The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill accepts external input via the $ARGUMENTS placeholder to receive code and requirements for refactoring. This represents a potential surface for indirect prompt injection where malicious instructions could be embedded in the data provided for analysis. However, this is inherent to the skill's primary purpose as a refactoring tool, and no specific automated command execution or data exfiltration capabilities were identified that could be exploited through this surface.
[SAFE]: The skill does not contain any executable scripts, obfuscated content, or unauthorized network operations. All external references are to legitimate GitHub repositories used for attribution and credit.

refactor-clean