Code Review Skill for OpenHarmony Distributed System Security

Overview

本技能提供OpenHarmony分布式业务安全代码检视的专业指导，包含18条安全设计规则和对应的检视要点。当检视分布式系统代码安全性时，在通用网络安全规则基础上，使用这些规则进行加强检视。

Trigger Phrases

• "检视代码安全实现"
• "代码安全审查"
• "安全代码review"
• "检查这段代码的安全性"
• "review分布式代码安全"
• "OpenHarmony安全检视"

Code Review Workflow

Step 1: Understand the Code Context

首先理解代码的业务场景和所在模块：

1. 识别关键模块: 确定代码是否涉及以下模块

- 分布式设备管理 - 分布式软总线 - 其他需要分布式能力的模块

1. 识别业务类型: 判断是否涉及以下安全敏感业务

- 设备间认证和授权 - 用户敏感数据传输 - 跨设备状态机管理 - 可信关系管理 - 硬件资源访问

1. 确定角色: 识别代码是主体侧(客户端)还是客体侧(服务端)

Step 2: Load Security Rules

根据代码涉及的业务类型，加载[securityrules.md](references/securityrules.md)中对应的规则：

快速索引关键词:

• 跨设备传输 → Rules 3, 8, 15, 17
• 状态机 → Rule 2
• 授权/鉴权 → Rules 1, 5, 6, 8, 12
• PIN码/秘钥 → Rules 3, 8, 9, 10
• 资源申请 → Rule 4
• 权限配置 → Rules 7, 13
• 开关标记 → Rule 14
• 用户切换 → Rule 16
• 兼容代码 → Rule 18

Step 3: Review Against Security Rules

对每个适用的安全规则，执行以下检视：

1. 定位相关代码: 使用Grep搜索关键模式

`` Grep patterns examples: - "auth", "authorize", "permission" for authorization checks - "PIN", "secret", "key" for sensitive data - "state", "status" for state machine - "random", "generate" for secret generation ``

1. 检查实现细节:

- 对照规则中的Check points逐项检查 - 查找潜在的违规模式 - 识别缺失的安全措施

1. 记录发现:

- 标记违规代码位置 (file:line) - 说明违反的具体规则 - 提供修复建议

Step 4: Apply General Security Best Practices

除了OpenHarmony特定规则外，还需检查通用安全实践：

1. 输入验证: 所有外部输入是否经过验证
2. 错误处理: 敏感操作是否有适当的错误处理
3. 日志安全: 是否记录了敏感信息
4. 资源管理: 是否有资源泄漏风险

Step 5: Generate Review Report

生成结构化的安全检视报告，包含：

1. 执行摘要: 发现的严重安全问题数量和等级
2. 违规清单: 按严重程度排序的违规项
3. 规则映射: 每个问题对应的安全规则
4. 修复建议: 具体的代码修改建议

Common Violation Patterns

Pattern 1: Client-controlled Authorization (违反规则1)

Bad Example:

// 客体侧直接使用主体侧传入的标志控制弹框
void handleAuthRequest(bool showPopup) {
    if (!showPopup) {
        // 直接跳过授权弹框
        grantAccess();
    }
}

Correct Approach:

// 客体侧独立决策是否需要授权
void handleAuthRequest() {
    if (isSystemBusinessAndRegistered()) {
        // 已注册的免授权业务
        grantAccess();
    } else {
        // 默认必须弹框
        showAuthorizationDialog();
    }
}

Pattern 2: Plaintext Sensitive Data (违反规则3)

Bad Example:

// 明文传输PIN码
message.pin_code = userPin;
sendToRemote(message);

Correct Approach:

// 加密后传输
encryptedPin = encryptPin(userPin, sessionKey);
message.encrypted_pin = encryptedPin;
sendToRemote(message);

Pattern 3: Custom Trust Verification (违反规则8)

Bad Example:

// 自行比对账号信息判断可信关系
bool isTrusted() {
    return localAccount == remoteAccount;
}

Correct Approach:

// 依赖HiChain查询
bool isTrusted() {
    CredentialType type = HiChain.queryCredentialType(remoteDevice);
    return type == CredentialType.SAME_ACCOUNT;
}

Pattern 4: Insecure Switch Defaults (违反规则14)

Bad Example:

// 默认值放通
bool enableSecurityCheck = true;  // 默认启用

Correct Approach:

// 默认值禁用
bool enableSecurityCheck = false;  // 默认禁用，需显式启用

Security Rule Categories

1. Authorization & Authentication

• Rule 1: Object-side Authorization Control
• Rule 5: Anti-Brute Force Protection
• Rule 6: Server-side Security Logic
• Rule 12: Sensitive Data Authorization and Audit

2. Data Protection

• Rule 3: No Plaintext Sensitive Data Transmission
• Rule 10: Secure Random Secrets
• Rule 17: Business-level Key Isolation

3. Trust Management

• Rule 7: Trusted Relationship Lifecycle Minimization
• Rule 8: Trusted Relationship Verification
• Rule 9: Trusted Relationship Persistence Timing
• Rule 15: Device Legitimacy Verification
• Rule 16: User Isolation for Distributed Trust

4. State Machine & Process Control

• Rule 2: State Machine Context Validation

5. Resource Management

• Rule 4: Resource Access Parameter Validation
• Rule 11: Resource Cleanup
• Rule 13: Minimal Permission Configuration

6. Code Quality

• Rule 14: Secure Switch Default Values
• Rule 18: Legacy Protocol Cleanup

Example Review Session

User request: "检视这段分布式设备管理代码的安全性"

Review process:

1. Load security rules → Read [securityrules.md](references/securityrules.md)
2. Identify relevant rules → Rules 1, 2, 7, 8, 9, 11 (设备管理相关)
3. Search code patterns → Grep for authorization, trust, state machine
4. Check each rule:

- ✓ Rule 1: 授权流程是否在客体侧独立控制 - ✗ Rule 2: 发现状态机未校验上下文 - ✓ Rule 7: 可信关系生命周期管理正确 - ✗ Rule 8: 发现自定义可信判断逻辑

1. Generate report → 列出违规点和修复建议

Tips

1. Start with keywords: 使用security_rules.md中的关键词快速定位可疑代码
2. Check both sides: 分布式业务需要同时检查主体侧和客体侧代码
3. Verify complete flows: 跟踪完整的业务流程，不要只检查单个函数
4. Consider edge cases: 检查错误处理、超时、重试等边界场景
5. Review logging: 确保日志中不泄露敏感信息

Resources

• Detailed Rules: See [securityrules.md](references/securityrules.md) for complete rule descriptions and check points
• Quick Reference: Use keyword mapping at the end of security_rules.md for fast rule lookup