The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted data with high-capability tools. Ingestion points: The wps_word_get_document_text tool reads content from external Word documents, which can be controlled by an attacker. Boundary markers: The instructions do not define delimiters or provide warnings to the agent to treat document content as data rather than instructions. Capability inventory: The agent has access to powerful tools like wps_word_open_document, wps_word_find_replace, and wps_word_insert_text. Sanitization: No validation or sanitization of document content is performed before it is processed.
[COMMAND_EXECUTION]: The skill exposes tools that interact with the host file system. The wps_word_open_document and wps_word_insert_image tools accept file paths as arguments; without strict path validation in the underlying MCP implementation, this could potentially be exploited to access or manipulate sensitive files on the user's system if the agent is tricked via prompt injection.

wps-word