Skip to main content
AI coding agents can integrate Claimr for you, as long as they know how the widget loader behaves. Without that context, agents tend to make the same mistakes: injecting the script from the component that shows the widget, using the wrong script ID, or calling destroy() on unmount. This page gives you three ways to prevent that.
  1. Add a rules file to your repository so the agent follows the integration rules on every task.
  2. Point the agent at the Markdown docs so it can read the full guides.
  3. Connect Claimr MCP if you also want the agent to create or edit campaigns.

Add a Claimr rules file to your project

Copy the block below into the instructions file your agent reads, fill in the Project values section, and commit it.
AGENTS.md

Point the agent at the docs

Every page on this site is also available as plain Markdown, which agents read more reliably than HTML: Agents with web access, such as Claude Code, can fetch these URLs on their own when the rules file lists them. If your agent runs without network access, open the page you need here, use the page menu at the top to copy it as Markdown, and paste it into the conversation.

Example prompts

Use prompts like these after adding the rules file. Name the framework, the page, and the sign-in model so the agent does not have to guess.

Configure campaigns with Claimr MCP

The rules file teaches an agent how to embed a campaign in your code. To let the same agent build or edit the campaign itself (quests, tasks, rewards, widget layout, translations), connect it to Claimr MCP.
For Cursor, VS Code, and other clients, see Connect an AI assistant. With both in place, a single request such as “create a three-quest onboarding campaign and add it to the /quests page” can cover the campaign setup and the code.