Unsafe Command

January 28, 2026 · View on GitHub

CodeSeverityi18nExperimental
unsafe-commandWarningsast_warnings.unsafe_command✅

Introduction

Detects potentially dangerous use of child process execution functions from the child_process module. Spawning system-level commands can introduce security risks, especially if user-controlled input is involved or if the command itself is sensitive (e.g., tools that query or change system configurations).

Note

This rule is experimental. The list of suspicious commands is not exhaustive and will evolve over time.

Detection Behavior

The probe has different detection modes depending on the sensitivity level:

Conservative Mode (default)

Detects usage of specific unsafe commands:

  • csrutil - System Integrity Protection configuration
  • uname - System information
  • ping - Network connectivity testing
  • curl - HTTP requests and downloads

Aggressive Mode

Flags any usage of child_process execution methods, regardless of the command.

Detected APIs

The probe monitors the following child_process methods:

  • child_process.spawn()
  • child_process.spawnSync()
  • child_process.exec()
  • child_process.execSync()

Examples

import { spawn, exec } from "child_process";

// Detected in conservative mode: Known unsafe commands
spawn("csrutil", ["status"]);
spawn("uname", ["-a"]);
exec("ping google.com");
exec("curl https://malicious.com");

// Detected only in aggressive mode: Any command
spawn("ls", ["-la"]);
exec("echo hello");

References