The Agent Skills Directory

[COMMAND_EXECUTION]: The skill generates and suggests shell commands (e.g., pytest [selected_tests]) where [selected_tests] is a list of file paths retrieved via git diff and import scanning. If the repository contains files with shell metacharacters in their names (e.g., ; rm -rf / ;.py), the generated command would execute unintended operations if run by the agent or user without further sanitization.
[DATA_EXFILTRATION]: The skill accesses and manages data files within the .claude/data/test-mapping/ and ~/.amplihack/ directories. While primarily used for caching test mappings and reliability scores, accessing the home directory is a sensitive operation that can lead to data exposure.

smart-test