# Command Injection (CWE-78)

User input is passed unsanitized to system shell commands, allowing attackers to execute arbitrary commands on the server.

**Stack:** Go

- Prevalence: Common Found in many applications
- Impact: Critical Full server compromise
- Prevention: Avoid shell Use execFile, not exec

**OWASP:** Injection (A03:2021-Injection) - #3

## Description

This could allow attackers to execute unexpected, dangerous commands directly on the operating system. This weakness can lead to a vulnerability in environments in which the attacker does not have direct access to the operating system.

## Prevention

基于 1 条 Shoulder 检测规则的 OS Command Injection 预防策略。

### Go

Use exec.Command with explicit arguments, never shell invocation

## Warning Signs

- [CRITICAL] user input flowing to os/exec command execution, enabling OS command injection

## Consequences

- 执行未授权命令
- 读取应用程序数据
- 绕过保护机制

## Mitigations

- 使用库调用而不是外部进程
- 如果使用 Runtime.exec(),请使用接受参数数组的版本
- 使用能自动强制实现数据与代码分离的结构化机制

## Detection

- Total rules: 3
- Critical: 3
- Languages: go, javascript, typescript, python

## Rules by Language

### Go (1 rules)

- **Command Injection via os/exec** [CRITICAL]: Detects user input flowing to os/exec command execution, enabling OS command injection.
  - Remediation: Use exec.Command with explicit arguments and validate input against an allowlist.

```go
allowed := map[string]bool{"file1.txt": true, "file2.txt": true}
if !allowed[userInput] {
    return errors.New("not allowed")
}
cmd := exec.Command("cat", userInput)
```

Learn more: https://shoulder.dev/learn/go/cwe-78/command-injection