The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes data from external Salesforce API endpoints, creating a potential surface for instructions embedded in source metadata to influence agent behavior.
Ingestion points: Data is fetched via sf api request rest from the agent-templates endpoint and via sf data query from BotDefinition records (Phases 1 and 2).
Boundary markers: The workflow uses explicit AskUserQuestion confirmation gates before performing any write operations and leverages standalone Node.js scripts to isolate parsing logic from the main prompt context.
Capability inventory: The skill is permitted to execute shell commands (sf, node), read/write local files, and prompt the user for input.
Sanitization: The build-create-body.mjs script implements yamlDoubleQuoteEscape to prevent user-supplied strings from breaking the YAML structure of the Agent Script and uses a multi-pass HTML entity decoder to normalize template content.
[COMMAND_EXECUTION]: The skill relies on the Salesforce CLI (sf) for all interactions with the target environment.
The instructions explicitly direct the use of sf api request rest and sf data query, which utilize the CLI's internal session management. This is a secure alternative to manual token handling, as it avoids exposing bearer tokens in shell command histories or environment logs.

service-itsm-agentic-setup-employee-agent-configure