The Agent Skills Directory

[PROMPT_INJECTION]: The skill retrieves data from ServiceNow GRC tables which could contain malicious instructions designed to alter the agent's behavior (indirect prompt injection).
Ingestion points: Data fetched via snow_query_table from tables such as sn_compliance_policy and sn_risk_risk.
Boundary markers: The skill does not use specific delimiters or instructions to ignore potential commands within the fetched data.
Capability inventory: The skill has the ability to execute server-side scripts using snow_execute_script_with_output.
Sanitization: There is no evidence of data sanitization or validation performed on the records retrieved from ServiceNow.
[COMMAND_EXECUTION]: The snow_execute_script_with_output tool allows the execution of arbitrary server-side JavaScript code. While this is necessary for the skill's intended purpose of ServiceNow development and automation, it represents a powerful capability that should be monitored for misuse.

grc-compliance