The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes a computed script path containing relative traversal (../skill-shaper/scripts/init_skill.py) to execute a script outside the skill's own directory scope. This relies on an assumed directory structure and can be used to execute unintended files if the environment is manipulated.
[PROMPT_INJECTION]: The skill is vulnerable to indirect injection and shell command injection via the $ARGUMENTS variable.
Ingestion points: Untrusted user input enters the agent context through the $ARGUMENTS placeholder in SKILL.md.
Boundary markers: None; the variable is interpolated directly into the bash shell command without quotes or delimiters.
Capability inventory: The skill uses the Bash tool to execute python3 with the provided arguments.
Sanitization: No escaping, validation, or filtering is performed on the $ARGUMENTS before they are passed to the shell, allowing for the execution of arbitrary commands via characters such as ;, &, or |.
[COMMAND_EXECUTION]: The execution of python3 "$SCRIPT" $ARGUMENTS is inherently unsafe as it allows shell metacharacters in $ARGUMENTS to break out of the intended command context.

init-skill