The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from external Excel workbooks, which creates a surface where malicious instructions embedded in cells could potentially influence the agent's behavior.
Ingestion points: The skill reads .xlsx workbooks using openpyxl and pandas for analysis and modification.
Boundary markers: No explicit boundary markers or instructions to ignore embedded prompts within spreadsheet data are defined.
Capability inventory: The skill possesses the ability to execute Python scripts (check_office.py), run a Node.js CLI tool (libreofficeKit), and perform file write operations in the workspace.
Sanitization: There is no mention of sanitizing, escaping, or validating the content of spreadsheet cells before they are processed or analyzed.
[COMMAND_EXECUTION]: The skill executes shell commands to perform document integrity checks and rendering tasks.
Evidence: Instructions direct the agent to run a shared validation script: <python> <skill-directory>/../scripts/check_office.py <workbook.xlsx>.
Evidence: The skill utilizes a Node.js-based CLI provided by the environment (libreofficeKit.node and libreofficeKit.cli) to render worksheets and convert files to PDF.
[DYNAMIC_EXECUTION]: The skill invokes executable tools and scripts using paths resolved at runtime.
Evidence: The agent is instructed to use a Python executable and a CLI tool from computed absolute paths to perform recalculation and rendering operations.

office-xlsx