The Agent Skills Directory

[SAFE]: The toolkit consists of standalone Python scripts (find_files.py, count_lines.py, file_stats.py, compare_files.py) that perform local file system operations using standard libraries. No external dependencies or network operations are involved.
[DATA_EXPOSURE]: The scripts read file metadata and content to provide summaries, statistics, and diffs. This behavior is consistent with the intended primary purpose of the skill and does not include mechanisms for exfiltrating data to external services.
[INDIRECT_PROMPT_INJECTION]: The skill processes file data which could potentially contain embedded instructions. However, the scripts do not possess dangerous capabilities (such as network access or shell execution) that could be leveraged by such instructions, and the agent's own safety guardrails remain the primary defense.

file-utils