The Agent Skills Directory

[PROMPT_INJECTION]: The skill's primary function involves reading and analyzing external source code which constitutes untrusted data. Malicious instructions embedded within comments, strings, or metadata of the target code could attempt to influence the agent's behavior during the refactoring process.
Ingestion points: Reads the file path provided in the $ARGUMENTS variable.
Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands within the code being refactored.
Capability inventory: The skill utilizes Bash and Edit tools, providing a high-privilege environment for potential exploitation.
Sanitization: No sanitization or validation of the ingested code content is performed before processing.
[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands (e.g., git checkout -b "refactor/$ARGUMENTS") that directly interpolate user-supplied arguments. This pattern presents a risk of command injection if the underlying platform does not sufficiently sanitize the $ARGUMENTS variable before shell execution.

refactor-code