base
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes instructions for running the LibreOffice binary (
soffice) in various modes, including headless mode with socket listening enabled for remote automation. It also provides a troubleshooting command to terminate the application process (killall soffice.bin).- [INDIRECT_PROMPT_INJECTION]: The skill features capabilities to import and process data from external sources such as CSV files, spreadsheets, and SQL databases, which represents an ingestion surface for potentially malicious instructions. - Ingestion points: Data import functions for CSV/spreadsheets and external database connections via JDBC/ODBC drivers.
- Boundary markers: No specific boundary markers or delimiters for untrusted input are defined in the provided templates.
- Capability inventory: Includes file system write operations (
doc.storeToURL), network connectivity for database access, and local process execution capabilities. - Sanitization: The documentation recommends using parameterized queries as a best practice to mitigate standard injection attacks.
Audit Metadata