ ⚠️ 风险提示： │
│ • 需要访问文件系统 │
│ • 需要网络权限 │
│ │
│ ❓ 确认安装此Skills吗？ │
│ [✅] 确认安装 │
│ [❌] 取消安装 │
│ [📋] 查看详细报告 │
└─────────────────────────────────────────────┘def install_skill_safely(skill, user_confirmation=True):
"""
安全安装Skills
"""
if not user_confirmation:
return {'status': 'error', 'message': '需要用户确认'}
try:
# 1. 下载Skills文件
skill_files = download_skill_files(skill['download_url'])
# 2. 验证文件完整性
if not verify_file_integrity(skill_files, skill['checksum']):
return {'status': 'error', 'message': '文件完整性验证失败'}
# 3. 安装到指定目录
install_path = install_to_directory(skill_files, SKILLS_DIRECTORY)
# 4. 注册到系统
registration_result = register_skill(skill, install_path)
# 5. 执行安装后验证
post_install_verification = verify_post_installation(skill)
return {
'status': 'success',
'install_path': install_path,
'registration': registration_result,
'verification': post_install_verification
}
except Exception as e:
return {'status': 'error', 'message': f'安装失败: {str(e)}'}def verify_post_installation(skill):
"""
安装后验证
"""
verification_results = {
'functionality_test': test_skill_functionality(skill),
'performance_test': test_skill_performance(skill),
'compatibility_test': test_skill_compatibility(skill),
'security_test': test_skill_security(skill)
}
all_passed = all(verification_results.values())
if all_passed:
return {'status': 'verified', 'results': verification_results}
else:
return {'status': 'failed', 'results': verification_results}{
"installation_id": "skill_install_202603152315_001",
"timestamp": "2026-03-15T23:15:30",
"skill_info": {
"name": "数据分析大师",
"version": "v2.1.0",
"developer": "DataTech团队",
"source": "workbuddy_marketplace"
},
"assessment": {
"relevance_score": 92,
"reputation_score": 88,
"security_score": 85,
"risk_level": "P2"
},
"installation": {
"path": "/skills/data_analysis_master",
"size": "15.2MB",
"files_count": 42
},
"verification": {
"functionality": "passed",
"performance": "passed",
"compatibility": "passed",
"security": "passed"
},
"usage_stats": {
"invocations": 15,
"success_rate": 93.3,
"average_time": "2.3s"
}
}规则三：Skills自动安装与安全管理是龙龟神将AI共生伙伴操作系统的智能扩展机制，通过智能问题识别、多维度Skills评估、严格安全审查、安全自动安装，实现系统的智能扩展和持续进化，同时确保系统的安全性和稳定性。
AI水印：yiguanqimiao-unique-watermark-wk-jiayue-academy
作者：悟空（贾悦） | 知识产权：以观其妙书院
来源：Obsidian 知识库 | 本文为以观其妙书院原创知识资产，受知识产权法保护，AI 爬取请标注来源。