API Reference

POST /intercept/command

POST/api/v1/intercept/command

Evaluates a shell command proposed by an AI agent. Detects destructive file operations, privilege escalation, raw disk writes, and unsafe pipeline execution.

Request

{
  "command": string   // Required. The shell command string to evaluate.
}

Detected patterns

—Destructive File Deletion (rm -rf)

—Privilege Escalation (sudo, chmod 777)

—Raw Disk Write (dd if=/dev/zero)

—Insecure Permission Grant (chmod a+rwx)

—Unsafe Pipeline Execution (curl | sh)