The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the dynamic context injection syntax !echo "$1" in SKILL.md to define the target path. This syntax executes the shell command at the moment the skill is loaded. Because the $1 parameter is sourced directly from user input and is not sanitized before being passed to the shell, an attacker can supply a malicious string containing shell metacharacters (such as semicolons, backticks, or subshell syntax) to execute arbitrary code. This represents a significant security risk for command injection and potential data exfiltration.

code-refactor