The Agent Skills Directory

[COMMAND_EXECUTION]: The skill integrates the snow_execute_script_with_output tool, enabling arbitrary server-side JavaScript execution in ServiceNow. This is a high-privilege function intended for development but requires strict input control.\n- [DATA_EXFILTRATION]: Functions like exportToCSV, exportToJSON, and exportToXML allow querying and extracting data from ServiceNow tables. This creates a risk of sensitive data exposure if table names or query parameters are controlled by untrusted sources.\n- [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface (Category 8).\n
Ingestion points: Untrusted data enters via csvContent in the importCSVData function and encodedQuery parameters in various data retrieval functions.\n
Boundary markers: No explicit delimiters or instructions are provided to the agent context to ignore embedded instructions within processed data.\n
Capability inventory: The skill possesses powerful write and execution capabilities, including snow_execute_script_with_output, bulkUpdate, and bulkDelete.\n
Sanitization: While CSV headers are normalized, there is no validation or escaping for row data values or query strings, permitting potential manipulation of system logic.

import-export