The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from user-supplied Excel workbooks, which can contain untrusted content designed to influence agent behavior.
Ingestion points: Spreadsheet files processed via the xlsx command-line tool.
Boundary markers: The instructions do not define specific delimiters or warnings to help the agent distinguish between data and potentially malicious instructions embedded within cell values or formulas.
Capability inventory: The skill utilizes a shell environment (bash) with capabilities for file parsing (xlsx), data processing (jq), and file system interaction (reading from /dropbox and /googledrive, writing to /tmp).
Sanitization: There are no guidelines for the agent to sanitize or validate spreadsheet content before processing the information.
[EXTERNAL_DOWNLOADS]: The skill directs users to install external plugins and connect to a remote server.
Instructions include adding the un-seen/hedwigai-skills plugin and connecting the https://mcp.hedwigai.com/ remote MCP server. These external resources are provided by the skill's author to enable the described functionality.
[COMMAND_EXECUTION]: The skill relies on executing shell commands to perform its core functions.
It instructs the agent to use the xlsx program for reading and editing workbooks, and jq for handling JSON output.
Operational workflows involve command-line file manipulations, such as copying workbooks to the /tmp directory for editing purposes.

excel-workbooks