writer

Pass

Audited by Gen Agent Trust Hub on Aug 10, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill employs subprocess.run to interact with system utilities such as unzip, zip, soffice, and killall. These commands are used to manage document file structures and control the LibreOffice process, which are standard for the described document automation workflows.
  • [DATA_EXFILTRATION]: The skill uses a local socket (localhost:8100) for the UNO bridge to automate LibreOffice. No sensitive data access or network transmission to external domains was detected.
  • [EXTERNAL_DOWNLOADS]: The skill suggests using odfpy and ezodf, which are standard, well-known Python packages for handling OpenDocument files, typically sourced from the official PyPI registry.
  • [PROMPT_INJECTION]: The skill possesses an ingestion surface through document files and template variables (Ingestion points) and uses them without explicit boundary markers or sanitization (Sanitization/Boundaries absent). While it has the capability to write files and run subprocesses (Capability inventory), its use of simple string replacement in templates is a common practice for document generation and no malicious patterns were identified.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 10, 2026, 07:59 AM