The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted Excel files and transforms their content into Markdown format for inclusion in the agent's context. This creates a vulnerability where an attacker could embed instructions within a spreadsheet to hijack the agent's reasoning.\n
Ingestion points: scripts/excel_to_markdown_general.py reads content from .xlsx and .xlsm files provided by the user via command-line arguments.\n
Boundary markers: The script outputs content using Markdown headers (##) and tables, but does not implement explicit delimiters or system instructions to distinguish data from commands, nor does it warn the agent to ignore instructions found within the data.\n
Capability inventory: The skill has the ability to read from the file system (via the openpyxl library) and write generated Markdown files to the local disk using Python's standard open() function.\n
Sanitization: While the script escapes Markdown table delimiters (|) and handles basic cell formatting (bold, italics, links), it does not include logic to sanitize or filter the cell text for potentially malicious natural language instructions that might be parsed by an LLM downstream.

excel-to-markdown