The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill documents methods for ingesting untrusted data from external spreadsheet files, which represents an attack surface for indirect prompt injection.
Ingestion points: SKILL.md examples demonstrate reading from files via load_workbook('data.xlsx'), pd.read_excel('data.xlsx'), and XLSX.readFile('data.xlsx').
Boundary markers: Absent; the examples do not include instructions for the agent to ignore potentially malicious content within the data cells.
Capability inventory: Demonstrates file read/write operations using standard libraries. No arbitrary command execution or network exfiltration patterns are present in the provided snippets.
Sanitization: Absent; the code snippets represent basic functionality without data validation or sanitization layers.

xlsx