The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external invoice files (PDF/XML) which could contain instructions meant to influence the agent's subsequent actions.
Ingestion points: The scripts/read_invoice.py script reads user-provided files from the filesystem.
Boundary markers: None. There are no explicit instructions or delimiters used to separate extracted invoice content from the agent's operational instructions.
Capability inventory: The skill allows the agent to execute Python scripts via the shell and write JSON output to the local filesystem.
Sanitization: The script extracts data into a structured JSON format but does not perform content-level sanitization on the strings extracted from the documents.
[EXTERNAL_DOWNLOADS]: The skill sends document data to the vendor's API at service.invoice-api.xhub.io. This network communication is a core part of the skill's functionality for processing e-invoices.
[CREDENTIALS_UNSAFE]: The skill script scripts/invoice_api.py reads configuration from a local .env file to retrieve the INVOICE_XHUB_API_KEY. This is a standard practice for managing credentials in development environments and is used to authenticate requests to the vendor's API.

invoice-read