技能 skill-creator-ms
📦

skill-creator-ms

低风险 ⚙️ 外部命令🌐 网络访问📁 文件系统访问🔑 环境变量

为 AI 代理创建 Azure SDK 技能

该技能指导 AI 代理为 Azure SDK 和 Microsoft Foundry 服务创建模块化知识包,确保一致的模式和正确的身份验证。

支持: Claude Codex Code(CC)
📊 70 充足
1

下载技能 ZIP

2

在 Claude 中上传

前往 设置 → 功能 → 技能 → 上传技能

3

开启并开始使用

测试它

正在使用“skill-creator-ms”。 Create a skill for Azure Key Vault secrets

预期结果:

A complete skill structure with SKILL.md containing: installation commands (pip install azure-keyvault-secrets), environment variables setup, DefaultAzureCredential authentication, core workflow examples for secret operations, best practices, and reference links.

正在使用“skill-creator-ms”。 What categories are available for organizing Azure SDK skills?

预期结果:

Foundry (AI Foundry, agents), Data (Storage, Cosmos DB), Messaging (Event Hubs, Service Bus), Monitoring (OpenTelemetry), Identity (authentication), Security (Key Vault), Integration (API Management), Compute (Batch), Container (Container Registry).

正在使用“skill-creator-ms”。 Create acceptance criteria for an Azure SDK skill

预期结果:

Acceptance criteria document with correct/incorrect import patterns, authentication patterns (DefaultAzureCredential vs hardcoded), client initialization, async variants, and common anti-patterns to avoid.

安全审计

低风险
v1 • 2/25/2026

Static scanner flagged 219 potential issues (external commands, network, filesystem, env access) but ALL are false positives. The skill is a documentation/guide containing markdown code examples for creating Azure SDK skills. No executable malicious code exists. The skill explicitly warns against hardcoding credentials as a security anti-pattern.

1
已扫描文件
619
分析行数
8
发现项
1
审计总数
低风险问题 (4)
External Command Patterns in Code Examples
Scanner detected 168 'external command' patterns at various lines. These are bash command examples in markdown code blocks (pip install, ln -s, pnpm commands) used to document the skill creation process. Not actual executable code.
Hardcoded URLs in Documentation
Three URLs detected: example Azure endpoint, Microsoft Learn docs URL, and GitHub repo URL. These are legitimate documentation URLs and example placeholders.
Filesystem References in Skill Structure
Path traversal and symlink patterns detected. These describe the skill's own directory structure (.github/skills/, skills/<language>/<category>/) used for organizing Azure SDK skills.
Environment Variable Examples
Environment variable access patterns (AZURE_ENDPOINT) in code examples. The skill explicitly warns against hardcoding credentials as a security anti-pattern.

风险因素

⚙️ 外部命令 (2)
🌐 网络访问 (3)
📁 文件系统访问 (2)
🔑 环境变量 (2)
审计者: claude

质量评分

38
架构
100
可维护性
87
内容
50
社区
82
安全
87
规范符合性

你能构建什么

新建 Azure SDK 技能

当开发人员需要为新 Azure 服务 SDK 创建技能时

更新现有技能

使用当前的 Azure SDK 模式和身份验证更新现有技能

技能测试和验证

创建验收标准和测试场景以验证技能质量

试试这些提示

创建 Azure AI 代理技能
Use the skill-creator-ms to create a new skill for the Azure AI Agents SDK. The SDK package is azure-ai-agents and the documentation is at https://learn.microsoft.com/en-us/azure/ai-services/agents/.
创建 Cosmos DB 技能
Create a skill for Azure Cosmos DB using the azure-cosmos SDK. Documentation: https://learn.microsoft.com/en-us/azure/cosmos-db/
添加身份验证部分
Add proper authentication documentation to the existing azure-storage skill using DefaultAzureCredential.
创建测试场景
Create acceptance criteria and test scenarios for the azure-eventhub skill following the skill-creator-ms patterns.

最佳实践

  • Always use DefaultAzureCredential for authentication instead of hardcoding API keys
  • Keep SKILL.md under 500 lines and use progressive disclosure with reference files
  • Include both correct and incorrect code patterns in acceptance criteria to help validate outputs

避免

  • Creating a skill without first obtaining the SDK package name and documentation URL from the user
  • Hardcoding credentials in example code - always use environment variables
  • Skipping the authentication section - agents will improvise poorly without guidance

常见问题

在创建 Azure SDK 技能之前我需要提供什么信息?
You must provide the SDK package name (e.g., azure-ai-agents), the Microsoft Learn documentation URL, and optionally the GitHub repository for code patterns.
如何按类别组织技能?
Create a symlink in skills/<language>/<category>/ pointing to .github/skills/<skill-name>. Categories include foundry, data, messaging, monitoring, identity, security, integration, compute, and container.
应该使用什么身份验证方法?
Always use DefaultAzureCredential for Azure SDKs. Never hardcode credentials. Use environment variables like AZURE_ENDPOINT.
SKILL.md 应该多长?
Keep SKILL.md under 500 lines. Use progressive disclosure - load metadata always, body when skill triggers, and reference files as needed.
需要为技能创建测试吗?
Yes, every skill must have acceptance criteria in references/acceptance-criteria.md and test scenarios in tests/scenarios/<skill-name>/scenarios.yaml.
支持哪些编程语言?
The skill covers Python, .NET (C#), Java, and TypeScript patterns for Azure SDKs.

开发者详情

文件结构

📄 SKILL.md