The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates the execution of dynamically generated Python code through the create_excel_spreadsheet and modify_excel_spreadsheet tools. This code is generated by the agent based on user instructions and utilizes specific libraries such as openpyxl, pandas, and numpy to interact with spreadsheet data.
[PROMPT_INJECTION]: The skill exposes an indirect prompt injection surface because it reads external content from spreadsheet files that may influence future agent actions or code generation.
Ingestion points: Data from external spreadsheets is ingested via the read_excel_spreadsheet and preview_excel_sheets tools.
Boundary markers: There are no specified delimiters or instructions to the agent to disregard instructions embedded within spreadsheet data.
Capability inventory: The skill allows the agent to execute Python code snippets and perform file system operations (list, read, write) via the provided toolset.
Sanitization: The instructions do not define any sanitization or validation processes for the contents of the spreadsheets before they are processed by the agent.

excel-spreadsheets