The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the SN-Execute-Background-Script tool to run server-side JavaScript on the ServiceNow instance. In the provided procedure, this is used to calculate performance and success rate analytics by aggregating data from the execution log table.
[COMMAND_EXECUTION]: The skill uses Bash commands to interact with ServiceNow REST APIs via curl for creating and managing process and bot definitions.
[CREDENTIALS_UNSAFE]: The skill interacts with the sn_rpa_credential table to manage bot authentication. The instructions promote security best practices by recommending the use of a credential vault and automatic rotation policies rather than hardcoding passwords within bot definitions.
[PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection as it transforms untrusted user-provided process descriptions into automated bot actions.
Ingestion points: Natural language process descriptions provided by the user (Step 1) are used to define the sequential logic and data entry fields of the RPA bot.
Boundary markers: The instructions do not define explicit boundary markers or warnings for the agent to ignore embedded instructions within the user input.
Capability inventory: The skill can create and update records across multiple ServiceNow tables (sn_rpa_bot, sn_rpa_action, etc.), allowing it to define complex workflows that interact with target applications like SAP GUI.
Sanitization: There is no explicit validation or escaping of the user-provided descriptions before they are interpolated into the RPA action records.

rpa-bot-generation