agent-builder

Create, configure, and deploy Frontline agents from the command line.

Updated Jul 27, 2026
One-click install
npx skills add https://github.com/api-evangelist/frontline --skill agent-builder-api-evangelist
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-builder
Source: https://github.com/api-evangelist/frontline/tree/main/skills/vendor/agent-builder
Command: npx skills add https://github.com/api-evangelist/frontline --skill agent-builder-api-evangelist

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Managing Frontline agents through a web UI is slow for repetitive setup tasks; this Skill lets you create, configure, deploy, and test agents entirely from the Frontline CLI with scriptable commands. ## Core Features & Use Cases - Agent Lifecycle Management: Create, list, update, soft-delete, and select an active agent context, with --agent-id overrides for any command. - Deployment & Testing: Publish or pause agents with billing checks, then smoke-test flows using terminal chat sessions. - Configuration Control: Update agent settings (instructions, model, temperature, tools), chat widget themes, and per-channel settings for livechat, WhatsApp, Instagram, and Messenger. - Use Case: A developer spins up a new support agent, sets its instructions and AI model, configures the WhatsApp channel payload, publishes it, and validates the flow with a test chat — all without leaving the terminal. ## Quick Start Authenticate with a USER API key using frontline auth login, then create and deploy a new agent named Support Agent.

Frequently Asked Questions about agent-builder

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create and deploy a Frontline agent from the CLI?▼

Run frontline agents create --name "Support Agent", which saves it as the active agent, then publish it with frontline agents deploy --offline false. Publishing runs the same billing checks as the internal assistant manager.

How do I update Frontline agent settings like instructions and model?▼

Use frontline agents agent-setting update --data with a JSON payload containing fields like instructions, temperature, and aiModelId. The CLI merges your data with the current setting before sending, so you only send the fields you want to change.

What API key type does the Frontline CLI require for agent mutations?▼

Mutations require a USER API key because ownership, billing, and account checks use the API key user context. Authenticate first with frontline auth login <api-key>.

Which channels does Frontline agent configuration support?▼

Supported channel names are livechat, whatsapp, instagram, and messenger. Each channel has its own validated payload schema, so start from settings get <channel>, edit the JSON, and send it back with settings update.

What are the temperature and maxIterations limits for Frontline agent settings?▼

Temperature must be between 0 and 1 using one decimal place, and maxIterations must be an integer between 3 and 30 with a default of 10. The aiModelId must be a non-zero TEXT model ID from frontline ai-models list.

Can the Frontline agent builder upload custom avatar images?▼

The first public builder cut does not upload image files. Use existing image URLs or null for the avatar and bubbleImage fields when updating the agent theme.