The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data in Step 1 (CSV, Excel, bank statements, or free-text lists). While the instructions focus on parsing structured records, there is a risk that malicious data within these files could attempt to influence the agent's behavior.
Ingestion points: Step 1 instructions (Gather expense data) explicitly tell the agent to read file paths provided by the user or parse provided text into structured records.
Boundary markers: The skill lacks explicit instructions for the agent to treat data content as purely passive or to ignore embedded instructions found within the processed files.
Capability inventory: The skill has access to tools such as Bash(python:*), Read, Edit, and Write, allowing it to read and write files and execute Python scripts for data processing.
Sanitization: There are no instructions provided for the sanitization or escaping of the ingested data before it is processed or used in report generation.

israeli-expense-categorizer