The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains instructions for agents to disregard user attempts to override system rules and to protect internal metadata, such as system prompts and function definitions, from being revealed to the user. It also includes specific guidance to suppress the use of the show_command platform tool in favor of direct text responses. These instructions are documented as defensive guardrails and stability measures to prevent prompt extraction and session state corruption.
[COMMAND_EXECUTION]: The skill guides the user through the use of the Salesforce CLI (sf) for managing the lifecycle of Agentforce agents, including metadata deployment, validation, and automated testing. These commands are executed locally within the user's development environment to interact with their authenticated Salesforce organizations.

agentforce-generate