The Agent Skills Directory

[SAFE]: The skill follows security best practices for its intended use case. Its operations are transparent and match the described functionality.
[DATA_EXFILTRATION]: The skill interacts with the local file system to process documents.
Evidence: scripts/format_word.py reads input data from user-specified paths and writes the formatted document to a target output path using doc.save().
Context: This behavior is the primary function of the skill. No code was found that attempts to send data to external servers or access sensitive system paths.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it processes content from external documents.
Ingestion points: The load_document function in scripts/format_word.py reads content from .docx, .txt, and .md files.
Boundary markers: Not present.
Capability inventory: The skill can write files to the disk but does not have network access or the ability to execute system commands.
Sanitization: The script treats the ingested text as literal strings for font and paragraph styling and does not pass the content to an evaluator or shell.

shmtu-word-formatter