The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill presents a vulnerability surface for indirect prompt injection because it interacts with external data sources that could be controlled by an attacker.
Ingestion points: The skill retrieves data from ServiceNow tables using SN-Query-Table and SN-Get-Record (Phase 7.4), which serves as an entry point for untrusted content into the agent context.
Boundary markers: No explicit delimiting instructions or boundary markers are provided to ensure the agent ignores potentially malicious instructions embedded in the retrieved table data.
Capability inventory: The agent has access to impactful tools including SN-Create-Record, SN-Update-Record, and local Bash for development workflows, creating a path where injected data could influence local or platform-level execution.
Sanitization: The skill lacks documentation or instructions for the validation, escaping, or sanitization of content ingested from external ServiceNow records before processing.
[DYNAMIC_EXECUTION]: The skill facilitates the generation and deployment of executable JavaScript code to a remote platform.
Evidence: The skill provides predefined templates for onLoad, onChange, and onSubmit scripts that are programmatically created in the sys_script_client table using the SN-Create-Record tool (Phase 1). This is consistent with standard development practices but involves script generation from templates.

client-scripts