The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from ServiceNow tables that could contain instructions influencing agent behavior.
Ingestion points: The snow_query_table tool is used to read data from tables like sn_ih_spoke and sn_ih_action as documented in SKILL.md.
Boundary markers: The instructions and code templates do not specify the use of delimiters or boundary markers to isolate untrusted data retrieved from the instance.
Capability inventory: The skill has access to powerful capabilities including server-side script execution (snow_execute_script), artifact management (snow_artifact_manage), and network connectivity testing (snow_test_rest_connection).
Sanitization: There is no evidence of input validation or sanitization for data retrieved from external ServiceNow tables before it is used by the agent.

integration-hub