The Agent Skills Directory

[SAFE]: The skill consists of legitimate administrative procedures for managing ServiceNow AI Search configurations, including search sources, profiles, and indexing.
[PROMPT_INJECTION]: The skill outlines the creation of a RAG pipeline that processes data from knowledge bases and incidents. The instructions include configuring a system prompt designed to ground the AI's responses to the provided context, which is a standard safety measure for RAG systems. * Ingestion points: Content is ingested from ServiceNow tables like kb_knowledge, sc_cat_item, and incident as described in SKILL.md. * Boundary markers: The generative AI configuration in Step 5 includes a system prompt instructing the model to use only the provided context. * Capability inventory: The skill uses SN-Create-Record and SN-Update-Record tools to modify the platform configuration. * Sanitization: The procedure relies on standard ServiceNow platform security and does not specify custom sanitization for retrieved content.

ai-search-rag