What problem does it solve? Building and maintaining Agentforce agents through the Setup UI and Agent Builder involves scattered metadata types, unclear deployment ordering, and confusing terminology between Prompt Builder templates and GenAiPromptTemplate metadata, leading to failed publishes and invisible agents. ## Core Features & Use Cases - Builder Metadata Guidance: Create and validate GenAiFunction, GenAiPlugin, and GenAiPromptTemplate metadata with correct folder structure, input contracts, and the 5-input flex template limit. - Agent Lifecycle Operations: Provision Service Agent running users with sf org create agent-user, configure Employee Agent visibility via Permission Sets with agentAccesses, and handle the separate publish and activate steps. - Apex AI Integration: Implement Models API (aiplatform.ModelsAPI) patterns with Queueable and Batch Apex, plus custom LightningTypeBundle UIs for structured agent action inputs and outputs. - Use Case: When registering an Apex discount calculator as an agent action, this Skill walks you through creating the GenAiFunction metadata, matching input/output contracts, deploying supporting metadata first, and activating the agent deterministically. ## Quick Start Ask the agent to set up a GenAiFunction for your Apex or Flow action and wire it into an Agentforce Builder agent.