The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands like find, grep, wc, and awk to perform static analysis on the local filesystem. While these are primarily read-only tools, the use of the $TARGET variable and $ARGUMENTS without explicit sanitization within the skill's instructions could potentially allow for command injection if the execution environment does not properly handle shell metacharacters.
[INDIRECT_PROMPT_INJECTION]: The skill processes external source code files which could contain malicious instructions designed to influence the agent's behavior during the refactoring analysis.
Ingestion points: Content of source code files (.ts, .js, .py, .rs) accessed via find and grep commands in SKILL.md.
Boundary markers: Absent; the instructions do not define delimiters or provide specific directives to ignore instructions embedded within the code being analyzed.
Capability inventory: Local filesystem read and search capabilities via find, grep, and wc as seen in SKILL.md.
Sanitization: Absent; file contents and file paths are processed directly without validation or escaping.

refactor