The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill performs runtime compilation and process injection to facilitate its formula recalculation feature. Specifically, scripts/office/soffice.py writes an embedded C source to a temporary file, compiles it using gcc into a shared library, and then uses the LD_PRELOAD environment variable to inject this shim into the soffice process. This is designed to redirect Unix socket calls to enable communication in restricted environments.
[COMMAND_EXECUTION]: The skill invokes several external binaries via subprocess.run. This includes soffice for spreadsheet processing, gcc for the shim compilation, gtimeout for execution limits, and git for comparing document versions.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted spreadsheet data, which represents a potential attack surface for indirect prompt injection. 1. Ingestion points: Data enters the agent's context through files read via pandas.read_excel and openpyxl.load_workbook. 2. Boundary markers: No explicit delimiters or instructions to ignore instructions within the data are provided to the model. 3. Capability inventory: The skill has the ability to execute shell commands and modify the filesystem. 4. Sanitization: It employs the defusedxml library for document parsing to protect against several XML-specific attack vectors.

xlsx