The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The scripts/extract_map.py script installs the openpyxl and pandas Python packages from the official PyPI registry at runtime if they are not already available in the environment.\n- [COMMAND_EXECUTION]: The SKILL.md instructions include shell commands for the agent to run, such as ls and python scripts, which use placeholders for user-supplied file paths. This pattern relies on the agent's ability to safely handle parameters within shell commands.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it extracts data from untrusted user-uploaded Excel files and presents it to the agent as an authoritative reference for decision-making.\n
Ingestion points: User-uploaded Excel workbooks in /mnt/user-data/uploads/ are parsed by the scripts/extract_map.py script.\n
Boundary markers: Absent. There are no delimiters or instructions that mark the extracted workbook data as untrusted or separate from the agent's primary instructions.\n
Capability inventory: The agent is authorized to execute shell commands and read/write to the file system.\n
Sanitization: Absent. The skill does not implement filtering or sanitization of cell values, column headers, or formulas to prevent malicious instructions from influencing the agent.

excel-mapper