The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted, user-generated data from ServiceNow HR tables which represents a potential injection surface.
Ingestion points: The skill reads from sn_hr_talent_feedback (content field) and sn_hr_talent_review (comments, strengths_noted, and areas_for_improvement fields) in Steps 3 and 4 of SKILL.md.
Boundary markers: There are no explicit boundary markers or instructions to the agent to treat the retrieved feedback or review comments as untrusted data.
Capability inventory: The skill has access to SN-Execute-Background-Script, SN-Query-Table, and Bash tools.
Sanitization: No sanitization or filtering is performed on the retrieved text content before it is processed by the agent or the background script.
[COMMAND_EXECUTION]: The skill instructs the agent to use the SN-Execute-Background-Script tool to run server-side JavaScript (GlideRecord) on the ServiceNow instance. This is a high-privilege operation used here to aggregate and calculate performance metrics.
[DATA_EXFILTRATION]: The skill is specifically designed to access and expose sensitive Human Resources data, including performance ratings, career aspirations, and risk of loss metrics from tables such as sn_hr_talent_profile, sn_hr_talent_review, and sys_user.

generate-talking-points