office-academic-skill
Pass
Audited by Gen Agent Trust Hub on May 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill exclusively uses the
defusedxmllibrary for all XML parsing and manipulation tasks in Python, effectively preventing XML External Entity (XXE) and Billion Laughs vulnerabilities. - [COMMAND_EXECUTION]: Local shell commands are invoked via
subprocess.runto leverage existing document tools including LibreOffice (soffice) for validation,pdftoppmfor thumbnail generation, andgitfor diffing tracked changes. These commands are executed without a shell (shell=False) and use validated internal paths. - [SAFE]: JavaScript logic in
html2pptx.jsuses Playwright to process local HTML files for slide layout generation, restricted to the local environment via thefile://protocol. - [SAFE]: No prompt injection attempts, hardcoded credentials, or data exfiltration patterns were detected in the skill's instructions or supporting scripts.
Audit Metadata