The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes a local Python script toshl/scripts/monthly_sync.py using the Bash tool to perform bulk data synchronization from the Toshl API. This script handles authentication via a token stored in a local .env file and manages data caching in CSV format.
[INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes external financial data from the Toshl Finance API, which is then used to generate markdown reports. This ingestion of untrusted external content represents a surface for indirect prompt injection.
Ingestion points: Transaction descriptions, category names, and tags are fetched from the external Toshl API and stored in local files such as toshl/data/{YYYY-MM}_entries.csv.
Boundary markers: The instructions for generating reports from the CSV data do not specify the use of delimiters or specific instructions to ignore malicious content embedded within the transaction data.
Capability inventory: The skill utilizes Bash for executing the sync script, Read for accessing cached data, and Write for creating the markdown reports.
Sanitization: No sanitization, escaping, or validation mechanisms are described for the external transaction data before it is processed by the language model for report generation.

toshl