The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data by parsing user-provided PayPal CSV exports, which could potentially contain malicious instructions embedded within text fields like 'Name', 'Subject', or 'Item Title'.
Ingestion points: The workflow in SKILL.md requires the user to provide a PayPal CSV file path for parsing.
Boundary markers: Absent; there are no instructions for the agent to use delimiters or ignore instructions that might be embedded within the CSV data fields.
Capability inventory: The skill uses transaction_search, categorize, and export_transactions tools. No high-risk capabilities such as arbitrary shell execution, sensitive file system access, or outbound network operations are present.
Sanitization: Absent; the skill does not specify any validation, filtering, or escaping of the imported data before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The documentation provides instructions for users to manually download their activity statements from the official PayPal domain (paypal.com). This reference to a well-known service is informative and does not involve automated remote code execution.

paypal-import