The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources and uses it in operations with real-world fiscal consequences.
Ingestion points: Data is retrieved from client and product databases via clientes_buscar and productos_buscar (SKILL.md).
Boundary markers: There are no specific instructions or delimiters provided to isolate untrusted data (like item descriptions or client names) from the agent's internal logic.
Capability inventory: The skill has the ability to emit legally binding documents through dte_emitir and modify records via cliente_crear (SKILL.md).
Sanitization: No explicit sanitization or validation of input string content is specified before these fields are used in document generation.

openfactura-facturar