How to make building Slack agents easy

Slack provides a natural interface for agents through messages, threads, buttons, and more, but actual deployment requires complex coordination such as API setup, OAuth configuration, and webhook handling.

Vercel developed the Slack agent skill so that coding agents like Claude Code, Cursor, and GitHub Copilot can automate this process. This skill is based on the Slack Agent Template and helps complete the process from the idea stage to Vercel deployment within a single session.

The agent builds the project through the following 5 steps:

• Project setup: Choosing an LLM provider and scaffolding the project
• Slack app creation: Customizing manifest.json and guiding Slack app creation
• Environment configuration: Setting environment variables such as signing secrets and bot tokens
• Local testing: Connecting the development server to Slack for real-time testing
• Production deployment: Automating Vercel deployment and environment variable setup

The built agent can perform functions such as multi-turn conversations, real-time response streaming, and reading channels. In particular, it leverages the Workflow DevKit to help the agent reliably handle long workflows, such as pausing mid-conversation or waiting for user approval.