The Agent Skills Directory

[SAFE]: The skill provides instructional prompts for code refactoring and does not contain any executable scripts, binary files, or network-enabled commands. The instructions align with industry standard best practices for software engineering.
[NO_CODE]: The skill consists entirely of Markdown instructions and metadata. There are no accompanying source code files, scripts, or package dependency files (e.g., package.json or requirements.txt) that could introduce execution-time risks.
[PROMPT_INJECTION]: The skill is designed to ingest and process user-provided code and requirements via the $ARGUMENTS placeholder. It represents a surface for indirect prompt injection as it lacks explicit boundary markers to separate user data from system instructions, although this is inherent to the skill's primary purpose of refactoring code.
Ingestion points: SKILL.md (via $ARGUMENTS and user-supplied code snippets).
Boundary markers: Absent.
Capability inventory: The agent is instructed to refactor code and update tests, which implies file-write and test-execution capabilities within the host environment.
Sanitization: Absent.

code-refactoring-refactor-clean