The Agent Skills Directory

Persistence Mechanisms (HIGH): The script writes a LibreOffice macro file (Module1.xba) to the user's application configuration directory (e.g., ~/.config/libreoffice/ or ~/Library/Application Support/LibreOffice/). This is a persistent modification of the host environment that remains across sessions.
Indirect Prompt Injection (HIGH): The skill processes untrusted external Excel files using complex system tooling.
Ingestion points: Untrusted Excel file paths provided as arguments to recalc.py.
Boundary markers: None. The skill assumes all content in the provided file is data to be processed.
Capability inventory: High-privilege file modification via soffice (LibreOffice) and file-read capabilities via openpyxl.
Sanitization: None; the script only performs a existence check via Path.exists() before processing.
Risk: Using a full office suite like LibreOffice to process potentially attacker-controlled binary files is a high-risk activity due to the complexity of document parsers and the history of RCE vulnerabilities in such software.
Command Execution (MEDIUM): The script uses subprocess.run to execute system binaries including soffice, timeout, and gtimeout. While it uses list-based arguments to mitigate shell injection, it executes system-level software with access to the filesystem.

xlsx