The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests text from ServiceNow incidents, problems, and changes to construct KB articles, creating a surface where malicious instructions in those source records could potentially influence the agent's behavior during content generation.\n
Ingestion points: Data is fetched from the incident, problem, and change_request tables via SN-Query-Table and REST API endpoints.\n
Boundary markers: The skill does not implement specific delimiters or instructions to treat ingested ticket data as untrusted or to ignore embedded commands.\n
Capability inventory: The skill can create and update records in the kb_knowledge table, create many-to-many relationships in m2m_kb_task, and execute bash commands.\n
Sanitization: There are no explicit instructions for sanitizing or escaping the content from ticket fields like description or close_notes before they are interpolated into the KB article HTML template.\n- [COMMAND_EXECUTION]: The skill documentation utilizes the Bash tool for demonstrating REST API calls to ServiceNow. While functional and restricted to the documented ServiceNow endpoints, standard execution environment security practices should be applied.

article-generation