What problem does it solve? Building Frontline assistant flows by hand is error-prone: invalid node types, missing required fields, bad references, and broken edges all return 400 errors or fail silently. This Skill encodes the exact rules, payloads, and validation steps needed to construct working flow graphs through the Frontline CLI. ## Core Features & Use Cases - Flow graph construction: Create flows, add nodes (START, TRIGGER_INTENT, SAY_AI, RESPONSE_AI, TOOLS_AI, API, CONDITIONAL_ROUTING, CREATE_RECORD_ACTIVITY), and connect them with edges using the correct handles. - Variable and intent management: Create agent variables and intents, capture values into variables from AI or API nodes, and wire intent-based or agentic routing triggers. - Validation guidance: Avoid 400 errors by pre-checking referenced resources (ai-models, custom tools, intents, record types) and respecting rules like required captureVariables, exitNodeWhen, and API node success/fail handles. - Use Case: Build a customer support flow that greets the user, captures their email into a Person record, calls an external API to look up an order, and routes the conversation based on the result. ## Quick Start Ask the assistant to create a new Frontline flow for your agent with a greeting node connected to an intent-triggered response, then inspect the graph to verify the structure.
