The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests schema and record data from ServiceNow tables using snow_query_table and snow_discover_table_fields, creating a surface where untrusted data could influence agent behavior.
Ingestion points: Data enters the context via snow_query_table and snow_discover_table_fields as shown in the 'MCP Tool Integration' section of SKILL.md.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the processing of ServiceNow data.
Capability inventory: The skill utilizes snow_execute_script which allows for script execution and record modification within the ServiceNow instance.
Sanitization: There is no evidence of sanitization, filtering, or validation applied to the data retrieved from ServiceNow tables.
[DYNAMIC_EXECUTION]: The skill generates and executes JavaScript snippets (ES5) at runtime using the snow_execute_script tool to perform operations such as table and field creation, as well as complex data policy logic.
[COMMAND_EXECUTION]: The snow_execute_script tool is used to execute arbitrary JavaScript code against the ServiceNow GlideRecord API, representing a powerful execution capability within the target environment.

data-policies