The Agent Skills Directory

[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection due to its core function of analyzing untrusted customer case data and journal entries (SKILL.md).
Ingestion points: Untrusted data is fetched from the sn_customerservice_case and sys_journal_field tables.
Boundary markers: Absent; the instructions do not define delimiters or specific warnings to ignore instructions that might be embedded in case descriptions.
Capability inventory: Includes powerful tools such as SN-Execute-Background-Script, Bash, and SN-Query-Table as defined in the frontmatter.
Sanitization: No sanitization or validation of the fetched text is performed before it is processed by the agent.
[COMMAND_EXECUTION]: The skill utilizes SN-Execute-Background-Script to execute server-side GlideScript for data aggregation and Bash for REST API interactions. These tools are used appropriately according to the documented procedures for reporting and analysis tasks and do not show signs of abuse.

trending-topics