The Agent Skills Directory

[SAFE]: The skill exhibits professional security standards throughout its implementation.
XML Security: It consistently employs the defusedxml library for all XML parsing and manipulation tasks in Python, which protects against XML External Entity (XXE) and quadratic expansion (Billion Laughs) attacks.
Safe Command Execution: System interactions for validation (via soffice) and comparison (via git) use subprocess.run with list-based arguments instead of shell strings, preventing command injection.
Least Privilege and Isolation: Operations are performed on temporary copies of documents within secure system temp directories.
Compliance Validation: The skill includes detailed validation modules (schema checking, ID uniqueness, relationship verification) to ensure that modifications produce valid, non-corrupt documents.

docx