The Agent Skills Directory

[COMMAND_EXECUTION]: The skill requests access to the Bash shell and utilizes the SN-Execute-Background-Script tool to run server-side GlideScript on ServiceNow instances for validation and deployment tasks.
[DYNAMIC_EXECUTION]: Scripts are dynamically generated and executed on the ServiceNow platform (e.g., in Phase 4.3 and 7.2) to automate conflict resolution and rollback analysis.
[INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted configuration data from ServiceNow update set tables (sys_update_xml), which could serve as a vector for indirect instructions.
Ingestion points: Metadata and payloads from the sys_update_xml, sys_remote_update_set, and sys_update_preview_problem tables (SKILL.md, Step 1.2, Step 4.2).
Boundary markers: None provided in the instruction templates.
Capability inventory: The skill possesses high-privilege capabilities including SN-Execute-Background-Script, SN-Update-Record, and local Bash access.
Sanitization: Validation logic exists in Step 1.2 to check for hardcoded identifiers, but there is no specific filtering for prompt injection patterns in the ingested data.

deployment-workflow