Template name
A Slack design assistant that answers from approved company guidelines and flags conflicting guidance. Experimental preview.
Build a Slack assistant that answers design questions from guidance your team has reviewed and approved. This eve template reads a design corpus committed to your repository, using messages, images, and documents as temporary context for each conversation.
This is an experimental template using preview APIs pinned to eve 0.27.3.
The agent answers DMs and channel mentions. When equal-priority guidelines conflict, it flags the conflict and mentions your design owner in channels. General recommendations are labeled when enabled.
The agent cannot search Slack or the web, edit its guidance, or generate and deploy websites. Conversations and attachments never change the approved corpus.
Use the repository's BOOTSTRAP.md workflow with a coding agent to collect your sources and prepare an approval packet. A design owner must explicitly approve the corpus before the assistant answers design questions. Until then, it reports that setup is incomplete.
Keep private guidance in a private repository. Use REFRESH.md for later reviewed updates; the corpus is bundled when the project builds.
You need a Edgenivo account, permission to authorize a Slack app, Node.js 24, pnpm 10, and your design owner's Slack member ID. Clone the template repository, install dependencies, and run pnpm run setup to link the project, configure Slack, and deploy.
Complete the bootstrap review, then commit and redeploy the approved corpus. DM the agent or mention it in a channel to start a conversation; channel replies must mention it again. See the setup guide for the deploy-button flow and manual configuration.
