The Agent Skills Directory

[DYNAMIC_EXECUTION]: The script scripts/fix_excel.py uses the zipfile.extractall() method to unpack Excel files. This function does not natively prevent path traversal attacks. A malicious .xlsx file containing members with filenames like ../../etc/passwd or similar relative paths could be used to write or overwrite files outside of the intended temporary directory when the repair script is run.\n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external files and has several exploitable capabilities.\n
Ingestion points: The fix_xlsx_empty_fills function in scripts/fix_excel.py accepts and processes .xlsx files provided via the filepath argument.\n
Boundary markers: Absent. The skill does not utilize delimiters or specific instructions to the agent to disregard instructions that might be embedded within the Excel metadata or content.\n
Capability inventory: The skill utilizes zipfile for file extraction, os.walk for filesystem traversal, shutil.copy2 for creating backups, and shutil.rmtree for directory deletion. Furthermore, SKILL.md demonstrates the use of subprocess.run.\n
Sanitization: Absent. The script fails to validate or sanitize the paths of the zip archive members before extraction, creating a vulnerability surface for path traversal.\n- [COMMAND_EXECUTION]: The SKILL.md file provides usage examples that encourage the agent to use subprocess.run to execute the fix_excel.py script. Although the examples use a list of arguments rather than a shell string (which reduces direct shell injection risk), it facilitates the execution of local system commands on user-provided file paths.

xlsx-fix