The Agent Skills Directory

[SAFE]: No security issues were detected during the analysis of the skill instructions and metadata.
Functional Scope: The skill is restricted to analyzing code patterns (duplicates, long functions, magic values) and proposing cleanup ideas.
No Execution: The instructions explicitly state the skill does not execute changes and only provides suggestions.
Tool Usage: The use of Bash, Read, Grep, and Glob is aligned with the stated purpose of searching and reading project files for analysis.
Data Handling: No network operations, credential access, or data exfiltration patterns were identified.
Indirect Prompt Injection Surface: While the skill reads untrusted local code files (an inherent property of code analysis tools), the risk is mitigated by the 'suggest-only' constraint and lack of automated execution of the processed content.

refactor-safe